Adaptive controller having optimal filtering

Claim 2

In a closed loop process control system for a process, the method of controlling process apparatus to maintain a controlled process variable (Y) at a predetermined value, said method comprising the steps of providing information in a stored mathematical model representing a functional relationship between said controlled variable (Y) and a manipulated process variable (X), deriving a predicted value (Y(k/k-1)) for the controlled variable from said mathematical model, obtaining measurements of the value of the controlled variable at periodic sampling intervals, deriving an estimated value (Y(k/k)) for said controlled variable from the predicted value and a measured value (k) of said controlled variable, calculating an adjusted value (X(k)) for said manipulated variable using the estimated value of said controlled variable, and adjusting the value of said manipulated variable to correspond to said adjusted value and thereby effect a corresponding change in the value of said controlled variable.

Claim 3

The method as set forth in claim 1 including the further step of adjusting the mathematical model at the time of deriving the estimated value (Y(k/k)) using the predicted value (Y(k/k-1)) last derived from the mathematical model and the estimated value (Y(k/k)) last derived to compensate for changes in operating conditions of the process.

Claim 4

The method as set forth in claim 2 wherein adjusting the mathematical model includes the steps of comparing the last predicted value (Y(k/k-1)) to a previous estimated value (Y(k-1/k-1)) which was derived from a preceding measurement (Z(k-1)) and adjusting the mathematical model whenever the last predicted value differs from the previous estimated value by more than a predetermined amount.

Claim 5

The method as set forth in claim 3, wherein the last predicted value (Y(k/k-1)) of the controlled variable is adjusted towards the last derived estimated value (Y(k/k)) of the controlled variable whenever the mathematical model is adjusted.

Claim 6

The method as set forth in claim 1, wherein the estimated value (Y(k/k)) for the controlled variable (Y) is defined by the relationship: Y (k/k) Y (k/k-1) + W (Z (k) - Y (k/k-1)) where Y (k/k-1) is the predicted value for the controlled variable Z(k) is the measured value for the controlled variable, and W is a noise weighting factor.

Claim 7

The method as set forth in claim 1, wherein calculating the adjusted value (X(k)) for said manipulated variable (X) includes the step of comparing the estimated value (Y(k/k)) to a desired set point value (R(k)) for the controlled variable (Y) to obtain a set point error value (E(k)) representing the difference between said estimated value and said set point value.

Claim 8

The method as set forth in claim 6, wherein calculating the adjusted value (X(k)) for said manipulated variable (X) includes the step of providing information in a stored process controller representing the functional relationship X (k) ((1 - theta )/K(1- eta )) (E(k) - eta E (k-1)) + theta X (k-1) + (1- theta ) X (k- Gamma -1) wherein E(k) is the value for the set point error calculated at the time of deriving the estimated value Y(k/k); E(k-1) is the value of the set point error calculated at the time of deriving a previous estimated value Y(k-1/k-1) which was derived from a preceding measurement; X(k-1) is the adjusted value calculated from the previous estimated value Y(k-1/k-1); X(k- Gamma -1) is the value of X(k) calculated ( Gamma + 1) measurements ago; K is the gain of the process controller, and eta and theta are exponential functions of the open and closed loop time constants respectively, of the process control system.

Claim 9

The method as set forth in claim 1, wherein said mathematical model stores information representing the functional relationship: Y(k/k-1) + Y(k-1/k-1) + K(1- eta ) Delta X(k- Gamma -1) + eta Y(k-1/k-1) + Y(k-2/k-2) eta wherein Y(k/k-1) is the predicted value for the controlled variable last derived from the mathematical model; Y(k-1/k-1) is a previous estimated value for the controlled variable derived at the time the last predicted value was derived; Y(k-2/k-2) is a further previous estimated value for the controlled variable derived prior to the time of deriving the last predicted value; Delta X (k- Gamma -1) is the difference between successive values of the manipulated variable calculated prior to the time of deriving the last predicted value; Gamma is a dead time factor; K is the gain of the mathematical process model, and eta is an exponential function of the open loop time constant of the process control system.

Claim 10

In a process control system for a process, measuring means for measuring a controlled process variable and providing signals representing a measured value of the controlled variable, means for storing data representing a predicted value for said controlled variable, process apparatus for controlling the value of said controlled variable as a function of the set point of said process apparatus, and process controller means responsive to said signals for controlling said process apparatus to compensate for changes in the value of said controlled variable caused by process upsets,said process controller means comprising signal processing means programmed to provide further signals from said stored data representing a predicted value and to calculate an adjusted set point value for said process apparatus from a weighted average of the predicted and measured values of said controlled variable represented by said signals.

Claim 11

In a process control system for a process, measuring means for measuring a controlled process variable and providing signals representing a measured value of the controlled variable, means for storing data representing a predicted value for said controlled variable, process apparatus for controlling the value of said controlled variable as a function of the set point of said process apparatus, and process controller means responsive to said signals for controlling said process apparatus to compensate for changes in the value of said controlled variable caused by process upsets,said process controller means comprising signal processing means programmed to provide further signals from said stored data representing a predicted value and to calculate an adjusted set point value for said process apparatus from a weighted average of the predicted and measured values of said controlled variable represented by said signals.

Claim 12

In a closed loop process control system for a process, measuring means for measuring the value of a controlled process variable (Y) and providing first output signals representing the value (Z(k)) of said controlled variable, system control means responsive to said first output signals for maintaining said controlled variable (Y) at a predetermined value (R(k)) by adjusting the value of a manipulated process variable (X) to effect a corresponding adjustment in the value of said controlled variable deviates from said predetermined value, said system control means comprising mathematical model means storing information for establishing a reference value for said controlled process variable and providing second output signals representing a predicted value (Y(k/k-1)) for said controlled variable, filter means responsive to said first and second output signals to provide third output signals representing an estimated value (Y(k/k)) for said controlled process variable and controller means responsive to said third output signals to calculate an adjusted value (X(k)) for said manipulated variable.

Claim 13

In a closed loop process control system for a process, measuring means for measuring the value of a controlled process variable (Y) and providing first output signals representing the value (Z(k)) of said controlled variable, system control means responsive to said first output signals for maintaining said controlled variable (Y) at a predetermined value (R(k)) by adjusting the value of a manipulated process variable (X) to effect a corresponding adjustment in the value of said controlled variable deviates from said predetermined value, said system control means comprising mathematical model means storing information for establishing a reference value for said controlled process variable and providing second output signals representing a predicted value (Y(k/k-1)) for said controlled variable, filter means responsive to said first and second output signals to provide third output signals representing an estimated value (Y(k/k)) for said controlled process variable and controller means responsive to said third output signals to calculate an adjusted value (X(k)) for said manipulated variable.

Claim 14

In a paper manufacturing process, process apparatus for manufacturing a continuous paper sheet having a predetermined weight, measuring means for measuring the weight of said paper and providing first output signals representing the weight (Z(k)) of said paper sheet, system control means including mathematical process model means storing information for establishing a reference value for the weight of said paper sheet and for providing second output signals representing a predicted value (Y(k/k-1)) for the weight of said paper sheet, filter means responsive to said first and second output signals to provide third output signals representing an estimated value (Y(k/k)) for the weight of said paper sheet, and controller means responsive to said third output signals for providing control signals to said process apparatus for effecting adjustment of the weight value of said paper sheet whenever the measured value of the weight of said paper sheet deviates from said desired value.

Claim 15

In a paper manufacturing process, process apparatus for manufacturing a continuous paper sheet having a predetermined weight, measuring means for measuring the weight of said paper and providing first output signals representing the weight (Z(k)) of said paper sheet, system control means including mathematical process model means storing information for establishing a reference value for the weight of said paper sheet and for providing second output signals representing a predicted value (Y(k/k-1)) for the weight of said paper sheet, filter means responsive to said first and second output signals to provide third output signals representing an estimated value (Y(k/k)) for the weight of said paper sheet, and controller means responsive to said third output signals for providing control signals to said process apparatus for effecting adjustment of the weight value of said paper sheet whenever the measured value of the weight of said paper sheet deviates from said desired value.

Claim 16

System control means as set forth in claim 11, wherein said mathematical process model means includes logic means for comparing the second output signals representing said predicted value (Y(k/k-1)) with the third output signals representing said estimated value (Y(k/k)) and effecting adjustment of a predicted value toward said estimated value.

Claim 17

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock, measuring means for measuring the weight of said paper sheet and providing first output signals representing the measured value of the weight of said paper sheet, valve means for controlling the flow rate of said liquid stock to said apparatus as a function of the set point of said valve means to establish a predetermined weight for said paper sheet, and system control means including mathematical process model means for providing second output signals representing a predicted value for the dry basis weight of said paper sheet, filter means responsive to said first and second output signals for providing third output signals representing an estimated value for said dry basis weight, controller means responsive to said third output signals to provide fourth output signals representing an adjusted set point value for the stock flow valve means, and output means responsive to said fourth output signals to provide control signals for said stock flow means for adjusting the set point of said stock flow valve means whenever the measured value of the dry basis weight deviates from a desired value due to a process upset.

Claim 18

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock, measuring means for measuring the weight of said paper sheet and providing first output signals representing the measured value of the weight of said paper sheet, valve means for controlling the flow rate of said liquid stock to said apparatus as a function of the set point of said valve means to establish a predetermined weight for said paper sheet, and system control means including mathematical process model means for providing second output signals representing a predicted value for the dry basis weight of said paper sheet, filter means responsive to said first and second output signals for providing third output signals representing an estimated value for said dry basis weight, controller means responsive to said third output signals to provide fourth output signals representing an adjusted set point value for the stock flow valve means, and output means responsive to said fourth output signals to provide control signals for said stock flow means for adjusting the set point of said stock flow valve means whenever the measured value of the dry basis weight deviates from a desired value due to a process upset.

Claim 19

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock including head box means for containing liquid stock and for depositing said liquid stock on conveyor means in a continuous paper web, measuring means for measuring the basis weight and moisture content of said paper web to provide first output signals representing a measured value of the dry basis weight of said paper web, stock flow valve means for controlling the flow rate of liquid stock through said head box means as a function of the set point of said stock flow valve means to thereby control the basis weight of said paper web, and system control means comprising mathematical process model means for providing second output signals representing a predicted value for the dry basis weight of said paper web, filter means responsive to said first and second output signals for providing third output signals representing an estimated value for said dry basis weight of said paper web, and controller means responsive to said third output signals for computing an adjusted set point value for said stock flow valve means to provide control signals for said stock flow valve means for adjusting the set point of said stock flow valve means whenever the measured value of the dry basis weight deviates from a desired value due to a process upset.

Claim 20

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock including head box means for containing liquid stock and for depositing said liquid stock on conveyor means in a continuous paper web, measuring means for measuring the basis weight and moisture content of said paper web to provide first output signals representing a measured value of the dry basis weight of said paper web, stock flow valve means for controlling the flow rate of liquid stock through said head box means as a function of the set point of said stock flow valve means to thereby control the basis weight of said paper web, and system control means comprising mathematical process model means for providing second output signals representing a predicted value for the dry basis weight of said paper web, filter means responsive to said first and second output signals for providing third output signals representing an estimated value for said dry basis weight of said paper web, and controller means responsive to said third output signals for computing an adjusted set point value for said stock flow valve means to provide control signals for said stock flow valve means for adjusting the set point of said stock flow valve means whenever the measured value of the dry basis weight deviates from a desired value due to a process upset.

Claim 21

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock including head box means for containing liquid stock and for depositing said liquid stock on conveyor means in a continuous web, dryer means for removing moisture from said paper web as said paper web is moved by said conveyor means from said head box means tO take-up reel means, steam valve means for regulating the drying process of said dryer means, measuring means for measuring the moisture content of said paper web to provide first output signals representing the measured value of the moisture content of said paper web, and system control means comprising mathematical process model means for providing second output signals representing a predicted value for the moisture content of said paper web, filter means responsive to said first and second output signals for providing third output signals representing an estimated value for the moisture content of said paper web, and controller means responsive to said third output signals for calculating an adjusted set point value for said steam valve means to provide control signals to said steam valve means for adjusting the set point of said steam valve means whenever the measured value of the moisture content of said paper sheet deviates from a desired value due to a process upset.

Claim 22

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock including head box means for containing liquid stock and for depositing said liquid stock on conveyor means in a continuous web, dryer means for removing moisture from said paper web as said paper web is moved by said conveyor means from said head box means tO take-up reel means, steam valve means for regulating the drying process of said dryer means, measuring means for measuring the moisture content of said paper web to provide first output signals representing the measured value of the moisture content of said paper web, and system control means comprising mathematical process model means for providing second output signals representing a predicted value for the moisture content of said paper web, filter means responsive to said first and second output signals for providing third output signals representing an estimated value for the moisture content of said paper web, and controller means responsive to said third output signals for calculating an adjusted set point value for said steam valve means to provide control signals to said steam valve means for adjusting the set point of said steam valve means whenever the measured value of the moisture content of said paper sheet deviates from a desired value due to a process upset.

Claim 23

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock, valve means for providing liquid stock to said apparatus at a predetermined flow rate, drying means for removing moisture from said paper sheet at a predetermined drying rate, measuring means for providing measurements of the basis weight and the moisture content of said paper sheet, and system control means including basis weight controller means for deriving from said measurements a first estimated value for the dry basis weight of said paper and comparing said first estimated value to a dry stock flow set point to provide control signals for said valve means for adjusting the flow rate of said liquid stock whenever the dry basis weight of said paper sheet deviates from a desired weight, and moisture controller means for deriving from said measurements a second estimated value for the moisture content of said paper sheet and comparing said second estimated value to a moisture content set point to provide further control signals for said drying means for adjusting the drying rate of said drying means whenever the moisture content of said paper sheet deviates from a desired amount.

Claim 24

In a paper manufacturing process, apparatus for manufacturing a continuous sheet of paper from liquid stock, valve means for providing liquid stock to said apparatus at a predetermined flow rate, drying means for removing moisture from said paper sheet at a predetermined drying rate, measuring means for providing measurements of the basis weight and the moisture content of said paper sheet, and system control means including basis weight controller means for deriving from said measurements a first estimated value for the dry basis weight of said paper and comparing said first estimated value to a dry stock flow set point to provide control signals for said valve means for adjusting the flow rate of said liquid stock whenever the dry basis weight of said paper sheet deviates from a desired weight, and moisture controller means for deriving from said measurements a second estimated value for the moisture content of said paper sheet and comparing said second estimated value to a moisture content set point to provide further control signals for said drying means for adjusting the drying rate of said drying means whenever the moisture content of said paper sheet deviates from a desired amount.

Claim 25

In a paper manufacturing process for manufacturing a continuous paper sheet from liquid stock, a closed loop process control system comprising dry stock flow controller means connected in a first control loop for comparing a value of dry stock flow to a dry stock flow set point and calculating an adjusted value for a set point for a stock flow valve whenever the dry stock flow differs from the dry stock flow set point, and basis weight controller means connected in a second control loop which includes said first control loop for comparing a value for the dry basis weight of said paper sheet obtained from measurements to a dry basis weight set point and calculating an adjusted value for the dry stock flow set point for said dry stock flow controller means whenever the dry basis weight differs from the dry basis weight set point.

Claim 26

In a paper manufacturing process for manufacturing a continuous paper sheet from liquid stock, a closed loop process control system comprising dry stock flow controller means connected in a first control loop for comparing a value of dry stock flow to a dry stock flow set point and calculating an adjusted value for a set point for a stock flow valve whenever the dry stock flow differs from the dry stock flow set point, and basis weight controller means connected in a second control loop which includes said first control loop for comparing a value for the dry basis weight of said paper sheet obtained from measurements to a dry basis weight set point and calculating an adjusted value for the dry stock flow set point for said dry stock flow controller means whenever the dry basis weight differs from the dry basis weight set point.

Claim 27

In a paper manufacturing process for manufacturing a continous paper sheet the method of controlling process apparatus to maintain the weight of said paper sheet at a predetermined value, said method comprising the steps of deriving a value of the dry basis weight of said paper sheet from measured values of the basis weight and the moisture content of said paper sheet, calculating a dry stock flow set point using the derived value for said dry basis weight, calculating the dry stock flow rate, comparing the dry stock flow set point to the dry stock flow rate, calculating an adjusted value for the set point of valve means of said process apparatus whenever the difference between the dry stock flow set point and the dry stock flow rate exceeds a predetermined amount, and providing control signals to said process apparatus for changing the set point of said valve means to said adjusted value.

Claim 28

In a paper manufacturing process for manufacturing a continous paper sheet the method of controlling process apparatus to maintain the weight of said paper sheet at a predetermined value, said method comprising the steps of deriving a value of the dry basis weight of said paper sheet from measured values of the basis weight and the moisture content of said paper sheet, calculating a dry stock flow set point using the derived value for said dry basis weight, calculating the dry stock flow rate, comparing the dry stock flow set point to the dry stock flow rate, calculating an adjusted value for the set point of valve means of said process apparatus whenever the difference between the dry stock flow set point and the dry stock flow rate exceeds a predetermined amount, and providing control signals to said process apparatus for changing the set point of said valve means to said adjusted value.

Claim 29

The method as set forth in claim 18, wherein calculating the dry stock flow set point includes providing the functional relationship between the dry basis weight and the dry stock flow set point in a stored mathematical model, deriving a predicted value for the dry basis weight from said mathematical model, calculating an estimated value for the dry basis weight from said predicted value and the dry basis weight value obtained by said measurements, and using said estimated value to calculate said dry stock flow set point.

Claim 30

In a paper manufacturing process for manufacturing a continuous sheet of paper, the method of controlling process apparatus to maintain the weight of said paper sheet at a predetermined value, said method comprising the steps of establishing in a stored mathematical model the functional relationship between the weight of said paper sheet and a set point (R(k)) for a control valve of said process apparatus, deriving a predicted value (Y(k/k-1)) for the weight of said paper sheet from said mathematical model, deriving an estimated value (Y(k/k)) for the weight of said paper sheet from the predicted value and a measured value (Z(k)) for the weight of said paper sheet obtained from a measurement, calculating an adjusted value (X(k)) for the set point for said valve means using said estimated value to obtain the set point required for said valve means to provide a desired weight for said paper sheet, and providing control signals for said process apparatus for changing the set point of said valve means to said adjusted value.

Claim 31

In a paper manufacturing process for manufacturing a continuous sheet of paper, the method of controlling process apparatus to maintain the weight of said paper sheet at a predetermined value, said method comprising the steps of establishing in a stored mathematical model the functional relationship between the weight of said paper sheet and a set point (R(k)) for a control valve of said process apparatus, deriving a predicted value (Y(k/k-1)) for the weight of said paper sheet from said mathematical model, deriving an estimated value (Y(k/k)) for the weight of said paper sheet from the predicted value and a measured value (Z(k)) for the weight of said paper sheet obtained from a measurement, calculating an adjusted value (X(k)) for the set point for said valve means using said estimated value to obtain the set point required for said valve means to provide a desired weight for said paper sheet, and providing control signals for said process apparatus for changing the set point of said valve means to said adjusted value.

Claim 32

The method as set forth in claim 20, which includes a further step of adjusting the mathematical model at the time of deriving the estimated value, including comparing the predicted value (Y(k/k-1)) last derived from the mathematical model to a previous estimated value (Y(k-1/k-1)) which was derived from a preceding measurement, and adjusting the mathematical model whenever the last predicted value (Y(k/k-1)) is less than or greater than the previous estimated value (Y(k-1/k-1)) by a predetermined amount.

Claim 33

The method as set forth in claim 21, wherein the resultant adapting of the mathematical model includes adjusting the last predicted value (Y(k/k-1)) towards the last derived estimated value (Y(k/k)).

Claim 34

The method as set forth in claim 22, wherein the mathematical model is characterized by a predetermined gain (G) and wherein the gain of said mathematical model is adjusted as a function of the value of the last predicted value (Y(k/k-1)) relative to the value of the last estimated value (Y(k/k)).

Claim 35

The method as set forth in claim 23, wherein the gain of the mathematical model is increased whenever the last predicted value (Y(k/k-1)) is less than the last estimated value (Y(k/k)) and the gain of the mathematical model is decreased whenever the last predicted value (Y(k/k-1)) is greater than the last estimated value (Y(k/k)) such that with each gain change a future predicted value (Y(k + 1/k)) derived from the stored mathematical model will approach the estimated value.

Claim 36

The method as set forth in claim 23, wherein the adjustment of said mathematical model includes calculating a gain factor and adjusting the gain in predetermined increments of said gain factor.

Claim 37

The method as set forth in claim 25, wherein the gain of said mathematical model is adjusted in increments of two per cent of the gain factor.

Claim 38

The method as set forth in claim 25, wherein the gain is adjusted within predetermined limits.

Claim 39

The method as set forth in claim 21, wherein the derivation of the last estimated value (Y(k/k)) for the weight of the paper sheet includes calculating a weighted average of the last predicted value (Y(k/k-1)) and the measured value (Z(k)) of the weight of said paper sheet.