Method and apparatus for process control with opimum setpoint determination

Claim 3

An apparatus according to claim 1, wherein said optimum setpoint means comprises A. batch-characteristic means for generating a statistical-characteristic signal representative of a statistical characteristic of each said batch of monitored values, and B. SIGXB means, coupled to said batch-characteristic means, for responding to statistical-characteristic signals of (ns) said batches of monitored values to generate a SIGXB signal representative of a standard deviation of arithmetic means of those (ns) batches.

Claim 4

An apparatus according to claim 2, wherein A. said batch-characteristic means comprises batch-range means for generating a batch-range signal representative of an arithmetic difference between a numerically minimum monitored value in each said batch of monitored values and a numerically maximum monitored value in that same batch, and B. said SIGXB means includes first generating means responsive to (ns) ones of said batch-range signals for generating said SIGXB signal.

Claim 5

An apparatus according to claim 3, wherein said first generating means includes means for generating said SIGXB signal in accord with the mathematical formula ##EQU10## where, SIGXB represents a value of said SIGXB signal, RB represents a numerical average of said plurality (ns) of range signals, d2 represents a Shewhart value corresponding to (ns), and n represents a quantity of said monitored values in each said batch.

Claim 6

An apparatus according to claim 2, wherein A. said batch-characteristic means comprises batch-variance means for generating a batch-variance signal representative of a standard deviation of values in each said batch of monitored values, and B. said SIGXB means includes second generating means responsive to (ns) ones of said batch-variance signals for generating said SIGXB signal.

Claim 7

An apparatus according to claim 5, wherein said second generating means includes means for generating said SIGXB signal in accord with the mathematical expression ##EQU11## where, SIGXB represents a value of said SIGXB signal, SB represents a numerical average of said (ns) batch-variance signals, c2 represents a Shewhart value corresponding to (n), and n represents a quantity of said monitored values in each said batch.

Claim 8

An apparatus according to claim 2, wherein A. said batch-characteristic means comprises batch-variance means for selectively generating a batch-variance signal representative of a standard deviation of values in each said batch of monitored values, B. said batch-characteristic means comprises batch-range means for selectively generating a batch-range signal representative of an arithmetic difference between a numerically minimum monitored value in each said batch of said monitored values and a numerically maximum monitored value in that same batch, C. said optimum setpoint means includes first generating means selectively responsive to (ns) batch-range signals for generating said SIGXB signal, and D. said optimum setpoint means includes second generating means selectively responsive to (ns) batch-variance signals for generating said SIGXB signal.

Claim 9

An apparatus according to claim 7, comprising selection means for invoking, in accord with an average quantity of monitored values in said (ns) batches of monitored values, at least a selected one of said batch-variance means, said batch-range means, said first generating means and said second generating means.

Claim 10

An apparatus according to claim 8, wherein said selection means includes A. means responsive to a numerical value of (ns) being less than or equal to ten for invoking said batch-range means and said first generating means, and B. means responsive to a numerical value of (ns) being greater than ten for invoking said batch-variance means and said second generating means

Claim 12

An apparatus according to claim 10, wherein said optimum setpoint means comprises summing means for generating said setpoint signal as a arithmetic sum of said scaled variance signal and a selected specification limit.

Claim 13

An apparatus according to claim 11, comprising means for receiving at least one of said selected scaling factor and said selected specification limit from an operator input device.

Claim 14

An apparatus according to claim 1, wherein said A. monitoring means includes measuring means for measuring a value of said characteristic of said process to generate a signal representative thereof, and B. said optimum setpoint means includes sampling means, coupled to said measuring means, for forming said batch of monitored values by sampling said characteristic value-representative signal (n) times and for generating a signal representative thereof, where (n) is an integer greater than one.

Claim 15

An apparatus according to claim 13, wherein said sampling means comprises consecutive sample means for forming said batch of monitored values by sampling said (n) characteristic value representative signals beginning every (n) values.

Claim 16

An apparatus according to claim 14, wherein said sampling means comprises skip-m sample means for forming said batch of monitored values by sampling said (n) first characteristic value representative signals beginning every (n+m) values, where (m) is an integer greater than or equal to one.

Claim 17

An apparatus according to claim 1, further comprising A. batch-mean means for generating a batch-mean signal representative of an arithmetic mean of a batch of said monitored values, B. XBB means for generating an XBB signal representative of an arithmetic average of arithmetic averages of a plurality of batches of monitored values, C. SIGXB means for generating a SIGXB signal representative of a standard deviation of arithmetic means of (ns) batches of monitored values, and D. rule means for detecting at least one of the following conditions and for generating an error signal in response thereto: i) the most recent value of the batch-mean signal XB(j) is a) greater than a value of the XBB signal plus three times a value of the SIGXB signal, or b) less than a value of the XBB signal minus three times a value of the SIGXB signal, ii) starting with the most recent values of each of three consecutive the batch-mean signals, XB(j), vary from one another in either direction by an amount equal to or greater than a value of three times a value of the SIGXB signal, iii) starting with the most recent values of two of three consecutive the batch-mean signals, XB(j), are a) greater than a value of the XBB signal plus two times value of the SIGXB signal, b) less than a value of the XBB signal minus two times a value of the SIGXB signal, iv) starting with the most recent values of four of five consecutive the batch-mean signals, XB(j), are a) greater than a value of the XBB signal plus the value of the SIGXB signal, b) less than a value of the XBB signal minus the value of the SIGXB signal, v) starting with the most recent values of eight consecutive batch-mean signals, XB(j), are greater than a value of the XBB signal, vi) starting with the most recent values of eight consecutive batch-mean signals, XB(j), are less than a value of the XBB signal, vii) starting with the most recent values of five consecutive batch-mean signals, XB(j), are of increasing value, viii) starting with the most recent values of five consecutive batch-mean signals, XB(j), are of decreasing increasing value, ix) starting with the most recent values of fifteen consecutive batch-mean signals, XB(j), are a) less than a value of the XBB signal plus the value of the SIGXB signal, and b) less than a value of the XBB signal minus the value of the SIGXB signal, x) starting with the most recent values eight consecutive batch-mean signals, XB(j), are a) greater than a value of the XBB signal plus the value of the SIGXB signal, b) less than a value of the XBB signal minus the value of the SIGXB signal, xi) starting with the most recent value, after a value of a first batch-mean signal, XB(j), that varies from the previous XB signal, in either direction, by an amount greater than three times the value of the SIGXB signal, then values of three consecutive batch-mean signals, XB(j), are compared to check if they are a) greater than 0.75 times a value of the SIGXB signal minus a value of the previous batch-mean signal, and b) less than 0.75 times a value of the SIGXB signal plus a value of the previous first batch-mean signal, and E. said optimum setpoint means includes error/bypass means responsive to said error signal for at least one of preventing generation of said setpoint signal and generating an operator alert signal for output on an operator output device.

Claim 18

An apparatus according to claim 1, wherein said optimum set point means comprises A. normality means for generating a normality signal representative of a statistical normality measure of a batch of measured values, and B. means coupled to said normality means for responding to a normality signal having a value outside a selected range for generating an error signal, and C. error/bypass means responsive to said error signal for at least one of preventing generation of said setpoint signal and generating an operator alert signal for output on an operator output device.

Claim 19

An apparatus according to claim 17, wherein said normality means includes at least one of A. skewness means for generating said normality signal as representative of a skewness of said batch of measured values, and B. kurtosis means for generating said normality signal as representative of a kurtosis of said batch of measured values.

Claim 20

An apparatus according to claim 17, further comprising means for receiving said selected range from an operator input device.

Claim 21

An apparatus according to claim 1, wherein said controller means includes closed-loop control means, coupled to said process-performing equipment, for applying said controlled variable signal to said process-performing equipment.

Claim 24

A method according to claim 21, wherein said setpoint generating step comprises A. generating a statistical-characteristic signal representative of a statistical characteristic of each said batch of monitored values, B. responding to statistical-characteristic signals of (ns) said batches of monitored values for generating a SIGXB signal representative of a standard deviation of arithmetic means of those (ns) batches.

Claim 25

A method according to claim 22, further comprising A generating a batch-range signal representative of an arithmetic difference between a numerically minimum monitored value in each said batch of monitored values and a numerically maximum monitored value in that same batch, and B. responding to (ns) ones of said batch-range signals for generating said SIGXB signal.

Claim 26

A method according to claim 23, further comprising generating said SIGXB signal in accord with the mathematical formula ##EQU12## where, SIGXB represents a value of said SIGXB signal, RB represents a numerical average of said plurality of range signals, d2 represents a Shewhart value corresponding to (ns), and n represents a quantity of said monitored values in each said batch of monitored values

Claim 27

A method according to claim 22, further comprising A. generating a batch-variance signal representative of a standard deviation of the values in each said batch of monitored values, and B. responding to (ns) ones of said batch-variance signals for generating said SIGXB signal.

Claim 28

A method according to claim 25, further comprising generating said SIGXB signal in accord with the mathematical expression ##EQU13## where, SIGXB represents a value of said SIGXB signal, SB represents a numerical average of said (ns) batch-variance signals, c2 represents a Shewhart value corresponding to (ns), and n represents a quantity of said monitored values in each said batch of monitored values.

Claim 29

A method according to claim 22, further comprising A. selectively generating a batch-variance signal representative of a standard deviation of values in each said batch of monitored values, B. selectively generating a batch-range signal representative of an arithmetic difference between a numerically minimum monitored value in each batch of said monitored values and a numerically maximum monitored value in that same batch, C. selectively responding to (ns) batch-variance signals for generating said SIGXB signal, and D. selectively responding to (ns) batch-range signals for generating said SIGXB signal.

Claim 30

A method according to claim 27, further comprising selectively executing, in accord with a quantity of monitored values in said batch, at least a selected one of said batch-variance signal generating step, said batch-range signal generating step, said batch-variance signal responding step, and said batch-range signal responding step.

Claim 31

A method according to claim 28, further comprising A. responding to a numerical value of (n) being less than or equal to ten for executing said batch-range generating and said batch-range responding steps, and B. responding to a numerical value of (ns) being greater than ten for executing said batch-variance generating and said batch-variance signal responding steps.

Claim 33

A method according to claim 30, further generating said setpoint signal as a arithmetic sum of said scaled variance signal and a selected specification limit.

Claim 34

A method according to claim 31, further comprising receiving at least one of said selected scaling factor and said selected specification limit from an operator input device.

Claim 35

A method according to claim 21, further comprising A. measuring a value of said characteristic of said process to generate a signal representative thereof, and B. forming a batch of monitored values by sampling said characteristic value-representative signal (n) times and for generating a signal representative thereof, where (n) is an integer greater than one.

Claim 36

A method according to claim 33, further comprising forming said batch of monitored values by sampling said (n) characteristic value-representative signals beginning every (n) values.

Claim 37

A method according to claim 34, further comprising forming said batch of monitored values by sampling said (n) first characteristic value-representative signals beginning every (n+m) values, where (m) is an integer greater than or equal to one.

Claim 38

A method according to claim 21, further comprising A. generating a batch-mean signal representative of an arithmetic mean of a batch of said monitored values, B. generating an XBB signal representative of an arithmetic average of arithmetic averages of a plurality of batches of monitored values, C. generating a SIGXB signal representative of a standard deviation of arithmetic means of (ns) batches of monitored values, D. detecting at least one of the following conditions and for generating an alarm signal in response thereto: i) the most recent value of the batch-mean signal XB(j) is a) greater than a value of the XBB signal plus three times a value of the SIGXB signal, or b) less than a value of the XBB signal minus three times a value of the SIGXB signal, ii) starting with the most recent values of each of three consecutive the batch-mean signals, XB(j), vary from one another in either direction by an amount equal to or greater than a value of three times a value of the SIGXB signal, iii) starting with the most recent values of two of three consecutive the batch-mean signals, XB(j), are a) greater than a value of the XBB signal plus two times value of the SIGXB signal, b) less than a value of the XBB signal minus two times a value of the SIGXB signal, iv) starting with the most recent values of four of five consecutive the batch-mean signals, XB(j), are a) greater than a value of the XBB signal plus the value of the SIGXB signal, b) less than a value of the XBB signal minus the value of the SIGXB signal, v) starting with the most recent values of eight consecutive batch-mean signals, XB(j), are greater than a value of the XBB signal, vi) starting with the most recent values of eight consecutive batch-mean signals, XB(j), are less than a value of the XBB signal, vii) starting with the most recent values of five consecutive batch-mean signals, XB(j), are of increasing value, viii) starting with the most recent values of five consecutive batch-mean signals, XB(j), are of decreasing increasing value, ix) starting with the most recent values of fifteen consecutive batch-mean signals, XB(j), are a) less than a value of the XBB signal plus the value of the SIGXB signal, and b) less than a value of the XBB signal minus the value of the SIGXB signal, x) starting with the most recent values of eight consecutive batch-mean signals, XB(j), are a) greater than a value of the XBB signal plus the value of the SIGXB signal, b) less than a value of the XBB signal minus the value of the SIGXB signal, xi) starting with the most recent value, after a value of a first batch-mean signal, XB(j), that varies from the previous XB signal, in either direction, by an amount greater than three times the value of the SIGXB signal, then values of three consecutive batch-mean signals, XB(j), are compared to check if they are a) greater than 0.75 times a value of the SIGXB signal minus a value of the previous batch-mean signal, and b) less than 0.75 times a value of the SIGXB signal plus a value of the previous first batch-mean signal, and E. responding to said error signal for at least one of preventing generation of said setpoint signal and generating an operator alert signal for output on an operator output device.

Claim 39

A method according to claim 21, further comprising A. generating a normality signal representative of a statistical normality measure of a batch of measured values, and B. responding to a normality signal having a value outside a selected range for generating an error signal, and C. responding to said error signal for at least one of preventing generation of said setpoint signal and generating an operator alert signal for output on an operator output device.

Claim 40

A method according to claim 37, further comprising at least one of the steps of A. generating said normality signal as representative of a skewness of said batch of measured values, and B. generating said normality signal as representative of a kurtosis of said batch of measured values.

Claim 41

A method according to claim 37, further comprising receiving said selected range from an operator input device.

Claim 42

A method according to claim 21, further comprising applying said controlled variable signal to said process-performing equipment.

Claim 43

In an apparatus for controlling a process performed with equipment that responds to a manipulated variable signal to vary a characteristic of that process, wherein said apparatus includes means for monitoring values of said characteristic of said process to generate a controlled variable signal representative thereof, and further includes controller means for responding to a setpoint signal for generating said manipulated variable signal for application to said process-performing equipment, the improvement comprising A. optimum setpoint means coupled to said monitoring means for generating a setpoint signal representative of a desired setting of said process-performing equipment, and B. said optimum setpoint means including means for generating each said setpoint signal in accord wtih a value of a standard deviation of arithmetic means of (ns) batches of said monitored process characteristic values identified by said controlled variable signals, where each said batch includes a plurality of monitored values, and where (ns) is an integer greater than one.

Claim 44

In an apparatus for controlling a process performed with equipment that responds to a manipulated variable signal to vary a characteristic of that process, wherein said apparatus includes means for monitoring values of said characteristic of said process to generate a controlled variable signal representative thereof, and further includes controller means for responding to a setpoint signal for generating said manipulated variable signal for application to said process-performing equipment, the improvement comprising A. optimum setpoint means coupled to said monitoring means for generating a setpoint signal representative of a desired setting of said process-performing equipment, and B. said optimum setpoint means including means for generating each said setpoint signal in accord wtih a value of a standard deviation of arithmetic means of (ns) batches of said monitored process characteristic values identified by said controlled variable signals, where each said batch includes a plurality of monitored values, and where (ns) is an integer greater than one.

Claim 45

An optimum setpoint control apparatus for controlling process-performing equipment that responds to a manipulated variable signal to vary a characteristic of that process and that includes monitoring means for monitoring values of said characteristic of said process to generate a controlled variable signal representative thereof, said optimum setpoint control apparatus comprising: A. optimum setpoint means coupled to said monitoring means for generating at least one setpoint signal representative of a desired setting of said process-performing equipment, B. said optimum setpoint means including means for generating each said setpoint signal in accord wtih a value of a standard deviation of arithmetic means of (ns) batches of said monitored process characteristic values identified by said controlled variable signals, where each said batch includes a plurality of monitored values, and where (ns) is an integer greater than one, and C. controller means, coupled to said optimum setpoint means, for responding to said setpoint signal for generating said manipulated variable signal for application to said process-performing equipment.

Claim 46

An optimum setpoint control apparatus for controlling process-performing equipment that responds to a manipulated variable signal to vary a characteristic of that process and that includes monitoring means for monitoring values of said characteristic of said process to generate a controlled variable signal representative thereof, said optimum setpoint control apparatus comprising: A. optimum setpoint means coupled to said monitoring means for generating at least one setpoint signal representative of a desired setting of said process-performing equipment, B. said optimum setpoint means including means for generating each said setpoint signal in accord wtih a value of a standard deviation of arithmetic means of (ns) batches of said monitored process characteristic values identified by said controlled variable signals, where each said batch includes a plurality of monitored values, and where (ns) is an integer greater than one, and C. controller means, coupled to said optimum setpoint means, for responding to said setpoint signal for generating said manipulated variable signal for application to said process-performing equipment.

Claim 47

An apparatus according to claim 42, wherein said optimum setpoint means comprises A. batch-characteristic means for generating a statistical-characteristic signal representative of a statistical characteristic of each said batch of monitored values, and B. SIGXB means, coupled to said batch-characteristic means, for responding to statistical-characteristic signals of (ns) said batches of monitored values to generate a SIGXB signal representative of a standard deviation of arithmetic means of those (ns) batches.

Claim 48

An apparatus according to claim 42, wherein said optimum setpoint means comprises sampling means, coupled to said measuring means, for forming said batch of monitored values by sampling said characteristic value-representative signal (n) times and for generating a signal representative thereof, where (n) is an integer greater than one.

Claim 49

An apparatus according to claim 42, wherein said optimum set point means comprises A. normality means for generating a normality signal representative of a statistical normality measure of a batch of measured values, and B. means coupled to said normality means for responding to a normality signal having a value outside a selected range for generating an error signal, and C. error/bypass means responsive to said error signal for at least one of preventing generation of said setpoint signal and generating an operator alert signal for output on an operator output device.