US 3578298 A
Description (OCR text may contain errors)
United States Patent CALCULATOR CONTROL I SYSTEM 3,469,828 9/1969 Lane Primary Examiner-John J. Camby Att0rneysWilliam S. Wolfe, Gerald R. Woods, Oscar B.
Waddell and Joseph B. Forman ABSTRACT: A rotary cement kiln control system and method. Burning zone temperature and torque measurements generate a total process error apportioned to fuel and speed control for the kiln. The control system responds to short-term process disturbances to maintain thermal stability in the kiln and the contributions of the burning zone temperature and torque measurements are modified in accordance with thermal stability. Feedback representing expected variations in the measurements is provided. Unusual or adverse conditions are sensed to generate override signals. The effect of torque in the chain section of the wet kiln is also considered in control.
CHECK LOGIC OXY INSTRUIENTATION KlLN Patented May 11, 1971 7 Sheets-Sheet l G53 85.: oo
MYRON R. HURLBUT ROBERT A, PHILLIPS LELAND C.TUPPER Patented Ma 11, 1971 I 3,578,298
7 Sheets-Sheet 5 MEMORY EFUEL H EFUEL m5) ESPD H EFUEL "-20 ESPD n EFUEL 5 W EFUEL 40 ESPD 0 EFUEL 5 ESPD EFUEL ESPD n CHAMP" ARITHMETIC UNIT TTTl EFUEL ESPD CHAMP F8" Patentga 'Ma 11, 1971 3,578,298
-7 Sheets-She et 4 CALCULATE THE TRENOEO VALUE MEASURE, CHECK, AND FILTER PROCESS VARIABLES FEEO FFEED TIGscqn FTIG DRswn FDRn OXYGEN FOXY ELscDn FFUEL AMPscon -a- Fm" TBZscon FTBZ FETSCOII 0- OF TSCF TO OBTAIN TSCTn CALCULATE THE CHANGE IN AMPS CHAMP DUE TO THE CHAIN SECTION AS A FUNCTION OF TSCFn AND TSCTn CALCULATE BURNING zoNE TEMPERATURE ERROR E152 FROM FTBZ AND 1N2 CALCULATE STABILITY GAIN FACTOR 86F" FROM FTBZ CALCULATE NEW TRENDED AMP SETPOINT, ANP FROM FAMP AND sen,
USE FB 1o CALCULATE THE NEw STEERING FUNCTION, STF AND NEw BASE FUEL, AND FEED rms BACK 10 THE PROCESS NoDEL CALCULATE THE AMP ERROR, DELAMP FROM AMP AND FAMP DETERMINE A BURNING ZONE TEMPERATIRE DEVIATION DELTBZ FROM ETBZn AND SGF COMBINE THE TEMPERATURE ERROR, DELTBZ THE AMP ERROR, OELAMP WHICH ARE FUNCTIONS OF SGF THE FEEDBACK FBn AND STEERING FUNC- TION STF TO OBTAIN A TOTAL PROCESS ERROR SIGNAL, ERR
CALCULATE ,THE TEMPERATURE OF THE FEED LEAVING THE CHAIN SYSTEM FOR THE CONDITIONS OVER THE LAST CONTROL PERIOD, TSCF FIG. 4
Patented May 11, 1971 3,578,298
'7 Sheets-Sheet 5 IS ERR" FERR ,LYES
CALCULATE. FUEL CHANGE CALCULATE FUEL CHANGE Rol ERR EERR FROM FERRW, EERR IS ERR GREATER THAN no SPEED DEAOBANO K YES CALCULATE SPEED CHANGE FOR ERR ABOVE DEADBAND,SERR
GAS SYSTEM IN ANP RATE FORCING YES LAST CONTROL PERIOD j R 7 CALCULATE A SPEED CHANGE FROM THE AMP RATE AND one AM as THE RATE OF DECREASE 0F PERCENT OF THE LAST SPEED No FAHP SUFFICIENT TO PUT IT CHANGE THEN USE THE LARGER BELOW SETPOINT m N SCANS AND CHANGE OF THE TWO IS IT DECREASING AT KRMm AMPS PER SCAN OR MORE 7 E L IS THE SELECTED CHANGE LESS A THAN A MINIMUM CHANGE YES SCAN TO-SATISFY THESE CONDITIONS YES (suns THE SECOND CONSECUTIVE TERNINATE THE AMP RATE FORCING CONDITION INITIATE ANP 'RATE FORCING AND CALCULATE A SPEED CHANGE PROPORTIONAL TO THE AMP RATE HMA I Pate ted-Ma 11,1971
7 Sheets-Sheet 6 IIETERHIHE .FORCIHG TEIIRERATuRE TEIIR AS A FUNCTION OF ETHz AND THE RATE OF GHAHGE 0F TBZ, IIsTT HAs THE sIrsTEIH IH HIGH TENP- YLS ..E.,. ERATuRE EoRcI IG LAST SCAN? I I I N0 ARE THE coHmTIoHs 0N TEHRn AND WSTTTI SATISFIED To TERTI- No ARE THE GoHoITIoHs 0N TEHP IHATE HIGH TEHPEHATuRE FORCING? AHo wsTT SATISFIED To Es- YES TABLISH HIGH EHR AND FORCING? YES I TERHIHATE HIGH TEMPERATURE FORCING GoHmTIoH ESTABLISH HIGH TEHRERATuRE FORCING CALCULATE A SPEED oEcREAsE FRON TEMP", HTEHH. AND GET FUEL/SPEED COUPLING T0 GAIGuLATE THE PROPER FUEL DECREASE (WAS THE SYSTEM IN LOW TENS YES ERATURE FORCING LAST SCAN? ARE THE CONDITIONS ON TENP AND WSTT SATISFIED. T0 TERMINATE No ARE THE C(NIDITIONS ON TENR AND LOW TENPERATURE FORCING? WSTT SATISFIED TO ESTABLISH YES LOW TENPERATURE FORCING? I YES TERNINATE LOW TEMPERATURE FORCING CONDITION ESTABLISH LON TENKRATURE FNCINC CALCULATE A SPEED DECREASE FRDN TENP LTENP AND SET FUEL SPEED COUPLING TO CALCULATE THE PROPER FUEL DECREASE 5 I FIG. 4B
Patented May 11, 1971 3,578,298
7 Sheets-Sheet 7 COMBINE ALL SPEED CHANGES CALCULATE FEED END TEMP- TO OBTAIN THE TOTAL SPEED ERATURE SETPOINT FET AS CHANGE DELSPDn A TRENDED VALUE OF T E FEET PUT TALL SPEED CHANGES, EXCEPT THAT A UUE To HIGH TEMPERATURE, DETERMINE FEED EMU TEMP- BACK INTO THE PROCESS MODEL ERATURE ERRoR UFET FRoM FET AND FFET I cALcULATE SPEED SETPOINT, KSPD FRoM BASE SPEED, KSPD IS THE MAGNITUDE 0F UFET AND SPEED cHAMGEs, DELSIDII GREATER THAN A DEADBAND? YES UETERMIIIE THE ToTAL DESIRED CALCULATE A FUEL cHAHGE FUEL OUTPUT-FROM THE FUEL FoR THE ERRoR OUTSIDE THE GHAHG oUE- To ERR, AND DEADBAND fDFET AMU ADD THE BASE FUEL, FUEL IT To OBTAIN THE FUEL SETPOINT, FUEL FEED THIS FUEL BACK INTO THE PROCESS MODEL LOGIC MEASURE THE FILTERED VALUE OF OXYGEN LEAVING THE KILN FOXYn, THE OUTPUT VALUES FOR EXIT GAS RATE COMPENSATE THE FUEL BY THE EXIT FUEL RATE SETPOINT SPEED FUNCTION TO REDUCE FUEL KILN SPEED KSPD FUEL AS A FUNCTION OF SPEED AND A FEED RATE RATIO SETPOINT FRSP (IS HIGH TEMPERATURE FORCING YES I IN EFFECT? DETERMINE NECESSARY CHANGES m IN FUEL SETPOINT FUEL;
- I AND EXIT GAS RATE EXI n TO MAINTAIN THE OXYGEN Is ETBZ ABOVE THE HIGH N0 LEAvIHG THE KILN AT A TEMPERATURE FUEL OVERRIDE sAFE LEVEL THREsHoLU 1 YES I I OUTPUT HEM SETPOINT VALUES cALcULATE FUEL DECREASE UUE oF KSPD FUEL AND To THE HIGH TEMPERATURE AND ExIT To THE PIIoEss SUBTRACT IT FRoM THE OUPUT FUEL FIG. 4
METHOD AND APPARATUS FOR CEMENT KTLN CONTROL BACKGROUND OF THE INVENTION This invention relates to the production of cement in rotary kilns and, in particular, to an improved method and apparatus for controlling and regulating the operation of rotary cement kilns to provide stable kiln operation with resulting uniformity of product quality and improved fuel efficiency.
Typical rotary kilns employed in the production of portland cement are steel cylinders 8 to 25 feet in diameter and between l and 700 feet long. The cylinders are lined with refractory brick and inclined 2 to 3 from the feed end to the discharge end. The steel cylinder is supported at spaced points and rotated through a gear drive by an electrical motor at speeds in the order of to l20 revolutions per hour. Cement raw material, such as finely ground limestone, clay or shale intermixed in the desired proportions and either in the form of a finely ground slurry or a dry pulverized, intermixed material are fed into the upper or feed end of the rotary kiln.
During rotation of the kiln, the raw materials move slowly down the kiln at a rate which is a function of the kiln rotational speed and pass through successive zones known as the drying or chain zone, the preheating zone, the calcining zone and the clinkering or burning zone. If the raw materials enter the feed end of the kiln in the form of a wet slurry, the moisture is evaporated in the chain zone which may extend for percent of the kiln length. Chains are suspended from the kiln to contact the slurry and serve as a heat exchanger to drive off moisture. The drying or chain zone is not necessary in a kiln which is specifically adapted to use only a dry mix. As the materials move down the kiln, they are slowly heated by a stream of hot gases which are produced by a burner positioned at the lower or discharge end of the kiln and which flow counter to the direction of material movement in the kiln. A fan at the feed end of the kiln creates a slightly negative pressure in the kiln and draws the hot combustion gases produced by the burner through the kiln to heat the raw materials mov-. ing in the opposite direction, causing the raw materials to undergo successive changes due to the steadily increasing temperature of the materials.
The temperature of the dried raw materials increases until the calcining temperature is reached at which time carbon dioxide is liberated from the raw materials, changing the carbonates to oxides. The calcining zone occupies the major portion of the kiln length. The temperature of the material changes little within the calcining zone since the calcining reaction is endothermic and requires heat. A measurement of the material temperature within this zone gives little indication of the degree of calcination. At a point down the kiln where calcination is complete, a large temperature difference exists between the solid materials and the counter-flowing hot gases. Thus, when calcination is complete, the temperature of the solid material begins to increase rapidly to the point where the exothermic clinkering reactions are initiated. The heat generated by these chemical reactions causes the solid material temperature to rapidly increase 700-800 F. The clinkering or burning zone is near the discharge end of the kiln and the material remains at or near the high temperature until it leaves the kiln and is thereafter cooled. The degree of completion of the chemical reaction in the clinkering or burning zone depends upon the feed composition, the temperature in this zone and the residence time of an increment of feed within the zone.
The kiln must be controlled in such a manner as to produce a clinker product having a satisfactory quality and preferably a uniform quality. The variables over which a kiln operator has immediate control and which directly influence the kiln operation are the kiln feed rate, i.e. the rate at which the raw materials are fed into the upper end of the kiln, the kiln rotational speed, the fuel rate, i.e. the rate at which fuel is injected into the kiln and burned, and the exit gas rate, i.e. the rate at which the combustion gases and other gaseous kiln products LII are drawn through the kiln and exhausted from the feed end into the atmosphere. The kiln operator attempts toselect values for each of these control variables which will result in stable kiln operation producing a desirable product at the required product volume.
In early rotary cement kilns, the operator visually observed the color of the burning zone, the position of the boundary between the-calcining and burning zones and the clinker size and consistency and took corrective action based upon these observations, using judgment gained bypast experiences. In general, kiln performance based on this type of control was poor in terms of product quality, product uniformity and fuel efficiency. More recently, elaborate instrumentation has been employed to sense various parameters during kiln operation. This provided the operator with more information of higher accuracy for determining proper control action. However, the results obtained are still a function of the operator's interpretation of the measurements and his judgment.
The range over which any controller is effective is a direct function of how well the relation between the manipulated variable and the controlled variable has been defined. If the relation is known exactly and never changes, then a controller can be defined which will perfomt satisfactorily for all deviations of the controlled variable from the set point value. However, in most processes, the relation between the manipulated and controlled variables can only .be approximated and a control relationship, for example, in the form of an equation,-is derived which fits this relationship over a reasonable range. Consequently, the more complicated the relation between the manipulated and controlled variables, the more sophisticated the equation required to provide a controller which will function over a reasonable range. The range of effective control of a controller may be defined by specifying the maximum allowable deviation of the controlled variable from a normal or setpoint value.
if the controlled value deviates outside the effective range of the controller, the controller may be incapable of exercising control over the process and external intervention is normally required to maintain control. In a cement kiln control system, control of the kiln may be lost if burning zone conditions deviate outside the effective control range. For example, if the length of the burning zone becomes too short due to a disturbance in kiln operation, the amount of dense clinker material in the burning zone may not be sufficient to resist the force of the feed dammed up behind the burning zone. Consequently, the feed may push forward to the discharge end of the kiln without completion of the clinkering reaction resulting in a poor quality product..The corrective action initiated by the control system may hot be sufficient to prevent this occurrence. Accordingly, it is desirable to detect and to anticipate deviations of the controlled variable outside the effective range of the controller and to correct the disturbance before control is lost.
Two diverse controllers have been used in the prior art. A
first controller responded to burning zone temperature measurements to control either the heat input to the kiln or the kiln speed because the burning zone temperature is the best representation of conditions in the kiln which is known.-Notwithstanding severe environmental difficulties which hamper the measurement, burning zone temperature has remained as the primary variable for process control. When it is used alone to control the heat input to the kiln, kiln speed or both, the measurement is additionally and independently susceptible to another ambiguity. A thermal gradient exists alongthe kiln length, and the burning zone temperature is defined at a specific kiln position. It is possible, however, for the material in the kiln to shift longitudinally. Such longitudinal shifts will alter the thermal gradient and may cause erroneous changes in the measured burning zone temperature. These errors can lead to incorrect control action and, in severe cases, positive feedback action with loss of kiln control.
The second controller is commonly known-as an amp controller. It responds to the motor torque variations required to turn the kiln by altering the heat input, kiln speed or both. This variable is easily measured and measurement techniques are generally not sensitive to the environment. It is also generally unaffected by variations in the thermal gradient. However, the .drive torque is not directly related to the burning zone temperature and is affected by changes in the kiln. While certain compensation techniques are available, it is difficult to compensate for buildup on the kiln walls and drive function variations especially when they are coupled with process variations such as changes in the feed composition.
Certain control systems have implemented both temperature controllers and amp controllers. However, each controller has been used independently. Hence, the problems of each controller remain in the total control system.
Therefore it is an object of this invention to provide an improved apparatus and method for controlling a rotary cement kiln.
Another object of this invention is to provide an improved apparatus and method for controlling a rotary cement kiln which compensates for certain measurement ambiguities.
Still another object of this invention is to provide an improved apparatus and method for controlling a rotary cement kiln which compensates for ambiguities in measuring burning zone temperature.
Yet another object of this invention is to provide a method and apparatus for controlling a rotary cement kiln which compensates for ambiguities in torque measurements.
Yet still another object of this invention is to provide a method and apparatus for controlling a rotary cement kiln which combines temperature and amp controllers in a complementary manner.
SUMMARY In accordance with one aspect of this invention, separate measurements of the torque required to turn the kiln and the burning zone temperature are made. A first signal is generated in response to the error in the actual burning zone temperature when compared to a predetermined burning zone temperature. A second signal is generated in response to the torque measurements. The first and second signals are then combined to control the input to the kiln to maintain the burning zone temperature at the predetermined value. Only heat input control is exercised over a narrow band to process disturbances. For larger process disturbances kiln speed is controlled'in conjunction with the heat input. Out of limit temperature conditions are controlled in a nonlinear fashion. A base fuel level for the kiln is also controlled to move in accordance with current operating conditions in the kiln.
The subject matter of the invention is particularly pointed out and distinctly claimed in the concluding portion of the specification. The above and further objects and advantages of this invention may be better understood by reference to the following description taken in connection with the. accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram depicting a rotary cement kiln embodying and utilizing the present invention;
FIG. 2 is a block diagram illustrating a control system incorporating the invention and employed to control the operation of the rotary cement kiln of FIG. 1;
FIG. 3 is a block diagram illustrating the organization of the process model in the control system of FIG. 2; and
FIGS. 4, 4A, 4B, and 4C are a flow diagram illustrating the operation of the control system of FIG. 2.
DESCRIPTION OF AN ILLUSTRATIVE EMBODIMENT Referring to FIG. I, a typical rotary cement kiln with its associated equipment is schematically illustrated. Rotary cement kiln has at its upper end or feed end a kiln feed hopper l1 and a kiln feed pipe I2 for feeding blended raw materials 13 into the upper end of kiln 10. The raw materials normally include Al 0 Si0 Fe l) MgCO and CaCO plus small amounts of K 0, Na O and sulfur. The blended raw materials may be in the form of a dry powder or a slurry and may be preheated in a heat exchanger utilizing the kiln exit gases. The specific embodiment shown is particularly adapted for use with a slurry. A chain section 16 is suspended along the kiln adjacent to the feed end 14 to remove moisturefrom the slurry. A chain section 16 may or may not be used if a dry mixture is inputed to the kiln. If a chain section is used in a dry kiln, then it preheats the dry mixture. Kiln l0, inclined downward at an angle of approximately 3 from feed end 14 to discharge end 15, is rotated by an electric motor 20 shown here driving a pinion gear 21 that engages a ring gear 22 encircling and attached to kiln 10. As kiln I0 is rotated by kiln drive motor 20 through gears 21 and 22, the kiln rotation causes the raw materials or feed to slowly cascade forward, the rate of forward progress of the feed within kiln 10 being approximately proportional to kiln rotational speed. Motor 20 is normally controlled to drive kiln 10 at a predetermined constant rotational speed.
At the discharge end of the kiln, a fuel supply line 25 and a primary air supply line 26 are connected to a fuel-air mixing chamber 27. Natural gas, pulverized coal, oil or combinations thereof may be employed as fuel, the fuel being fed into line 25 from a suitable source. The primary air is forced through line 26 and into chamber 27 by fan 28.
The interior of kiln I0 is lined with a refractory material (not shown) which is capable of absorbing heat from flame 30 and transmitting it to the gases and feed travelling through kiln 10. The combustion gases and other gaseous kiln products are drawn through the kiln by an induced draft fan 31 which exhausts the gases through a dust collector and stack 32. Induced draft fan 31 creates a slightly negative pressure in the kiln drawing secondary air from clinker cooler 35 through the kiln. The gases emerging from feed end 14 of kiln 10 pass through a series of dust collectors 37 which recover the dust and an exit gas damper 38. The dust may be reintroduced to the kiln through a conduit 33 to a dust feeder 34.
As the feed proceeds slowly down the kiln, it is heated by hot gases flowing counter to it and also by the heated refractory walls of the kiln. The temperature of the dry feed increases until the calcining temperature is reached. At this point, the calcium carbonate CaCO and the magnesium carbonate MgCQ, begin to decompose, forming CaO and MgO. The released carbon dioxide CO joins the combustion gas and is drawn from kiln 10 by fan 31. The zone of kiln 10 where this reaction occurs is called the calcining zone. This reaction continues over a major portion of the kiln length. The temperature of the feed changes very little within this zone since the calcining reaction is endothermic and requires heat. A measurement of feed temperature within this zone will not give a meaningful indication of the degree of calcination of the feed.
At the point in kiln 10 where calcination of the feed is complete, a large temperature difference exists between the feed and the combustion gases and, therefore, a rapid increase in feed temperature results. The temperature at which the exothermic clinkering reaction occurs is reached quickly and the heat generated by the clinkering reaction causes the temperature of the feed to increase still further to the point where the solids become partially liquefied. The clinkering reaction for the formation of (CaO) '(SiO (CaO) -(AI O (CaO) -(A1 09- 0 which are the crystalline compounds that determine the physical properties of the cement, occurs rapidly. The resulting partly fused mass of varying size continues to move down the burning zone of the kiln and remains near its maximum temperature until it nears discharge end 15 of the kiln. While at this temperature, most of the remaining CaO combines with the (CaO) -(SiO-,) to form (CaO) -(SiO,). The degree of completion of this clinkering reaction depends upon the feed composition, the temperature in the burning zone and the residence time of an increment of feed within the zone.
As the hot clinker material approaches the end of the kiln, it begins to lose some of its heat to the incoming secondary air.
At the exit end of the kiln, the clinker drops onto a travelling grate 40 usually reciprocated by a motor 4011. Air is blown through the grate 40 by fan 41 to cool the clinker. Part of the resulting heated air becomes secondary air which is drawn through kiln by fan 31, the remainder being exhausted by fan 42 through dust cyclone 43 to the atmosphere. The cooled clinker is transported by conveyor 45 to grinding apparatus (not shown) which pulverizes the clinker to form cement.
Process Variable Measurement A number of sensors are provided to monitor various parameters of kiln operation and to generate electrical signals representing the values of these parameters. These signals are employed by the control system of the invention to direct the operation of the kiln. As illustrated in FIG. 1, a signal indicating the rate at which feed is being supplied to the kiln is provided for control system 50 by a feed rate detector 51 associated with the' hopper 11 on a line 52. A temperature measuring device 53, for example a thermocouple, is provided near feed end 14 of kiln to provide a signal indicating the feed end and temperature transmitted to control system 50 on line 54. Analyzer 55 is also provided near the feed end 14 of the kiln to measure the oxygen content of the gases being exhausted from the kiln, the signal representative of the oxygen content being applied to control system 50 on line 56. A second temperature measuring device 57 is provided adjacent to termination of the chain section 16 of the kiln to provide a signal transmitted to control system 50 on line 58 indicating the temperature of the gases flowing through the kiln at that point.
A signal representing the rate of dust flow through the conduit 33 and dust feeder 34 is provided to control system 50 by a device 59 for measuring the dust flow rate, the signal being applied to control system 50 on line 60. A sensor 61 associated with kiln drive motor provides a signal to control system 50 on line 62 representing the rotational speed developed by the motor 20 as it rotates the'kiln. Sensor 63 as sociated with kiln drive motor 20 provides asignal to control system 50 on line 64 representing the torque developed by motor 20 as required to rotate kiln at a predetermined rotational speed. The temperature at the burning zone is determined by a burning zone temperature sensor 65 such as a color ratio pyrometer which optically determines the temperature at a prefocused point. The signal from the burning zone temperature sensor 65 is supplied to the control system 50 on line 66. A signal representing the rate of fuel flow to mixing chamber 27 is provided to control system 50 by fuel supply sensor 67 associated with the fuel supply line 25, a signal being supplied on line 68.
Control system 50 utilizes the information concerning kiln operation provided on lines 52, 54, 56, 58, 60, 62, 64, 66 and 68 to produce kiln control signals on lines 7071 and 72. The
control signal on line 70 represents a fuel rate set point and is applied to controller 73 to control the'rate of fiow of fuel to mixing chamber 27 and, therefore, the heat input to kiln 10. The control signal on line 71 represents an exit gas rate set point and is applied to controller 74 to control the speed of induced draft fan 31 and, therefore, the exit gas flow rate. The
signal on line 71 representing an exit gas rate set point may, alternatively, be employed to control the position of damper 38, thereby adjusting the exit flow rate. The control signal on line 72 represents a kiln speed set point and is applied to controller 75 to control a rotational speed of kiln drive motor 20 and, therefore, the rotational speed of kiln 10. Controllers 73, 74 and 75 are standard analog controllers as known in the art and will not be described in detail.
Signal Conditioning FIG. 2 illustrates the details of control system 50 shown in FIG. 1. Referring to FIG. 2, the signal on line 52 representing the drive feed rate to the kiln from the hopper 11 is applied to check logic 101. This signal FEED therefore represents the instantaneous feed rate to the kiln. Check logic 101 compares the present signal FEED with the previous signal F EED if the present and previous values differ by more than a given amount, it is assumed that some unusual conditions exist within the kiln and the output signal from check logic 101 is not used until it returns to within a reasonable range of the previous value of FEED,,,-,,,.. Check logic 101 serves to mask momentary or short-term disturbances and to indicate a failure of the sensor 5-]. Continued failure of check logic 101 to pass the signal FEED to filter 102 may also be used to implement an alarm function, Check logic 101 may be implemented with a digital computer.
Filtering and smoothing of the FEED signal to remove noise and other signal variations unrelated to the feed rate is performed in filter 102. The output signal of filter 102 is FFEED,.. The filtering action of filter 102 is described in the equation:
Where FFEED is the new filtered value,
FFEED, is the last filtered value,
FFEED is the present scan value, and
K is the filter constant.
The function of filter 102 may conveniently be performed in a I digital computer with K FFEED and FFEED,, being stored in the computer memory. This calculation is performed at short intervals, for example every 5 seconds, to insure that signal FFEED, represents the current condition of the feed rate. K is selected to be small enough to eliminate noise and other irrelevant effects, but not so small as to damp out the signal.
Check logic 103 receives successive values of a signal TlG. mm representing thegas temperature in the kiln adjacent the discharge end of chain section 16 which appears on line 58 from sensor 57. If two successive signals differ by more than a given amount, it is assumed that the measuring device, such as a thermocouple, has failed and the present value of TlG is not used. Rather, the previous value of TlG is used. Check logic 103 can also respondto a continued failure to pass the signal TlG to filter l04by initiating an alarm function. Filter 104 receives TlG,,,-,,,, and filters it in accordance with the following equation:
Where FllG, is the present filtered value,
F'TlCv is the previous filtered value,
TIG is the present measured value, and
K is the filter constant.
Typically, the FTIG can be calculated once every minute. The functions of filter 88 and check logic 89 may be conveniently performed in a digital computer with signals F1'lG,,, PUG, and TlG and the constant K being stored in the computer memory.
A signal representing the'rate at which dust is admitted to the kiln 10 through thedust feeder 34 is generated by dust rate sensor 59 and coupled to control system 50 on line 60. The signal DR is monitored by check logic 105 to determine whether the signal is accurate representation of the actual dust rate. As with other check logic, check logic 105 can mask momentary disturbances and, through alarm functions, can detect a sensor failure. The signal DR when coupled through check logic 105, is applied to filter 106 which generates the signal FDR, in accordance with the following equation:
Where FDR, is the present filtered value,
FDR, is the last filtered value,
DR is the present measured value, and
K is the filter constant.
Filter 106 thereby performs similar functions to filters 102 and 104 with respect to the dust rate signal. The functions of check logic105 and 106 may conveniently be performed in a digital computer.
Check logic 107 receives on line 54 a signal representing gas temperature at the feed end 14 of the kiln, as measured by device 53. The gas temperature information represented by the signal is coupled through check logic 107 to mask out mo mentary disturbances and verify the operation of the sensor 53. The signal, FET is then applied to filter 108 to be filtered in accordance with the following equation:
Where FFET, is the present filtered value,
FFET is theprevious filtered value,
FET is the present measured value, and
K is the filter constant.
The filter constant K,,,, is selected to be small enough to eliminate noise and other effects on the signal for the rate of calculation, an example being a calculation each minute. The functions of check logic 107 and filter 108 may be con veniently perfonned in a digital computer with signals FFET FFET,,,,, FET and K being stored in the computer memory.
The oxygen content of the exit gas is determined by analyzer 55 to generate OXY on line 56 which is coupled through check logic 109 to filter 110. Check logic 109'can perform one or more of the functions of check logic 101 for this particular signal. Filter 110 provides an output signal FOXY, in accordance with the equation:
Where FOXY,, is the present filtered value,
FOXY is the previous filtered value,
OXY is the present measured value, and
K is the filter constant. a The functions of checklogic 109 and filter 110 can also be conveniently performed in a digital computer with FOXY,,, FOXY OXY and the filter constant K being stored in computer memory.
Another process parameter which is used in certain embodiments of the control system 50 is the rate of fuel flow to the fuel air mixing chamber 27. Fuel supply sensor 67 generates a signal FUEL on line 68. After being examined in check logic 111, the signal FUEL is applied to filter 112. Filtering action of filter 112 can be described by the following equation:
Where FFUEL, is the present filtered value,
FFUEL,,,, is the previous filtered value,
FUEL,,,.,,,, is the present measured value, and
K is the filter constant.
The value of constant K,,,,,, will be dependent upon various process parameters and the rate at which the signal FFUEL, is
calculated. The function of check logic 111 and filter 112 may be conveniently performed in a computer with the signals FFUEL,,, FFUEL, and FUEL and the filter constant K,,,,,, being stored in computer memory.
Signal AMP generated by sensor 63 on motor 20 and coupled to control system on line 64 can represent the instantaneous heat state within the kiln while changes in the value of AMP can indicate corresponding changes in the condition of the burning zone. If kiln drive motor 20 is an AC motor, assuming a constant speed of rotation of kiln 10, the signal on line 64 is a measure of the kilowatt power input to motor 20 which represents the torque developed by motor 20 to rotate the kiln 10. 1f kiln drive motor 20 is a constant field, DC motor, the signal on line 64 is a measure of the armature current of kiln drive motor 20 which represents the torque developed by motor 20 to rotate kiln 10, with constant field and supply voltages. For purposes of this description, motor 20 is assumed to be a constant field, DC motor; and the signal AMP on line 64 represents armature current in and torque developed by motor 20. Various means for obtaining a torque signal from other motors are known in the art. The use of torque measurements represented by the signal AMP to effeet control of kiln operation independent of actual burning zone measurements is claimed in U.S. Pat. (Ser. No. 678,851), now U.S. Pat. No. 3,469,828,-issued (filed Oct. 30, 1967) and granted to James W. Lane and assigned to the same assignee of the present invention. Filtering and smoothing of the AMP signal to remove noise and other signal variations unrelated to the condition of the burning zone, for example of kiln rotation on the signal, is performed, after examining the signal AMP and check logic 113, by a filter 114. The output signal of filter 114 FAMP is described in the equation:
Where FAMP,, is the new filtered value,
FAMP is the last filtered value,
AMP,,,., is the present scan value, and
K,,,,,,, is the filter constant. The function of filter 114 may be conveniently performed in a digital computer with K,,,,,,,, F AMP, and FAMP being stored in the computer memory. This calculation is performed at short intervals, for example every 5 seconds, to insure that signal FAMP, represents the current condition of motor torque forming an accurate basis for control action. K,,,,,,, is selected to be small enough to eliminate noise and the effect of kiln rotation on the signal but not so small as to damp out the signal. The function of check logic 113, which compares successive values of the signal AMP may also be performed in a digital computer with AMP being stored in memory.
The burning zone temperature sensor 65 generates a signal TBZ on line 66. This sensor may take several forms although the use of optical pyrometers or other optical measuring devices is widely accepted. Burning zone temperature signal TBZ is coupled on line 66 to check logic 115. Check logic 115 compares the present signal AMP with the previous signal. If the present and previous signals differ by more than a given amount, it is assumed that some unusual condition exists within the kiln or kiln instrumentation and the value of TBZ is not used until it returnsto within a reasonable range of the previous value. Check logic 115 thereby serves to mask momentary or short-term disturbances. It can also indicate the failure of the burning zone temperature sensor 65 and initiate an alarm function. The function performed by check logic 115 may be conveniently implemented in a digital computer with successive signals from the kiln instrumentation being stored in computer memory.
The output signal of filter 116, FTBZ,,, which is energized by the signal TBZ,,,.,,,,, is described by:
Where FTBZ, is the new filtered value,
FTBZ is the last filtered value,
TBZ is the present scan value, and
K is the filter constant.
The function of filter 116 may conveniently be performed in a digital computer with FTBZ,,, FTBZ and TBZ and filter constant K being stored in computer memory. The value of K depends upon the repetition rate of calculations or FIBZ, and other circuit parameters and selected to be small enough to eliminate noise and other irrelevant effects on the signal TBZ Control system 50 responds to actual process measurements represented by the signals FFEED,., FTlG FDR,,, FFET,,, FOXY,,, FFUEL,,, FAMP,,, and FTBZ, to perform various control functions on the heat input, kiln speed and exit gas rate in the kiln. in accordance with one aspect of this invention, the signals FAMP, from filter 114 and F'TBZ, from filter 116 are combined to control the fuel rate. In response to the filtered signals FAMP, and FTBZ, control system generates an output signal FUEL on line 70 which is coupled to controller 73 to thereby control the quantity of fuel supplied to fuel-air mixing chamber 27.
Basic Controller The control system of FIG. 2 is shown in a preferred embodiment. As will become obvious in the following discussion, certain of the objects and advantages of this invention may be obtained by eliminating certain sections or control loops in the control system 50. A basic embodiment of a control system utilizing this invention would primarily respond to burning zone temperature represented by signal FTBZ, which is converted to an error signal ETBZ, summing amplifier 120 additionally responsive to a burning zone temperature set point signal TBZ The burning zone temperature set point represented by the signal TBZ,,,, is controlled by the operator by means of a potentiometer, a value stored in a digital computer or other equivalent means and will normally be based upon the chemical analysis of the kiln product periodically reported to the operator. For example, if free lime (uncombined CaO) content of the kiln product is too low, the operator will decrease the burning zone temperature set point. Whereas if the free lime content is too high, the operator will increase the burning zone temperature set point. Typical burning zone temperature set point values for a particular type and rate of feed and for a particular type kiln product based on past experience would be employed by the operator to select an ini- The burning zone temperature signal ETBZ, is positive if the I present filtered value of burning zone temperature is less than the burning zone temperature set point and is negative if the present filtered value of burning zone temperature exceeds the burning zone temperature set point. The function of summing amplifier 120 may conveniently be performed in digital computer. The signal ETBZ, is then coupled to a summing amplifier 121 as signal DELTBZ,,. As specifically shown, that this signal is coupled through a multiplier 122 which is additionally responsive to the burning zone temperature value FlBZ, and function generator 123. However, in accordance with this specific embodiment being described multiplier 122 and function generator 123 are not necessary so DELTBZ,,=ETBZDn. Their utilization in the control system of FIG. 2 will be described hereinafter.
Another summing amplifier 124 responds to the difference between a kiln amp set point AMP and the present filter value of kiln amps represented by signal FAMP,,. A filter 125, energized by the present filter value of kiln amps FAMP, normally controls the value of the amp set point signal AMP,,,,. The filter 125 may be a fixed value filter which provides a trend set point signal in accordance with a trending function. In the preferred embodiment the rate at which the set point is trended is a function of a stability gain factor explained hereinafter. Generally, the faster the set point is trended the less the influence of kiln amps on the control system. The operator may override the signal produced by filter 125. Summing amplifier 124, therefore, produces a signal DE- LAMP representing the difference between the kiln amps set point AMP and the present filtered value of kiln amps represented by signal FAMP,,, as expressed by the equation:
Amp error signal DELAMP, is positive if the present filter 7 value of kiln amps is less than the kiln amp set point and is negative if the present filter value of kiln amps exceeds the kiln amps set point. The function of summing amplifier 124 may conveniently be performed in a digital computer.
The outputs of summing amplifier 120, ETBZ,,, and summing amplifier 124, DELAMP,,, are applied to the summing amplifier 121 to be combined with a feedback signal.
through changeover circuit 127, summing amplifier 126, additional summing amplifiers 134 and 135 and filter 136. The output of filter 136 is a signal EFUEL, which represents a fuel error at that time in the-process. In accordance with this embodiment of the invention, the fuel error signal EFUEL, is placed in process model 133 -to be delayed for a time equivalent to the process delay inreacting to a change in fuel. At the time of the delay, the process model 133 uses a previous value of the fuel error signal to generate a feedback signal FB,, which accounts for previous control actions on fuel and this is applied to summing amplifier 121. The output of summing amplifier 121 can, for purposes of this specific discussion, be represented by the equation:
ERR,,=DELTBZ,,+DELAMP,,+FB,, where ERR represents the present total composite error signal. The function of summing amplifier 121 may conveniently be performed in a digital computer. The composite error signals ERR,, may, in a basic controller, constitute the fuel set point FUEL,,,,, the fuel set point signal FUEL on line 70'then being applied to controller 73 to adjust the fuel set point. In this basic controller, therefore, both the burning zone temperature and the motor torque required to turn the kiln have been combined to generate, with appropriate feedback, a total error signal which'controls the heat input to the kiln by controlling the fuel to the primary air fuel mixing chamber 27.
Improved operation of this basic controller can be obtained by implementing the functions of multiplier 122 and function generator 123 to control filter 125 and thereby trend the kiln amp set point AMP in accordance with changes in the burning zone temperature represented by the filter value FTBZ,,. The output of summing amplifier 120, ETBZ,,, is applied to the multiplier 122 together with the output of function genera tor 123 which is a signal SGF,,. If signal SGF is zero, there is no contribution of temperature error to the controller and all control is based on kiln amps. The output of function genera- I tor 123 is a stability gain factor and is used because the burning zone temperature apparently becomes more unreliable as a control variable when the kiln becomes upset and cycles due to unusual disturbances. Conversely, as the kiln becomes stable, the burning zone temperature becomes more reliable; the amp set point for proper kiln operation drifts and attempts to hold it at a constant value cause over or under heating of the kiln. As previously indicated, during upsets it is possible for the burning zone temperature measurement to provideunreliable indications of the conditions in the kiln as the temperature profile of the kiln shifts longitudinally as a result of upsets. In these cases, the kiln amp signal can give an earlier indication of actual changes in kiln condition so that kiln amps become a better control variable to stabilize an upset kiln. Multiplier 122 and function generator 123 combine with the filter 125 to provide improved operation of the basic controller by dynamically controlling the selection of or the relative emphasis on the amp and temperature signals based on the stability of the process.
Process stability can be measured by a signal WSTI), which is variable in accordance with the filtered derivative of the burning zone temperature signal in accordance with:
Where K through K are the filter constants, and the value of m is one less than the-number of terms in the filter,
FTBZ is the present filter value,
FTBZ, is the previous filter value, and
FIBZ is the filter value m times before the present filter value.
Normally, five filter sections are used so that m=4 to provide a rate of change of temperature calculated as a weighted average of the change over the last five calculations of the burning zone temperature. The weighted scan trend of temperature signal generated within function generator 123 is then filtered o obtain a filtered signal in accordance with equation:
Where FAWST is the present filter value,
FAWST is the previous filter value,
/WSTl",,/ is the absolute value of the weighted scan trend of temperature, and
K is a filter constant.
The stability gain factor SGF is limited to a positive value and normally should not be permitted below a minimum value. For example, values of the signal SGF, between 0.2 and L indicate a normal range. 1
The stability gain factor signal SGF,, is then applied both to multiplier 122 and filter 125. As a result, the error value from multiplier 122 is the product of the stability gain factor SGF and the burning zone temperature error signal ETBZ,,. Filter 125 in this form is energized by the signal FAMP, and the sta bility gain factor signal SGF, to normally produce an amp set point signal AMP in accordance with:
Where I AMP,,,,,, is the new kiln amp setpoint,
AMP is the last kiln amp setpoint,
' FAMP, is the present filtered kiln amp value,
SGF,, is the present stability gain factor, and
K is the filter constant.
Therefore the stability gain factor varies the kiln amp setpoint. It also varies the burning zone temperature deviation represented by DELTBZ, in accordance with the equation:
Where DELTBZ, isithe new compensated temperature deviation,
ETBZ, is the present burning zone temperature error, and
SGF, is the present stability gain factor.
As the process becomes unstable, the gain on the burning zone temperature error DELTBZ, is decreased to that the more stable amp signals are emphasized. However, when the system is stable, the gain factor is maximized so the burning zone temperature signals become the primary control variable. In this manner, the control emphasis of the two input signals is constantly varied to achieve control on the fuel set point FUEL which is responsive to the more reliable control variable existing at a particular time.
As will be obvious the functions of multiplier 122, function generator 123 and filter 125 may also be conveniently perfonned in a digital computer.
Cement kilns are known to be nonlinear processes that must be maintained within relatively narrow controllable regions. When a cement kiln becomes cold, the clinker quality is unacceptable. More extreme conditions can stop the clinkering process and extinguish the flame. Whenever a trend to a cold condition is detected, it is necessary to take control action that will correct the trend as quickly as possible without upsetting the kiln. In addition to the feed rate and the fuel rate control, it is also possible to control the speed of the kiln. The process response to speed changes is faster than the response to fuel changes. Hence, another embodiment of the basic controller responds to large disturbances by controlling kiln speed while using fuel control in response to smaller disturbances which can take longer to correct. In accordance with this addition to the basiccontroller, such a procedure is possible which then allows the production rate, which otherwise is reduced when the kiln speed is reduced, to be maximized. Further, speed changes are implemented only as necessary thereby avoiding the generation of other upsets in other portions of the kiln, especially wet kilns where a speed change can cause an upset adjacent the feed end which propagates through to upset the burning zone conditions and thereby cause subsequent cycling.
In accordance with the control system shown in FIG. 2, the output signal from summing amplifier 121 ERR, is applied to changeover circuit 127 and to changeover circuit 140. The output signal from changeover circuit 127 is FERR and is calculated in accordance with equation: I
Where FERR, is the present effective fuel system error,
ERR, is the present composite error, and
K is a proportionality constant.
Changeover circuit 127 also limits the value of FERR, to a maximum value FERR Changeover circuit generates a signal SERR in accordance with the equation:
Where SERR is the present effective speed error,
ERR,, is the present composite error signal,
K,,,.,, is a proportionality constant, and
K,,,, is a dead band constant (i.e., K,,,,=FERR,,,,,,). Changeover circuits 127 and 140 permit the switching of control modes between speed and fuel. The system errors represented by signal ERR up to a threshold value, indicated by the change to a zero slope line in changeover circuit 127, and the change to a negative slope line in changeover circuit 140, pass through the changeover circuit 127 and are blocked from passing through the changeover circuit 140. Therefore, signal errors up to the threshold value only affect fuel set point. Errors above the threshold value pass through the changeover circuit 140 and affect the kiln speed.
The signal SERR,, which represents the portion of the composite error signal from summing amplifier 121 which will be used to control the kiln speed is applied to the input of summing amplifier 141. The output signal shown in FIG. 2 is DSERR,,. lf amp rate forcing and low temperature speed control are not utilized, then DSERR equals SERR and, in fact, summing amplifier 141 is not necessary. The amp rate forcing and low temperature speed control concepts will be discussed hereinafter. The output of summing amplifier 141 is then coupled to the process model133as a speed error through filter 142. Filter 142 produces an error speed signal in accordance with:
Where ESPD is the present filtered value,
ESPD is the previous filtered value,
DSERR, is the present speed error signal, and
K,.,',,,,, is a filter constant.
The function of filter 142 may be conveniently performed in a digital computer with the signals ESPD,,, ESPD DSERR,, and proportionality constant K being stored in the computer memory. With this feedback signal, it can be seen that the feedback signal FB,, from the process model 133 is a function of the fuel and speed error signals EFUEL, and ESPD,,.
Signal DSERR is then coupled through summing amplifier 143 to generate a signal DELSPD, which is equal to the signal DSERR, if high temperature speed control is not utilized. in summing amplifier 144 the speed error signal DELSPD, is combined with the base speed which is inputed by the operator KSPD,,,, to generate the output signal on line 72 KSPD, in accordance with:
During normal control operations when the composite error signal ERR is of a magnitude which permits fuel control, the kiln speed is maintained at the base speed KSPD However, for positive errors exceeding the threshold value of changeover circuit 140, the fuel signal FUEL remains at a maximum, except as provided hereinafter, while the kiln speed is varied to cause more rapid control action and reaction by the kiln. Changeover circuits 127 and 140 and summing amplifiers 143 and 144 may be conveniently implemented by digital computers.
It has been found, however, that when kiln speed is altered it is advisable to reduce the fuel rate. To accomplish this, function generator 145 is energized by the signals KSPD and DELSPD, and generates an output signal tSPD, which is coupled as an input to multiplier 130. The multiplier fSPD is generated by function generator 145 in accordance with:
- K (DELSPD,,+K i KSPD Where fSPD is a present value of a coupling factor,
DELSPD, is a negative number representing the present speed error,
KSPD is the present base speed,
K and K are functions of the required forcing function which vary depending on whether the particular function generator 145 is adapted for high or low speed forcing.
For purposes of further understanding function generator 145,
assume that it is adapted for low temperature forcing. The output of function generator 145 has been coupled to multiplier 130 to produce an output signal XTFERR, according to:
Where XFERR, is the present value of the factored or multiplied signal -TFERR,,.
This produces coupling between the fuel and speed calculations. Whenever the kiln speed is reduced, the total feed entering the kiln is decreased which means that less total heat will be required at the feed end to sustain the same conditions. This problem is especially critical in a wet-process kiln where about 40 percent of the total heat is exchanged in the chain section for evaporation of water in thefeed. If too much heat is put in the chains for the amount of feed going through, overheating occurs as the feed progresses through the kiln; By adding function generator 145 and multiplier 130, both of which may conveniently be implemented by a digital computer, fuel is decreased whenever speed is decreased with the proportion of fuel decrease being specified by constants K and K,,,. In this manner, the kiln adjacent the feed end is controlled without waiting for the feed end temperature to react.
As previously indicated, the operation of the control system shown in FIG. 2 is enhanced by the use of process model i 133. Process model 133 permits control of the process by accounting for all changes caused by control actions. Any difference between actual process response and this predicted response is then interpreted as a process disturbance to becorrectcd by the controller. Therefore, predicted effects of the speed changes must be entered into the model to prevent these from looking like the process disturbance to the fuel controller in addition to fuel changes. As indicated earlier, the kiln speed changes are fed into process model 133 as the signal ESPD while fuel changes are fed as EFUEL,,. The error signal EFUEL, can be generated by summing amplifiers 126 and 134 and filter 136. However, such a system is effective only if the average operating level of the kiln does not change. The error signal ERR, reflects short-term disturbances as well as longterm changes. this is undesirable because changeover or threshold circuits 127 and 140 must operate on changes above the current operating level of the kiln. To eliminate the effect of long-term changes and permit the threshold circuits 127 and 140 to be responsive to short-term changes, a base fuel calculator 150, summing amplifier 151 and filter 152 are added to the circuit. This section of the circuit steers base fuel to an average operating value on a continuous or dynamic basis and, therefore, may be denoted as a dynamic steering unit. To accomplish dynamic steering; the feedback signal FB, is applied to filter 152 to generate an output STF',, representing a steering function in accordance with:
Where STF is the present value of the steering function,
STF,, is the previous value of the steering function,
FB, is the present feedback signal, and
K is a filter constant.
The filter constant is chosen so that the feedback signal P8,, is provided with a time constant delay. The output STF is then applied to summing amplifier 121, summing amplifier'135, and summing amplifier 151. This affects the composite error signal ERR, which is generated in the summing amplifier 121 in'accordance with:
ERR,,==SGF,,DELTBZ,,+DELAMP,,+FB,,-STF, The steering function also affects the calculation of the signal FUEL the output of the summing amplifier 151. The second input to summing amplifier 151 is a signal BASE generated by base fuel calculator 150. Base fuel calculator 150 generates a signal BASE,,,=FFUEL,, when control of the system is initiated. Therefore, the signal FUEL,,,,,,,. is generated in summingamplifier 151 in accordance with:
FUEL,,,,,,..=BASE,-,,+STF,, thereby causing the base fuel signal FUEL to vary in accordance with changes caused by previous control actions. The base fuel signal FUEL,,,,,,,. is applied to summing amplifier 126 so the signal representing the total fuel error TFERR, is calculated in accordance with:
The output of summing amplifier 134 represented by signal DFUEL is generated in accordance with:
DFUEL,,=TFERR,,FUEL,,.,,,,. The steering factor also affects the signal DFUEL, by being added thereto in summing amplifier 135 to generate a steered total fuel error STFUEL, in accordance with:
The input to process model 133 EFUEL, is then generated in filter 136 in accordance with:
Where EFUEL is the new'filtered value,
EFUEL,, is the previous filtered value,
STFU EL, is the present steered total fuel error, and
K is a filter'constant.
It will be obvious that the function of filter 136 can be implemented in a digital computer with signals'EFUEL,,, EFUEL,, and STFUEL,, and constant K being stored in computer memory. Likewise the functions of summing amplifiers 126, 134 and 135, base fuel calculator and filter 152 may conveniently be implemented in a digital computer with signals STF,,, STF and FB,, and constant K, for filter 152 being stored in computer memory.
Through these functions, dynamic steering effectively transfers long-term changes in the composite error signal ERR to permanent changes in the base fuel signal FUEL,,,,,,,,. ln accordance with this circuitry, the feedback signal P8,, is passed through filter 152 to generate a time constant. Although a quantity of fuel is addedin response to the feedback signal, a similar quantity is subtracted and added to the base fuel in summing amplifier 151'. it is also subtracted from the output of summing amplifier 121 leaving only shorter term changes in the composite error. As process model 133 muststill be responsive to total system changes in order to maintain the output of filter 152 at the changed value, the output of filter 152 is returned to process model 133 through summing. amplifier 135. 3
Process model 133, identified by reference symbol 133 in FIG. 2 and shown in detail in FIG. 3, comprises two delay tables in which the error signals EFUEL, and ESPD are stored each time a control action is taken. If, for example, a control value is calculated every 5 minutes to initiate a control action, if required, this control value is stored in the process model and control values previously stored are shifted through one storage position in the process model each time a new value is entered. Assuming an interval of 5 minutes between control value calculations, the fourth storage position on the table will contain the control value calculated 20 minutes earlier The actual delay incorporated in the process model is the time period which elapses between initiation of a control action and the kilnresponse to the control action as reflected by a change in burning zone condition, this delay being a function of the characteristics of a particular kiln. In a typical kiln, the delay between a control action causing a change in fuel set point and the response thereto in the burning zone may be of the order of 30 to 35 minutes or more. On the other hand, responses to control actions changing the speed set point and the response thereto in the buming zone of the kiln may be less. The delay table of the process model contains a sufficient number of storage positions so that the delay range available in the delay table for both error signals encompasses delay the process model table correspondingto the delay' between a control action for fuel and the reaction in the burning zone characteristic of the particular kiln, and ESP-D is the value of speed error signal calculated and stored in process model table corresponding of the delay between a speed control action and the reaction in the burning zone which is characteristic of the particular kiln.
If a wet kiln is being implemented and an improvement to be described hereinafter is utilized for generating a feedback input CHAMP, which is responsive to the effect of the change on torque, the feedback signal F8, is calculated in accordance with the equation:
Signal CHAMP, may be stored in memory 154 and coupled thereto by arithmetic unit 153 as shown in FIG. 3.
At this point, a control system for a rotary cement kiln has been described which senses motor torque in burning zone temperature and in response thereto controls the fuel set point and kiln speed set point. The parameter having the greatest effect on the subsequent control action is chosen in response to kiln stability with the burning zone temperature being the better control variable when the process is stable and the torque signal being the better control variable when the process is unstable. Means have been shown which respond to these two signals to determine the stability'by smoothly shifting emphasis from one to the other of the control variables. As a result, the composite error signal, which, through the use of dynamic steering and a dynamic process model, reflects only errors due to process disturbances, is then analyzed to determine its magnitude in comparison with a threshold magnitude. If the magnitude is less than the threshold, it indicates the disturbance can be corrected by varying fuel. If the disturbance is greater than the threshold, it indicates that more drastic action is necessary and the kiln speed is altered. Such speed changes, however, are coupled to the fuel control to decrease the fuel level to thereby tend to stabilize the feed end temperature when kiln speed changes.
Feed End Temperature Control As previously indicated, about 40 percent of the heat transfer in a wet kiln takes place in the chain section which has a definite influence on changes in torque. In addition, equipment constraints exist which dictate that the temperatures in the feed end of the kiln be maintained within given limits. Good control of the temperatures at the feed end should minimize disturbances that are propagated to upset burning zone conditions. Further if the feed temperature leaving the chain section is constant the remainder of the system should remain constant and the control should be minimized. However, the feed temperature in wet kilns as the feed leaves the chain section and dry kilns as the feed. leaves the heat exchange is'not available. A feed temperature can be calculated on the basis of other variables which are subject to error and which are unreliable. The feed end temperature, however, is a reliable and accurately obtainable signal and has indicated itself to contribute to good overall kiln control if the feed end temperature is held constant. Such control can be accomplished in a control loop comprising check logic 107 and filter 108 which generate the signal FFET, together with summing amplifier 160, filter 161, dead band amplifier 162, and gas temperature control 163. This control serves to maintain a relatively constant gas temperature near feed end 14 to provide a relatively constant source of heat for the feed entering the kiln and a relatively constant temperature profile from the discharge end to the feed end of the kilnv The gas temperature at the feed end of the kiln is thereby decoupled from control actions which vary the rate of fuel flow and, therefore, the rate of heat input into the kiln due to control actions initiated in the torque control loop or the burning zone control loop previously described. Normally as the fuel rate is increased or decreased to adjust burning zone conditions, the feed end temperature control loop adjusts the exit gas flow rate to maintain sufficient heat availability in the feed preparation section of the kiln comprising the drying end or preheating zones.
A feed end temperature error signal DFET,, is generated in summing amplifier by comparing the actual feed end temperature represented by filtered signal FFET, with a trended set point signal PET, generated by filter 161. Filter 161 produces an output signal FET,,,, in accordance with:
Where FET,,,,,,, is the present trended set point value,
FET is the previous trended set point value,
FF ET, is the present value of the measured variable, and
K is a filter constant.
The function of filter 161 may be conveniently performed in a digital computer with the signals FET,,,,,,,, FET,,,,, and FFET, and the constant FK being stored in computer memory. In addition, the set point signal may be overridden by the operator who can input a signal FET I An error is determined by summing amplifier 160 which generates an output signal representing the error DFET, in accordance with:
The function of summing amplifier 160 also may be conveniently performed in a digital computer.
Gas temperature controller 163 determines a desired exit gas flow rate in the kiln and provides both proportional and integral modes. The function of gas temperature controller 163, which may be implemented by a digital computer, is represented by the equation for output signal EXIT, of controller 163, as follows:
Where EXlT is the desired exit gas flow rate,
EXIT, is the previous exit gas flow rate,
DFET, is the present feed end temperature error signal,
DFET,, is the previous feed end temperature error signal,
K, and K are control constants.
Output signal EXIT, of gas temperature controller 163 is applied to line 71 for application to controller 74 through logic switch 164. In the illustrated embodiment, signal EXIT is employed to control the speed of fan 31 but may, as an alternative, serve to control the position of damper 38. Logic switch 164 normally connects temperature controller 163 to controller 74, as illustrated, but may also serve to interrupt the connection, as subsequently described. If the function of gas temperature controller 163 is performed in a digital computer, signals EXIT,,, EXIT,,,,, DFET and DFET and constants K and K would be stored in the computer memory.
Excessive values of error DFET indicate the kiln will be upset in a short time if corrective action is not taken. For instance, the discharge end of the kiln could be hot with fuel being taken off to cool it down. An excessive feed end temperature decrease indicates that too much fuel has been taken off. Such a condition can cause an excessive temperature drop when the colder feed gets to the front .of the kiln. When this happens, dead band-amplifier 162 forces the fuel set point to tend to override'the main fuel set point thereby attempting to correct the problem before it really develops. Such an action is especially helpful in decreasing the fuel level when the kiln speed is decreased to maintain the feed end temperature within limits. Therefore, dead band amplifier 162, which may be conveniently implemented by a digital computer, generates a signal FDFET in accordance with the equation:
Where fDFET, is the present function,
DFET is the present error,
K is an amplifying factor, and
K is a dead band constant which causes dead band amplifier 162 to generate a signal fDFET,,= for values of DFET less than K,,,,,,,,.
Therefore, fDFET, affects the fuel set point only during cxcessive deviations of the feed end temperature represented by the signal FFET,,. The fuel set point is affected because dead band amplifier 162 is coupled to an input of summing amplifier 131. Therefore, when this function of feed end temperature control is added, the fuel set point output from summing amplifier 131 represented by signal FUEL is generated in accordance with equation:
Where FUEL, is the fuel set point,
XTFERR, is the present fuel error signal compensated for speed changes, previous control actions and long-term changes, and
fDFET, is the feed end temperature control function. Hence, if the feed end temperature is cold by a first amount less than the threshold value of the dead band amplifier 162, EXIT, is increased to increase the gas flow through the kiln and the feed end temperature. However, if the feed end temperature becomes sufficiently low, the fuel rate will be increased. Similarly, if the feed end temperature becomes too high the draft will initially be lowered and then dead band amplifier 162 will cause the fuel set point FUEL to be decreased.
Not only does this control loop respond to fuel changes and speed changes caused by differences-in burning zone conditions, it also controls the system in response to the characteristics of raw materials entering the feed end or to changes in feed rate. For example, if the raw, materials require a greater quantity of heat, the gas temperature at the kiln feed end will decrease. The control loop will increase exit gas flow to transfer more heat to the feed end of the kiln thus maintaining the desired temperature profile. Further, if a sufficient feed change occurs, this control loop will cause the fuel set point to be altered to increase the heat input. If such action were not taken, the resulting decrease in gas temperature at the kiln feed end could eventually affect burning zone conditions and appear as a disturbance requiring more drastic corrective action by the torque control loop previously described. This feed end temperature control loop, therefore, tends to compensate for disturbances and for the effects of other control actions taken by the control system.
Amp Rate Forcing Amp rate forcing logic 165 detects kiln upsets and tends to take corrective action earlier than the rest of the control system would detect it. Therefore, it improves the basic controller system by allowing corrective'action to be initiated sooner to thereby minimize the magnitude of the upset.
A good measure of the magnitude of an upset is the rate at which the signal FAMP, changes. By previous definition,.the signal FAMP, represents the torque and is a more reliable process parameter during nonstable conditions. Therefore, amp rate forcing logic responds to DELAMP, representing the amp error signal and FAMP to generate a forcing function represented by signal FF,,. The signal FF, is calculated by first determining a rate of kiln amp change DE- RAMP, in accordance with the equation:
Therefore, the value of the signal DERAMP, varies as the rate of change of the signal FAMP,,. DERAMP, is compared with a function of DELAMP,, which varies so that amp rate forcing is required if DERAMP, is large enough to indicate that FAMP will go below a set point ina predetermined number of scans or control actions and if DERAMP, is greater than a constant. If such a condition exists, no immediate action is taken. Rather after a time delay the DELAMP, and FAMP are reexamined to determine if the condition still exists. If the condition persists, amp forcing occurs. In accordance with a preferred embodiment, two speed factors are determined. In one, a speed change is calculated which is proportional to the rate and is compared with a given proportion of the previous forcing signal. Therefore, two signals exist, one proportional to the present rate of change and the other proportional to the last forcing function. The greater is used to generate the signal FF,,.
With this amp rate forcing kiln speed can be decreased as much as required by rate on each scan. However, kiln speed will return only a given percent during each scan. Kiln speed changes which would otherwise result from transient rate changes are limited. For example, a large speed change may cause the signal FAMP, to stop decreasing for one scan and then to continue decreasing in successive scans. lf speed changes responded only to DERAMP,,, the kiln speed would return to its base value during the second scan and then decrease during the third scan. lmplementing the dual signal technique avoids this condition. The forcing function will also be responsive to other changes. In normal system operation where kiln amp forcing is utilized, regular speed logic will tend to reduce kiln speed as the rate logic tends to increase kiln speed with the net effect of holding speed and thereby providing a lead on the speed change. The forcing function signal FF, is applied to summing amplifier 141 so that, with the addition of amp rate forcing through the system, the signal DSERR, representing the composite error apportioned to speed changes is generated in accordance with:
DSERR,,=SERR,,+FF, with the value FF, being stored in a computer memory if the function of the summing amplifier 141 is implemented in a digital computer. Therefore, the kiln speed'KSPD, will be responsive to proper combinations of the rate of change of torque determinedby comparing the values of the torque signal represented by FAMP and the magnitude of'the total deviation represented by the signal DELAMP, to generate speed forcing.
High Temperature Speed Forcing And Override 1f the burning zone temperature represented by FTBZ becomes excessive, error ETBZ, goes negative. ETBZ, in addition to being applied to multiplier 122 is also applied to a temperature controller which is also coupled to function generator 122 to be responsive to the trended temperature value WS'I'I, to generate an output signal TEMP, in accordance with:
Where TEMP is the present value of the temperature control function,
ETBZ, is the present burning zone temperature error,
WS'IT, is the present weighted temperature function, and
K is a proportionality constant.
During normal operation ETBZ, represents normally control- I lable burning zone temperature errors with errors caused by high temperature being represented by negative values of TEMP,,. These negative values will not be sufficient to cause threshold circuit 171 to generate high temperature forcing function signal HTEMP, during normal operation. However, if excessively high temperatures exist, output signal HTEMP, is generated in accordance with:
Where HTEMP, is the present speed forcing function,
TEMP,, is a present modified temperature error, and
K,,,, is a proportionality constant. The output of threshold circuit is such that HTEM P, is always less than a minimum value HTEMP,,,,,,. When generated, the signal HTEMP, is then applied to summing amplifier 143 to modify the speed error signal DELSPD,, generated thereby so that the output of summing amplifier 143 is represented by DELSPD,,=DSERR,,+HTEMP, Therefore, the speed and, through function generator 145, the fuel are modified for temperature excursions above a particular value. That'is, kiln speed will decrease and will cause a decrease in the fuel. In this manner an overternperature condition may be quickly corrected without causing long-term upsets in the kiln. If a small speed increase were implemented, it could result in a fairly gentle action and could make the kiln respond in an orderly way to bring the temperature down. It could, however, cause a temperature rise with a final large burning zone temperature drop causing the kiln to go cold.
Such a situation could exist if the load in the kiln moved toward the discharge end 15. Therefore, thermal runaway of the kiln is prevented by significantly decreasing speed and fuel while leaving the gas flow rate through the kiln as high as possible. Once the high temperature condition is corrected, fuel can be added. At the lower temperature, the flame position stabilizes to maintain the proper heat distribution or temperature profile in the kiln. By affecting the speed, feed is decreased to help maintain thermal balance and allow greater fuel changes to break thermal runaway while tending to avoid upsets causing succeeding cycles. Therefore, the percent fuel decrease is greater than the percent speed decrease so there will still be a decrease in the total heat balance of the kiln.
A fuel override is also provided when the temperature becomes too high. This is generated by a switch 172 which couples the burning zone temperature error ETBZ, to function generator 173 until signal TEMP, reaches the threshold value of threshold circuit 171. Function generator 173 generates an output signal fETBZ, which is applied to summing amplifier 131. The signal fETBZ, is generated in accordance with:
2 fE n=Krm l fr Where ETBZ, is the present value of the override function,
ETBZ, is the present burning zone temperature error,
K is a proportionality constant,
K is a dead band function, and
K;, is a proportionality constant. Function generator 173 also is formed so that the signal fEBTZ, is zero until ETBZ, signifies that the burning zone temperature has reached a level indicated by K and the set point TBZ,,,,.
fETBZ, is the present value of the override function. ln this manner, the fuel set point is responsive to the effect of the excessive burning zone temperatures in addition to the basic burning zone temperature error. If high temperature forcing is being utilized, threshold circuit 171 opens switch 172, so the front end override and high temperature signals cannot exist simultaneously.
Low Temperature Speed Forcing If the temperature of the burning zone represented by the signal FIBZ, goes below the-burning zone temperature set point represented by the signal TBZ,,,,, the error signal ETBZ, goes positive. At a positivethreshold value, threshold circuit 174 generates a low temperature forcing signal LTEMP,, applied to summing amplifier 141. The value of the low temperature forcing function represented by the signal LTEMP, which must be greater than a minimum value, is calculated in accordance with the equation:
Where LTEMP, is the present value of the forcing function,
TEMP, is the present modified temperature error,
K is a proportionality constant, and
K is a dead band constant. When the low temperature forcing function is added, the speed error signal DSERR, is generated by summing amplifier 141 in accordance with: v
DSERR,,=SERR,,+FF,,+LTEMP,, so that the output signal is a function of that portion of the composite error attributable to speed, the forcing function and, where excessive low temperatures exist, the low temperature forcing function. The addition of the low temperature forcing function has the effect of slowing the kiln and, as it is added to summing amplifier 141 the coupling effect on fuel also exists. Furthermore, the low temperature conditions are fed back to process model 133 and become a part of the model as a composite speed error signal ESPD,,.
Oxygen Override Oxygen override logic 175 receives the fuel set point signal FUEL on line 70 and the speed output signal KSPD, on line 72 in addition to the output EXIT, of controller 163. Signal FOXY,,, the filtered value of the oxygen from the feed end generated by the oxygen sensor 55 on line 56 is inputed to the oxygen override logic. Finally, a feed ratio set point is applied to the oxygen override logic to generate, in conjunction with the speed, a feed rate. Basically, oxygen override logic 175 has the capability of calculating predicted oxygen content and employing this as a substitute for measured oxygen content. One embodiment of oxygen override logic has been described in US. Pat. (Ser. No. 678,851) issued (filed Oct. 30, 1967) by James W. Lane and assigned to the same assignee as the present invention. Basically, oxygen override logic 175 calculates the present exit gas rate and calculates a predicted oxygen content of the exit gas based on new conditions within the kiln. If the predicted oxygen content is less than this rate minimum, a new overriding exit gas rate is calculated. If the recalculated gas rate is greater than the capacity of fan 31, a new overriding fuel rate which will be less than the fuel set point signal FUEL is determined using the maximum capacity of the fan 31. The recalculated, overriding fuel rate set point results in a minimum safe exit gas oxygen content when the exit gas rate is at the maximum. A signal representing the overriding exit gas rate is calculated by override logic 175; or a signal representing the maximum exit gas rate, if required, is applied from oxygen override logic 175 to switch 164 and takes precedence over the exit gas rate EXIT, determined by gas temperature controller 163. Similarly, if the new fuel overriding set point is calculated by oxygen override logic 175, a signal representing this new fuel set point is applied to logic switch 132 and takes precedence over fuel rate set point F UEL,,,, determined by summing amplifier 131. The functions of oxygen override logic 175 may be conveniently performed in a digital computer, with the computer memory being employed to store the signals required'for computations.
Oxygen override logic 175 assures that the oxygen content is above a minimum safe level and that no combustibles or carbon monoxide appear in the exit gases. Oxygen override logic 175 monitors the oxygen content of the exit gas and determined what the new oxygen content will be after the contemplated control actions are taken. if the predicted oxygen content is less than the prescribed minimum level overriding action is taken. Priority is such that the exit gas rate calculated by gas temperature controller 163 is sacrificed first to prevent the desired fuel rate determined by summing amplifier 131, override logic 175 calculating a new exit gas rate set point which will result in the predicted oxygen content of the minimum safe level of the desired fuel rate. However, if the exit gas cannot adjust sufficiently to provide the required minimum oxygen logic, the fuel rate is also adjusted by override logic 175 to produce the safe minimum oxygen content of the maximum exit gas rate. Oxygen override logic 175 thus prevents dangerous conditions from occurring by preventing the selection of fuel and exit gas flow rates which reduce the oxygen content of the exit gas below minimum safe level.
Wet Kiln Controller While the above described embodiment with its various commutations is directly adapted for control in a dry kiln, it has been found that in wet kilns where a slurry mixture is added to the kiln at the feed end 14, a significant effect on 1 becomes sticky and rides up on the walls giving a large angle of repose. As the feed dry, the weight of water is lost and the powder that remains flows down the kiln with a small angle of repose requiring very little torque. The total torque required to turn the chain section has been very much a function of the amount of wet heavy feed that must be moved. The faster the feed is dried, the less wet material there will be to turn and the less torque will be required. If the material is dried sooner, the temperature of the dried feed leaving the chain section should be greater and a longer time will be available to heat the material. Therefore, the temperature of the feed at the exit from the chains is a function of the torque required to turn the chain section. As the temperature goes up, the required torque goes down. Therefore, as heat input to the chain section 16 is increased, the torque requirements for the chain section will decrease. if this change is.interpreted as being due to burning zone change, it will look like fuel should be added. However, in such a condition fuel should be decreased. Thus, the effective changes at feed end of the kiln in the chain section 16 when interpreted as coming from the burning zone will cause wrong control actions to be taken. lt has been deter-- mined that in some kilns the changes in torque resulting from changes in the chain section can be of the same order ofmagnitude as changes caused by changes in the burning zone region. This effect can be properly handled by generating a signal CHAMP, which represents the effect on torque of changing conditions in the chain section.
The filtered signals representing the dry feed rate FFEED,,, the intermediate gas temperature FTlG, the dust ratio FDR,,, the front end temperature FFET,,, the kiln speed as indicated by the set point KSPD, and, from the operator, a signal PMIF representing the percent moisture in the feed and the signal TMPF representing the input temperature of the feed are applied to temperature calculator 180. initially a feed temperature function at the exit from the chain section is calculated by implementing a heat balance around the chain section using the dust, feed and the intermediate gas temperature as indicating heat inputs. Heat is carried from the chain section by the dry gas exiting the feed end, moisture driven from the slurry as steam, dust leaving the chain section and the feed leaving the chains. Therefore the temperature function of the feed leaving the chains represented by signal TSC, can be approximated by:
TSC,,= [K,,,,,+( FFEED, C FI'MP)/( lPMlF)+ PMIF FFEED,,(C,,,.,,,,+1000 FFETUK l-PMIF)] /(FFEED,, C
Where FIMP and PMIF are operator inputs representing the feed temperature and percent moisture in the feed respectively, I
C C,,.,,, C C and C are the specific heat for the feed, incoming gas to the chain section, gas exiting the chains and dust and heat of vaporization respectively,
K is a radiation constant, and
K and K, are constants related to determining the actual dust in the kiln. The instantaneous value TSC, is filtered in temperature calculator to obtain a value TSCF, in accordance with:
Where TSCF is the present filtered value,
TSCF is the previous filtered value,
TSC, is the new calculated value, and
K is a filter constant.
This filtered instantaneous number is trended to obtain an output signal TSC'l", from filter 181 in accordance with the equation:
Where i TSCT, is the present trended filtered value,
TSCT is the previous trended filtered value,
TS(F,, is the present filtered value, and
K is a constant.
An error function DSCT,, is then determined by summing amplifier 182 energized by the trended temperature signal TSCT, and the instantaneous filtered signal TSCF, in accordance with the equation:
Where DSCT, is the present difference value,
TSCT, is the present'trended value, and
TSCF is the present instantaneous value.
This difference signal is then applied to filter 183 to generate the output signal CHAMP, coupled to process model-133 in accordance with the equation:
CHAMP,,=CHAMP,,,,+K,,,,,,,,,, (DSCT,, CHAMP Where CHAMP, is the present calculated change in torque due to changes in the chain section,
CHAMP is the previous change in torque,
DSCT, is the present difference in the temperature, and
K is a proportionately constant.
Therefore, a change in the chain section which immediately appears in the signal FAMP will, without the addition of the chain amp section, appear as a burning zone temperature disturbance. However, the feedback signal FB from process model 133 is, with the addition of the chain amp, calculated with the equation:
F B,,==EFUEL,,.,,+ESPD,,,,,+CHAMP, so it immediately appears as an offsetting factor in the composite error signal ERR, from summing amplifier 121 and the effect of changes are thereby minimized.
FIGS. 4, 4A and 4B (hereinafter FIG. 4) illustrate a flow chart of the operation of the control system of FIG. 2'. Referring to FIG. 4, initially several parameters to include dry feed rate, intermediate gas temperature, dust rate, feed end temperature, discharge oxygen rate, fuel rate, motor torque and burning zone temperature are measured, checked and filtered. Each of these signals are scanned on a continuous basis and made available to the filters to obtain periodic filtered values FFED,,FTIG, FDR FFET,,, FOXY,,, FUEL,,, FAMP and- FTBZ The check logic associated with each input compares successive'inputs and if two values differ by more than a predetermined value, the previous value is saved and used in lieu of the present value.
Error signal ETBZ at the end of the control period is determined from the filtered value of burning zone temperature FIBZ and the set point TBZ,,. A stability gain factor SGF is produced from the burning zone temperature FIBZ, to indicate the stability of the process. The signal representing motor torque FAMP, is trended to obtain a set point signal AMP,,,,,,, which is also a function of the stability gain factor signal SGF,,. Thetrended amp set point and the filtered value FAMP, are then combined to generate an error DELAMP,,. Also the stability gain factor is combined with the burning zone temperature error ETBZ to obtain a burning zone temperature deviation DELTBZ,,.
1n the preferred form of the wet kiln control system, it is necessary to next calculate the temperature of the feed leaving the chain system for the conditions which have existed over the last control period, the signal being TSCF,, and then trending this to obtain a signal TSCT,,. With these two values it is possible to calculate that portion of a change in kiln amps which is due to changes inthe chain section conditions represented by signal CHAMP,,. This signal CHAMP, is a function of TSCF, and TSCT,,. Once the various errors have been obtained, it is possible to update the process model. Predicted process changes due to previous control actions in speed and fuel are obtained and added to changes related to variations in the chain section conditions to generate a total feedback signal PB The feedback signal FB, is filtered to determine a new steering function STF and a new base fuel signal FUEL Both these signals are fed back in combination with other signals, to the process model as fuel error input EFUEL,,.
In addition, the burning zone temperature deviation DELTBZ,,, the kiln amp error DELAMP,,, and the feedback signal P8,, are combined with the steering function STF, being subtracted therefrom to obtain a total composite error signal representing short term disturbances in the burning zone ERR The short term disturbance represented by ERR, is compared with a threshold value FERR representing a maximum error for fuel control to limit the portion of ERR, is also compared with a speed dead band to implement a kiln speed variation if it exceeds the dead band, the portion of ERR, which is apportioned to speed being designated as SERR,,.
If amp rate forcing is utilized, the control system determines whether amp rate forcing existed during the previous control action. If it did, a first speed decrease as a function of the rate of change of kiln amps and a second speed decrease as a function of the percent of the last speed change are calculated. The larger speed decrease is compared with a minimum acceptable change. lf the predicted change is less than the minimum, amp rate forcing is terminated. [f it is not, the larger Speed decrease is implemented. lf amp rate forcing was not implemented during the prior control action, the rate of change of amps is examined. If the rate of change will cause the signal FAMP representing motor torque to fall below a set point within a predetermined number of scans and if the rate of change exceeds a constant value, additional tests are made. If these two conditions do not exist, amp rate forcing is not required. If the conditions are met and if the control system indicates that the present scan is the second consecutive scan to satisfy the conditions, amp rate forcing is initiated and a speed change proportional to the amp rate is calculated and implemented.
lf high or low temperature-speed forcing or high temperature fuel override are utilized, a temperature function is determined from the burning zone temperature error ETBZ,,. High or low temperature forcing are implemented only if the burning zone temperature error leaves a dead band region. Considering high temperature forcing first and assuming that high temperature forcing was implemented, the temperature function TEMP, and the weighted, trended temperature WSTT are analyzed to determine whether the forcing can be terminated. lf termination is not feasible, a high temperature speed forcing function HTEMP, is generated. Whenever this speed function is implemented, the fuel is decreased in accordance with fuel speed coupling function fSPD,,. lf forcing were not implemented on the previous scan, TEMP and WSTT, would be examined to determine if forcing were necessary. lf it were HTEMP would be generated.
Similar action is taken for low temperature forcing. If forcing was previously practiced, the conditions of TEMP and WSTY are examined to determine the feasibility of termination. if the forcing function is not terminated, a speed decrease and appropriate fuel decrease are determined. if forcing was not previously practiced, the conditions of TEM P, and WSTT,, are examined and low temperature forcing is implemented if necessary.
After these fuel and speed changes are determined based upon the various process measurements, all speed changes are combined to obtain a speed deviation signal DELSPD,,. Further, all speed changes except the speed change determined by the high temperature forcing function HTEMP are combined and filtered to obtain a feedback signal ESPD, which is applied to the process model. Alternatively, speed and fuel forcing functions might be added into the control signals but not into the feedback signal. This feedback signal is generated on the basis of past history with each factor being delayed in accordance with process characteristics. The resulting total speed change DELSPD, is added to base speed KSPD to obtain the kiln speed set point signal KSPD,,.
Fuel feedback is based upon a base fuel level FUEL and total process error apportioned to fuel TFERR, to obtain a signal representative of the change in fuel DFUEL to which is added the steering function STF, to achieve a steered fuel change representation STFUEL,,. When this function is filtered, the output signal EFUEL, is an input to the process model. The signal TFERR,, is then compensated by fuel-speed coupling fSPD to achieve a factored fuel change XTFERR,,.
if the burning zone temperature is above the set point but below a value which would cause high temperature forcing an override function fEBTZ is added to the factored fuel level XTFERR,,. A feed end temperature set point PET is obtained as a function of the filtered feed end temperature FFET, and an error DFET, is calculated. lf DFET, lies outside a dead band, a fuel change fDFET, is combined with the signal fEBTZ,, and XTFERR, to obtain a fuel set point FUEL which is utilized to set the fuel supply controller.
Feed end temperature error DFET,, is also utilized to determine a new exit gas rate. The oxygen leaving the kiln represented by FOXY,,, present exit gas rate EXlT,,, the fuel rate set point FUEL,,,, and kiln speed KSPD, and a feed ratio set point are also measured for determining the new exit gas rate. The predicted exit gas rate is then compared with flow capacities and predicted oxygen flow from the kiln. The new exit gas rate is then determined to assure a safe oxygen level. If necessary, previously determined exit gas rates and fuel set points are modified.
After these functions are completed, the new kiln speed KSPD,,, fuel level FUEL and exit gas rate EXIT are coupled from the control system to their respective controllers to effect kiln control. After a predetermined time increment, the control system shown in H6. 4 determines a new set of conditions.
Accordingly, there has been described herein a method and apparatus for kiln control embodying the instant invention. All the principles of the invention have now been made clear in the illustrated embodiment, and there will be immediately obvious to those skilled in the art many modifications in structure, steps, arrangement, proportions, elements, materials and components used in the practice of the invention.
The appended claims are, therefore, intended to cover and embrace any such modifications, within the limits only of the true spirit and scope of the invention.