IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0126768
(2002-04-19)
|
발명자
/ 주소 |
- Wojsznis, Wilhelm K.
- Blevins, Terrence L.
- Thiele, Dirk
- Gudaz, John A.
|
출원인 / 주소 |
- Fisher-Rosemount Systems, Inc.
|
대리인 / 주소 |
|
인용정보 |
피인용 횟수 :
13 인용 특허 :
10 |
초록
▼
An adaptive process controller drives a process variable to be substantially equivalent to a set point and adapts the controller gain, the controller reset, and/or the controller rate, based on model free adaptation. The adaptive controller combines a controller gain computed from an oscillation ind
An adaptive process controller drives a process variable to be substantially equivalent to a set point and adapts the controller gain, the controller reset, and/or the controller rate, based on model free adaptation. The adaptive controller combines a controller gain computed from an oscillation index with a controller gain computed from a steady state estimate and that adapts the controller reset/rate by forcing the ratio of two of the controller proportional, integral or derivative terms to be equal to a predetermined value.
대표청구항
▼
1. A tuner for use in tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; a gain adaptation
1. A tuner for use in tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; a gain adaptation module which uses the error signal to develop a gain for use in tuning the process controller, wherein the gain adaptation module calculates the gain as a combination of an oscillation index gain and of a gain calculated from a process steady state gain; and a second module which monitors the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, modifies the gain to prevent the process variable from oscillating. 2. The tuner of claim 1, wherein the first module further causes the gain adaptation module to develop the gain when the error signal is greater than a predetermined threshold.3. The tuner of claim 2, further comprising an excitation generator which produces an excitation signal and a summer which sums the error signal with the excitation signal to cause the error signal to be greater than the predetermined threshold.4. The tuner of claim 1, wherein the second module detects when the process variable oscillates by detecting when the error signal crosses a zero point.5. The tuner of claim 1, wherein the second module modifies the gain to prevent the process variable from oscillating by multiplying the gain by a dampening constant.6. The tuner of claim 5, wherein the dampening constant is a positive constant less than one.7. The tuner of claim 1, wherein the gain adaptation module further develops one of a proportional-integral-derivative (PID) controller gain, a proportional-integral (PI) controller gain, or a proportional-derivative (PD) controller gain.8. The tuner of claim 1, wherein the gain adaptation module computes the gain as a weighted summation of the gain calculated from the oscillation index gain and of the gain calculated from the process steady state gain.9. The tuner of claim 1, wherein the process controller is a fuzzy logic controller and further comprising a translation module which uses the gain developed by the gain adaptation module to develop a change in controller output scaling factor for use in tuning the fuzzy logic controller.10. The tuner of claim 9, further comprising a second module which monitors the process variable to detect when the process variable oscillates, and upon the detection of the process variable oscillating, modifies the change in controller output scaling factor to prevent the process variable from oscillating.11. The tuner of claim 10, wherein the second module detects when the process variable oscillates by detecting when the error signal crosses a zero point.12. The tuner of claim 10, wherein the second module modifies the change in controller output scaling factor term to prevent the process variable from oscillating by multiplying the change in controller output scaling factor term by a dampening constant.13. The tuner of claim 12, wherein the dampening constant is a positive constant less than one.14. The tuner of claim 9, wherein the translation module calculates the change in controller output scaling factor as:SΔu=XSΔeKc wherein:SΔu=the change in controller output scaling factor; X=a preset constant between the values of 2 and 4; SΔe=a change in error scaling factor; Kc=a gain for use in tuning the process controller; wherein:SΔu=βsfΔSP wherein:ΔSP=a change in the set point relative to a nominal set point 1% change; βsf=a determined value. 15. The tuner of claim 14, wherein the translation module calculates the determined value (βsf) as: wherein:D=a process dead time; and Tc=a process lag. 16. A tuner for use in tuning a process controller which set a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; a gain adaptation module which uses the error signal to develop a gain for use in tuning the process con controller, wherein the gain adaptation module calculates the gain as a combination of oscillation index gain and of a gain calculated from a process steady state gain; and a second module which monitors the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, modifies the gain to prevent the process variable from oscillating; wherein the gain adaptation module computes the controller gain from the process steady state gain as wherein:KcSS=the gain calculated from the process steady state gain; PV0=the process variable before a change in the set point; OUT0=a controller output before the change in the set point; PV1=a steady state process variable after the change in the set point; OUT1=a steady state controller output after the change in the set point; λ=a lambda controller tuning factor; DT=a process deed time; and Ti(k)=a controller reset at a time k. 17. A tuner for use in tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point an the process variable; a gain adaptation module which uses the error signal to develop a gain for use in tuning the process controller, wherein the gain adaptation module calculates the gain as a combination of an oscillation index gain and of a gain calculated from a process steady state gain; and a second module which monitors the process variable to detect when the process variable oscillate, and upon detection of the process variable oscillating, modifies the gain to prevent the process variable from oscillating; wherein the gain adaptation module computes the controller gain from the oscillation index as KcOSC=KcOSC(k)+ΔKcOSC(k) wherein:KcOSC(k)=the gain calculated from the oscillation index at a time k; KcOSC(k?1)=the gain calculated from the oscillation index at a time k?1; ΔKcOSC(k)=a change in the gain calculated from the oscillation index at a time k; wherein:ΔKcOSC(k)=χKcOSC(k)(W(k)+Wref) wherein:χ=a predetermined constant between the values of 0.02 and 0.05; Wref=a preset value chosen from the interval ?1 to +1; W(k)=a an oscillation index at a time k; wherein: wherein:i=int((k?k0)/2); ko=a time at which adaptation started; and e(k)=the error signal at a time k. 18. The tuner of claim 17, wherein the gain adaptation module further reduces the effect of noise by changing the preset value Wref by a determined value dependent upon the noise level.19. A tuner for use in tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; an adaptation module which uses the error signal to calculate at least two of a proportional term, a controller integral time, and a derivative term for use in tuning the process controller, wherein the adaptation module calculates a first and a second of the proportional, integral and derivative terms so that a ratio of a change in the first term and of a change in the second term is equal to a preset constant; and a second module which monitors the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, modifies at least one of the proportional term, the controller integral time, and the derivative term to prevent the process variable from oscillating. 20. The tuner of claim 19, wherein the first module further causes the adaptation module to calculate the first term and the second term when the error signal is greater than a predetermined threshold.21. The tuner of claim 20, further comprising an excitation generator which produces an excitation signal and a summer which sums the error signal with the excitation signal.22. The tuner of claim 19, wherein the adaptation module uses the preset constant such that the product of the preset constant and a lambda-tuning constant is equal to approximately one.23. The tuner of claim 19, wherein the adaptation module uses the preset constant such that the product of the preset constant and a lambda-tuning constant is greater than one.24. The tuner of claim 19, wherein the adaptation module uses the preset constant such that the product of the preset constant and a lambda-tuning constant is less than one.25. The tuner of claim 19, wherein the second module further detects when the process variable oscillates by detecting when the error signal crosses a zero point.26. The tuner of claim 19, wherein the second module modifies the at least one of the proportional term, the controller integral time, and the derivative term to prevent the process variable from oscillating by multiplying the at least one of the proportional term, the controller integral time, and the derivative term by a dampening constant.27. The tuner of claim 26, wherein the dampening constant is a positive constant less than one.28. The tuner of claim 19, wherein the process controller is a fuzzy logic controller and further comprising a translation module which uses the first and second of the terms calculated by the adaptation module to develop an error scaling factor for use in tuning the fuzzy logic controller.29. The tuner of claim 28, wherein the translation module calculates the error scaling factor as wherein:Se=the error scaling factor; Ti=the controller integral time; SΔe=a change in error scaling factor; and Δt=a controller scan period. 30. A tuner for use in tuning a process controller which set a set point and a profess variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; an adaptation module which uses the error signal to calculate at least two of a proportional term, an controller integral the and a derivative term for use in tuning the process controller, wherein the adaptation module calculates a first and a second of the proportional, integral and derivative terms so that a ratio of a change in the first term and of a change in the second term is equal to a preset constant; and a second module which monitors the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, modifies at least one of the proportional term, the controller integral time, and the derivative term to prevent the process variable from oscillating; wherein the adaptation module calculates the controller integral time as: Ti(k+1)=Ti(k)+ΔTi(k) wherein:Ti(k+1)=an controller integral time at a time k+1; Ti(k)=the controller integral time at the time k; ΔTi(k)=a change in the controller integral time at the time k; wherein: wherein:γ=a preset constant controlling the speed of reset adaptation which may be, for example, between 0.05 and 0.5; and wherein:ΔPk=the change in the proportional term at the time k; ΔIk=the change in the controller integral time at the time k; and α=a preset ratio. 31. A tuner for use in tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; a gain adaptation module which uses the error signal to develop a gain for use in tuning the process controller, wherein the gain adaptation module calculates the gain as a combination of an oscillation index gain and of a gain calculated from a process steady state gain; and an adaptation module which uses the error signal to calculate at least two of a proportional term, an controller integral time, and a derivative term for use in tuning the process controller, wherein the adaptation module calculates a first and a second of the proportional, integral, and derivative terms so that a ratio of a change in the first term and of a change in the second term is equal to a preset constant. 32. The tuner of claim 31, wherein the first module further causes the gain adaptation module to develop the gain and to cause the adaptation module to calculate the first term and the second term when the error signal is greater than a predetermined threshold.33. The tuner of claim 32, further comprising an excitation generator which produces an excitation signal and a summer which sums the error signal with the excitation signal.34. The tuner of claim 31, wherein the adaptation module compensates the controller integral time when the product of the change in the controller gain and the controller integral time is greater than zero.35. A tuner for use in tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuner comprising:a first module which develops an error signal representative of the difference between the set point and the process variable; a gain adaptation module which uses the error signal to develop a gain for use in tuning the process controller, wherein the gain adaptation module calculates the gain as a combination of an oscillation index gain and of a gain calculated from a process steady state gain; and an adaptation module which uses the error signal to calculate at least two of a proportional term, an controller integral time, and derivative term for use in tuning the process controller, wherein the adaptation module calculates a first and a second of the proportional integral, and derivative terms that a ratio of a change in the first term and of a change in the second term is equal to a preset constant; wherein the adaptation module compensates the controller integral time when the product of the chance in the controller gain and the controller integral time is greater than zero; and wherein the adaptation block compensates the controller integral time as: whereinΔTi(k)corr=a compensated change in the controller integral time at a time k; ΔTi(k)=a change in the controller integral time at a time k; ΔKc(k)=a change in the gain at the time k; Kc(k)old=the gain at the time k before the adaptation module compensates; and κ=a degree of reset compensation. 36. A method of tuning a process controller which uses a set point and a process variable to generate a process control signal, the method comprising the steps of:developing an error signal representative of the difference between the set point and the process variable; and calculating at least two of a proportional term, an controller integral time, and a derivative term for use in tuning the process controller, wherein a ratio of a change in first term and of a change in the second term is equal to a preset constant. 37. The method of claim 36, further comprising the step of calculating at least two of the proportional term, the controller integral time, and the derivative term when the error signal is greater than a predetermined threshold.38. The method of claim 37, further comprising the steps of introducing an excitation signal and summing the error signal with the excitation signal.39. The method of claim 36, further comprising the step of monitoring the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, modifying at least one of the proportional term, the controller integral time, and the derivative term to prevent the error signal from oscillating.40. The method of claim 36, further comprising the step of developing a gain for use in tuning the process controller, wherein the gain is a combination of an oscillation index gain and of a gain calculated from a process steady state gain.41. A tuning system for tuning a process controller which uses a set point and a process variable to generate a process control signal, the tuning system comprising:a memory; a first logic routine stored in the memory and executed on a processor to S develop an error signal representative of the difference between the set point and the process variable; and a second logic routine stored in the memory and executed on a processor to develop a gain for use in tuning the process controller, wherein the gain is a combination of a gain calculated from an oscillation index gain and of a gain calculated from a process steady state gain. 42. The tuning system of claim 41, further comprising a third logic routine stored in the memory and executed on a processor to develop the gain when the error signal is greater than a predetermined threshold.43. The tuning system of claim 41, further comprising a third logic routine stored in the memory and executed on a processor to calculate at least two of a proportional term, an controller integral time, and a derivative term for use in tuning the process controller, wherein a ratio of a change in the first term and of a change in the second term is equal to a preset constant.44. The tuning system of claim 43, further comprising a fourth logic routine stored in the memory and executed on a processor to develop the gain when the error signal is greater than a predetermined threshold.45. The tuning system of claim 43, further comprising a fourth logic routine stored in the memory and executed on a processor to monitor the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, to modify the gain to prevent the process variable from oscillating.46. A process control system comprising:a process controller which uses a set point and a process variable to generate a process control signal; a tuner for tuning the process controller; a first module which develops an error signal representative of the difference between the set point and the process variable; and a gain adaptation module which uses the error signal to develop a gain for use in tuning the process controller, wherein the gain adaptation module calculates the gain as a combination of a gain calculated from an oscillation index gain and of a gain calculated from a process steady state gain. 47. The process control system of claim 46, wherein the first module further causes the gain adaptation module to develop the gain when the error signal is greater than a predetermined threshold.48. The process control system of claim 46, further comprising a second module which monitors the process variable to detect when the process variable oscillates, and upon detection of the process variable oscillating, modifies the gain to prevent the process variable from oscillating.49. The process control system of claim 48, wherein the second module modifies the gain to prevent the process variable from oscillating by multiplying the gain by a dampening constant.50. The process control system of claim 48, further comprising an adaptation module which uses the error signal to calculate at least two of a proportional term, an controller integral time, and a derivative term for use in tuning the process controller, wherein the adaptation module calculates a first and a second of the proportional, integral, and derivative terms so that a ratio of a change in first term and of a change in the second term is equal to a preset constant.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.