Preemptive constraint control
OwnerAlbert J. Waltzinventor
Life of this patent
Estimating the term…
Abstract
A means of controlling a process with a multiplicity of measurements dependent on one process parameter, so that only the most constraining of the measurements manipulates the process in such a way that the performance of each individual controller selected is equivalent to the performance if there were a single controller controlling only one measurement, the most constraining variable. The means of controlling also manipulates the process to prevent a measurement, that is within a process constraint limit, from excessively exceeding that process constraint limit when it is projected that it would do so without the corrective manipulation.
Claims36
A method for controlling a process operating under supervision of a plurality of feedback control loops, said method comprising:
- (a) providing at least one constraint controller including a combination of operatively connected and interrelated override controllers, each of said override controllers including at least one individual controller, said override controllers combined in a preemptive constraint control system which system provides responsive regulation of interrelated process conditions attendant to said process being controlled;
- (b) selecting a group of variables to be monitored;
- (c) establishing operating parameters for said process being controlled based upon said group of variables, including
- (1) establishing a setpoint for each variable within said group of variables, said setpoint and its corresponding variable being controlled by a selected one of said individual controllers,
- (2) establishing a limit value for each of said variables within said group of variables,
- (3) establishing limit violation direction for each operating parameter, and
- (4) choosing a common process parameter to receive output from said individual controllers;
- (d) providing a plurality of monitoring means, said monitoring means being selectively connected with said override controllers and positioned to monitor said group of variables;
- (e) monitoring by selected ones of said monitoring means, each of said variables within said group of variables, and measuring at an instant in time each of said variables within said group, thus establishing a measured value for each of said variables;
- (f) storing each of said measured values;
- (g) then computing a pseudo-value of each of said variables within said group of variables, each of said pseudo-values corresponding to a respective one of said measured values which said pseudo-value and said measured value form a corresponding pair of values;
- (h) storing each of said pseudo-values;
- (i) then selecting one of said measured values and comparing said selected measured value with its corresponding pseudo-value, and repeating such comparison for each of said corresponding pairs of values;
- (j) establishing an array of adjustment values by utilizing sets of control values, each set consisting of one of said setpoints and said corresponding pair of values and computing for each of said control value sets a corresponding adjustment value, said adjustment value being used to bring said measured value of said control value set to a corresponding one of said setpoints;
- (k) selecting one of a plurality of said individual controllers by choosing a final adjustment value for said plurality of feedback control loops by selecting a maximum value from said adjustment values computed in step (j) if a positive adjustment value moves said process away from one of said limit values, otherwise, by selecting a minimum value from said adjustment values computed in step (j) if a negative adjustment value moves said process away from a selected one of said limit values, thereby leaving a number of non-selected individual controllers;
- (l) changing said common process parameter by said final adjustment value;
- and
- (m) repeating steps (e) through (l) in a continuous manner.
A method for controlling a process by interconnecting a plurality of constraint controllers in a cascade control structure, each constraint controller including a plurality of individual controllers, each individual controller controlling a selected variable chosen from the group consisting of a monitored process variable and a variable computed from a set including monitored process variables and constants, such that one constraint controller selects a maximum positive adjustment to a common process parameter of said process, while a different constraint controller selects a minimum negative adjustment to said common process parameter of said process, said method comprising the steps of:
- passing said adjustments developed by a first constraint controller of said constraint controllers to a second constraint controller of said constraint controllers with said adjustment from said first constraint controller treated as an adjustment developed by one of said individual controllers in said second constraint controller;
- communication between said constraint controllers so that no one of said adjustments is made to said common process parameter whenever opposing adjustments, excluding signals cascaded from other controllers, are selected in two different constraint controllers adjusting said common process parameter;
- and
- producing an alarm when no adjustment is made due to said opposing adjustments.
A method of modifying constraint control adjustments to a common process parameter of a multiplicity of individual controllers, which individual controllers constitute a constraint controller, to compensate for an expected effect on measured variables controlled by said individual controllers, of recently implemented adjustments to said common process parameter, said method comprising the steps of:
- (a) establishing an array of said recently implemented adjustments to said common process parameters;
- (b) establishing an array of sets of factors, each of said factors in each of said sets corresponding to one of said recently implemented adjustments and representing an unrealized expected change in one of said measured variables controlled by said individual controller, at current time, resulting from said implemented adjustment, and each of said sets of factors corresponding to one of said individual controllers;
- (c) utilizing each of said sets of factors to compute an adjustment modification representing a total adjustment required to product said unrealized expected change, and computing said adjustment modification for each of said individual controllers;
- (d) modifying each of said constraint control adjustments by a corresponding one of said adjustment modifications;
- and
- (e) selecting a final adjustment value for said common process parameter from said modified constraint control adjustments.
A method of controlling a process by computing adjustments to a common process parameter of a multiplicity of individual controllers, which individual controllers constitute a constraint controller, each of said individual controllers responding to a rate of approach of a measured variable to setpoint for each of said individual controllers, said method comprising the steps of:
- (a) establishing an array of adjustment values for said constraint controller, each value in said array resulting from a control response from one of said individual controllers constituting said constraint controller;
- (b) establishing a limit violation direction for said measured variable deviations for each of said individual controllers;
- (c) establishing a measurement standard variation range by adding to the mean of said measured variable measured during a recent past period, a positive factor multiplied by the standard deviation of said measured variable and a negative factor multiplied by said standard deviation, said standard deviation having been computed for said recent past period, said measurement standard variation range being established for each of said multiplicity of individual controllers;
- (d) determining a selected portion of each one of said array of said adjustment values, said selected portion resulting from said rate of approach of said measured variables;
- (e) selecting from said array of adjustment values a maximum value when a positive adjustment value moves the process in a direction opposite to said limit violation direction and a minimum value from said array when a negative adjustment moves said process opposite to said limit violation direction;
- (f) adding to said selected adjustment value, each of said rate selected portions corresponding to each of said individual controllers constituting said constraint controller, when said measured variable corresponding to said individual controller is beyond said measurement standard variation range in said limit violation direction;
- (g) summing for each individual controller said rate selected portions for each respective one of said individual controllers;
- (h) subtracting from said selected adjustment value said sum of said rate selected portions when said rate of approach corresponding to said sum of said rate selected portions changes sign when said measured value is within said measurement standard variation range;
- and
- (i) zeroing said sum of said selected portions.
A method of arranging and interconnecting constraint controllers to optimize excess air addition to a combustion process, said constraint controllers consisting of a multiplicity of individual controllers, each of said individual controllers controlling a measured variable of said combustion process which reflects the condition of said combustion process, said method comprising the steps of:
- (a) measuring selected combustion process conditions;
- (b) establishing a setpoint for each of said controlled measured variables;
- (c) measuring the carbon monoxide concentration of a stream of combustion products
- (d) measuring oxygen concentration of said stream of combustion products;
- (e) computing the natural logarithm of said carbon monoxide concentration;
- (f) utilizing said natural logarithm of said carbon monoxide concentration as said measured variable signal for a selected one of a said individual controllers in a said constraint controller;
- (g) utilizing said oxygen concentration as said measured variable for a maximum oxygen limit controller and a minimum oxygen limit controller, said maximum and said minimum oxygen limit controllers being a part of said constraint controller system;
- (h) adjusting setpoints of said oxygen limit controllers based upon preprogrammed acceptable values correlated to furnace load as indicated by a measurement inferring amount of combustion such as steam flow from a steam boiler;
- (i) controlling the ratio of air flow to total firing load;
- (j) limiting air flow control through said constraint controllers utilizing said individual controllers for (1) minimum air flow as determined by minimum air to fuel flow ratio, (2) minimum fire box pressure, and (3) minimum air flow valve position;
- and
- (k) limiting fuel flow control through said constraint controller utilizing said individual controllers for (1) maximum fuel flow as determined by maximum fuel to air flow ratio, (2) maximum tube temperature, (3) maximum fuel flow valve position.
A method for eliminating unnecessary individual controller manipulation of process parameters, said unnecessary manipulation being caused by control response to small amplitude variation of a controlled measured variable wherein said unnecessary manipulation prevents the process from reaching a quiescent steady state, said method comprising the steps of:
- (a) inhibiting said unnecessary manipulation when said measured variable is within a factor of a statistical standard deviation of the mean of a first-in/first-out set of recent measured variables;
- (b) inhibiting particular control responses;
- and
- (c) developing derivative control response based upon change in said measured variable.
A method of controlling a calculated process variable, said method comprising the steps of:
- (a) monitoring a process to establish process variables;
- (b) performing calculations by utilizing constants, said process variables and setpoints of said process variables;
- (c) establishing an integral only controller through a digital control algorithm with an adjustable integral time and scaling gain;
- (d) using the result of said calculations as the variable controlled by said integral only controller;
- (e) using a setpoint adjusted by said integral only controller as said setpoint used in said calculations;
- (f) setting said scaling gain equal to the inverse of said calculations;
- and
- (g) setting said adjustable integral time equal to the duration between executions of the digital control algorithm of said integral only controller.
A method for controlling a process utilizing a plurality of individual controllers and a plurality of feedback control loops to provide an output to manipulate a common process parameter and to regulate a plurality of variables, comprising the steps of:
- (a) providing an output and an output adjustment to said common process parameter;
- (b) for each individual controller,
- (1) measuring the value of one variable of said plurality of variables, and determining a measured value characteristic,
- (2) determining a setpoint and a setpoint characteristic for said variable,
- (3) determining a limit violation direction for said variable,
- (4) deriving a deviation characteristic by determining the difference between said setpoint characteristic and said measured value characteristic,
- (5) developing a control characteristic utilizing said deviation characteristic,
- (6) modifying said control characteristic, and
- (7) determining an individual adjustment utilizing said modified control characteristic, said individual adjustment being that adjustment to said common process parameter tending to reduce the deviation of a corresponding one of said plurality of variables from said corresponding setpoint;
- (c) selecting an individual adjustment having the maximum value if an increased adjustment to said common process parameter would change a corresponding one of said variables in the opposite direction of said limit violation direction of said variable, otherwise, selecting an individual adjustment having the minimum value if a decreased adjustment to said common process parameter would change said corresponding variable in the opposite direction of said limit violation direction of said variable;
- and
- (d) selecting the individual controller corresponding to said selected individual adjustment, the remaining individual controllers being non-selected, the individual adjustment of said common process parameter.
A method for controlling a combustion process utilizing a plurality of individual controllers and a plurality of feedback control loops to provide an output to manipulate a combustion process parameter affecting the air to fuel ratio and to regulate a plurality of variables, comprising the steps of:
- (a) providing an output and an output adjustment to said combustion process parameter;
- (b) developing measurement characteristics for said plurality of variables;
- (c) measuring the oxygen concentration of the combustion gas products to develop a said measurement characteristic,
- (d) measuring the carbon monoxide concentration of said combustion gas products to develop a said measurement characteristic,
- (e) computing the logarithm of said carbon monoxide concentration,
- (f) for each individual controller,
- (1) establishing one of said measurement characteristics as an individual measurement characteristic,
- (2) determining a setpoint and a setpoint characteristic for said individual measurement characteristic,
- (3) determining a limit violation direction for said individual measurement characteristic,
- (4) deriving a deviation characteristic by determining the difference between said setpoint characteristic and said individual measurement characteristic,
- (5) developing a control characteristic utilizing said deviation characteristic;
- (6) modifying said control characteristic;
- (7) determining an individual adjustment utilizing said modified control characteristic, said individual adjustment being that adjustment to said combustion process parameter tending to reduce the deviation of a corresponding one of said individual measurement characteristics from said corresponding setpoint;
- (g) establishing said logarithm as said individual measurement characteristic and establishing increasing carbon monoxide concentration as said limit violation direction for a said individual controller forming a carbon monoxide controller;
- (h) establishing said measurement characteristic developed from oxygen concentration as said individual measurement characteristic and establishing decreasing oxygen concentration as said limit violation direction for a said individual controller forming a minimum oxygen controller;
- (i) establishing said measurement characteristic developed from oxygen concentration as said individual measurement characteristic and establishing increasing oxygen concentration as said limit violation direction for a said individual controller forming a maximum oxygen controller;
- (j) selecting a maximum of said individual adjustments developed by said carbon monoxide controller and said minimum oxygen controller as a maximum select adjustment;
- (k) selecting a minimum of said maximum oxygen controller and said maximum select adjustment as said adjustment to said combustion process parameter, the remaining individual controllers being non-selected.
A method for controlling a process utilizing a feedback controller to provide an output to manipulate a process parameter to regulate a variable, comprising the steps of:
- (a) measuring the value of said variable, and determining a measured value characteristic;
- (b) determining a setpoint and a setpoint characteristic for said variable;
- (c) deriving a deviation characteristic by determining the difference between said setpoint characteristic and said measured value characteristic;
- (d) developing a control characteristic utilizing said deviation characteristic;
- (e) establishing an array of most recent said deviation characteristics;
- (f) computing the standard deviation of said array;
- (g) calculating a standard variation range by multiplying said standard deviation by a factor reflecting the range of said measurement characteristic variation during steady process operation;
- (h) modifying said control characteristic based on comparison of said measured value characteristic relative to said standard variation range;
- (i) utilizing said modified control characteristic to develop an adjustment to said process parameter, reducing said adjustment when said measured value characteristic is within said standard variation range;
- (j) manipulating said process parameter by said reduced adjustment.
Apparatus for controlling a process by controlling a common process parameter, comprising:
- (a) a plurality of individual controllers for providing an output to manipulate a common process parameter and to regulate a plurality of variables and an output adjustment for controlling the common process parameter;
- (b) means for measuring the value of one variable of said plurality of variables and determining a measured value characteristic;
- (c) means for determining a setpoint and a setpoint characteristic for said variable;
- (d) means for determining a limit violation direction for said variable;
- (e) means for deriving a deviation characteristic by subtracting said setpoint characteristic from said measured value characteristic;
- (f) means for developing a pseudo-value characteristic;
- (g) means for modifying said deviation characteristic by subtracting said pseudo-value characteristic from said deviation characteristic;
- (h) means for determining an individual adjustment utilizing said modified deviation characteristic, said individual adjustment being that adjustment to said common process parameter tending to reduce the deviation of a corresponding one of said plurality of variables from said corresponding setpoint;
- (i) means for selecting an individual adjustment having the maximum value if an increased adjustment to said common process parameter would change a corresponding one of said variables in the opposite direction of said limit violation direction of said variable, otherwise, selecting an individual adjustment having the minimum value if a decreased adjustment to said common process parameter would change said corresponding variable in the opposite direction of said limit violation direction of said variable;
- and
- (j) means for selecting the individual controller corresponding to said selected individual adjustment, the remaining individual controllers being non-selected, the individual adjustment of said selected individual controller becoming the adjustment to said common process parameter.
An apparatus for controlling a combustion process by manipulating a combustion process parameter, comprising:
- (a) a plurality of individual controllers for providing an output to manipulate a combustion process parameter affecting the air to fuel ratio and to regulate a plurality of variables and an output adjustment for manipulating the combustion process parameter;
- (b) means for measuring the values of said plurality of variables, the variables including oxygen and carbon monoxide concentration, and determining measured value characteristics;
- (c) means for computing the logarithm of said carbon monoxide concentration to develop a said measurement characteristic;
- (d) means for individual controller for establishing one of said measurement characteristic as an individual measurement characteristic;
- (e) means for individual controller for determining a setpoint and a setpoint characteristic for said individual measurement characteristic;
- (f) means for each individual controller for determining a limit violation direction for said individual measurement characteristic;
- (g) means for each individual controller for deriving a deviation characteristic by determining the difference between said setpoint characteristic and said individual measurement characteristic;
- (h) means for each individual controller for developing a control characteristic utilizing said deviation characteristic;
- (i) means for each individual controller for modifying said control characteristic;
- (j) means for each individual controller for determining an individual adjustment utilizing said modified control characteristic, said individual adjustment being that adjustment to said combustion process parameter tending to reduce the deviation of a corresponding one of said individual measurement characteristics from said corresponding setpoint;
- (k) means for establishing said measurement characteristic developed from said logarithm as said individual measurement characteristic and establishing increasing said carbon monoxide concentration as said limit violation direction for a said individual controller forming a carbon monoxide controller;
- (l) means for establishing said measurement characteristic developed from said oxygen concentration as said individual measurement characteristic and establishing decreasing said oxygen concentration as said limit violation direction for a said individual controller forming a minimum oxygen controller;
- (m) means for establishing said measurement characteristic developed from said oxygen concentration as said individual measurement characteristic and establishing increasing said oxygen concentration as said limit violation direction for a said individual controller forming a maximum oxygen controller;
- (n) means for selecting a maximum of said individual adjustments developed by said carbon monoxide controller and said minimum oxygen controller as a maximum select adjustment;
- (o) means for selecting a minimum of said individual adjustment developed by said maximum oxygen controller and said maximum select adjustment as said adjustment to said combustion process parameter, said individual controller developing said individual adjustment corresponding to said adjustment to said combustion process parameter is the selected controller, the remaining individual controllers being non-selected;
- (p) means for manipulating said combustion process parameter by said selected minimum.
Dependencies are inferred from the claim text.