Patent EP0059506A2 successfully refreshed from Google Patents.
Patent JPS53136649A successfully refreshed from Google Patents.
Patent US6792336B1 successfully refreshed from Google Patents.
US6792336B1
Expired / Lapsed
US
⏱️ 0.0 years remaining
Learning-based controller for biotechnology processing, and method of using
Assignee: Battelle Energy Alliance LLC
•
Filing: 1999-05-13
•
Granted: 2004-09-14
•
Expires: 2019-05-13
Abstract
Abstract The present invention relates to process control where some of the controllable parameters are difficult or impossible to characterize. The present invention relates to process control in biotechnology of such systems, but not limited to. Additionally, the present invention relates to process control in biotechnology minerals processing. In the inventive method, an application of the present invention manipulates a minerals bioprocess to find local exterma (maxima or minima) for selected output variables/process goals by using a learning-based controller for bioprocess oxidation of minerals during hydrometallurgical processing. The learning-based controller operates with or without human supervision and works to find processor optima without previously defined optima due to the non-characterized nature of the process being manipulated.
Inventors
- 👤 John A. Johnson
- 👤 Daphne L. Stoner
- 👤 Eric D. Larsen
- 👤 Karen S. Miller
- 👤 Charles R. Tolle
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-07 20:44
CPC Classifications
G; G05; G05B; G05B13/00; G05B13/02; G05B13/0265; G05B13/0275; C; C22; C22B; C22B3/00; C22B3/18; Y; Y02; Y02P; Y02P10/00; Y02P10/20
Claims Analysis
Total Claims: 30
Claim 1
Independent Claim
283 words
A method of controlling a bioreactor comprising: establishing in an expert controller a split data distribution for at least one input variable, the split data distribution having a mean of N and a standard deviation, an upper limit, and a lower limit; running the bioreactor according to the at least one input variable as controlled by a biocontroller and first measuring at a diagnostic system at least one output variable from the reactor; changing in the expert controller a value of the at least one input variable from the mean to a new input variable value, N+1, to create a difference (N+1)−N; running the bioreactor according to the value of the it least one input variable as controlled by the biocontroller and second measuring at the diagnostic system the at least one output variable from the bioreactor, and if the at least one output variable is improved, in the expert controller: changing the mean of the at least one input variable to the new input variable value, N+1; changing the upper limit; and changing the lower limit, wherein a one of the upper and lower limit that is approached by the direction of the difference is changed more than a one of the upper and lower limit that is distanced by the direction of the difference; and if the at least one output variable is not improved, in the expert controller: not changing the mean of the at least one input variable; changing a one of the upper and lower limit that is approached by the direction of the difference by decreasing it; and not changing a one of the upper and lower limit that is distanced by the direction of the difference.
Claim 2
Independent Claim
283 words
A method of controlling a bioreactor comprising: establishing in an expert controller a split data distribution for at least one input variable, the split data distribution having a mean of N and a standard deviation, an upper limit, and a lower limit; running the bioreactor according to the at least one input variable as controlled by a biocontroller and first measuring at a diagnostic system at least one output variable from the reactor; changing in the expert controller a value of the at least one input variable from the mean to a new input variable value, N+1, to create a difference (N+1)−N; running the bioreactor according to the value of the it least one input variable as controlled by the biocontroller and second measuring at the diagnostic system the at least one output variable from the bioreactor, and if the at least one output variable is improved, in the expert controller: changing the mean of the at least one input variable to the new input variable value, N+1; changing the upper limit; and changing the lower limit, wherein a one of the upper and lower limit that is approached by the direction of the difference is changed more than a one of the upper and lower limit that is distanced by the direction of the difference; and if the at least one output variable is not improved, in the expert controller: not changing the mean of the at least one input variable; changing a one of the upper and lower limit that is approached by the direction of the difference by decreasing it; and not changing a one of the upper and lower limit that is distanced by the direction of the difference.
Claim 3
Dependent Claim
75 words
The method of controlling a bioreactor according to claim 1 , wherein when the at least one output variable comprises a plurality, the plurality including product productivity, P product , and biomass density, P biomass , further comprising: ascribing a relationship between the product productivity and the biomass density according to P=f(P product )+(1−f)(P biomass ), wherein P is overall process productivity and f is a weighting factor in a range from 0 to 1.
Claim 4
Dependent Claim
155 words
The method of controlling a bioreactor according to claim 2 wherein the weighting factor, f, is varied comprising: when a present productivity of a first product is greater than its previous maximum productivity, then the weighting factor is increased; but when the weighting factor, as increased, is greater than 1, then the weighting factor is equal to 1; when the present productivity of the first product is not greater than its previous maximum productivity, the weighting factor is decreased; when a present productivity of a second product is greater than its previous maximum productivity, then the weighting factor is decreased; when the present productivity of the second product is not greater than its previous maximum productivity, the weighting factor is increased; but when the weighting factor is greater than 1, then the weighting factor is equal to 1; and repeating the method at least once, wherein the weighting factor is a highest retained value thereof.
Claim 5
Dependent Claim
32 words
The method of controlling a bioreactor according to claim 1 , wherein changing the value of the at least one input variable includes ascribing a random value thereto within a predetermined range.
Claim 6
Dependent Claim
22 words
The method of controlling a bioreactor according to claim 1 , wherein one of the at least one input variable is pH.
Claim 7
Dependent Claim
38 words
The method of controlling a bioreactor according to claim 4 , wherein the predetermined range is determined by selecting at least one microorganism and determining a pH range within which the at least one microorganism will remain viable.
Claim 8
Dependent Claim
22 words
The method of controlling a bioreactor according to claim 1 , wherein one of the at least one input variable is temperature.
Claim 9
Dependent Claim
38 words
The method of controlling a bioreactor according to claim 4 , wherein the predetermined range is determined by selecting at least one microorganism and determining a temperature range within which the at least one microorganism will remain viable.
Claim 10
Dependent Claim
24 words
The method of controlling a bioreactor according to claim 1 , wherein one of the at least one input variable is microorganism flow rate.
Claim 11
Dependent Claim
41 words
The method of controlling a bioreactor according to claim 4 , wherein the predetermined range is determined by selecting at least one microorganism and determining a microorganism flow rate range within which the at least one microorganism will resist reactor washout.
Claim 12
Dependent Claim
28 words
The method of controlling a bioreactor according to claim 1 , wherein changing the value of the at least one input variable includes ascribing a random value thereto.
Claim 13
Independent Claim
221 words
A method of controlling a reactor comprising: establishing in an expert controller a split data distribution for at least one input variable, the split data distribution having a mean of N, a standard deviation, an upper limit, and a lower limit; running the reactor according to the at least one input variable as controlled by a biocontroller and first measuring at a diagnostic system at least one output variable; increasing in the expert controller a value of the at least one input variable from the mean to a new input variable value, N+1; running the reactor according to the value of the at least one input variable as controlled by the biocontroller and second measuring at the diagnostic system the at least one output variable, and if the at least one output variable is improved, in the expert controller: changing the mean of the at least one input variable to the new input variable value, N+1; changing the upper limit; and changing the lower limit by an amount equivalent to the difference (N+1)−N, wherein the upper limit is changed by a value greater than N+1−N; and if the at least one output variable is not improved in the expert controller: not changing the mean of the at least one input variable; not changing the lower limit; and decreasing the lower limit.
Claim 14
Independent Claim
221 words
A method of controlling a reactor comprising: establishing in an expert controller a split data distribution for at least one input variable, the split data distribution having a mean of N, a standard deviation, an upper limit, and a lower limit; running the reactor according to the at least one input variable as controlled by a biocontroller and first measuring at a diagnostic system at least one output variable; increasing in the expert controller a value of the at least one input variable from the mean to a new input variable value, N+1; running the reactor according to the value of the at least one input variable as controlled by the biocontroller and second measuring at the diagnostic system the at least one output variable, and if the at least one output variable is improved, in the expert controller: changing the mean of the at least one input variable to the new input variable value, N+1; changing the upper limit; and changing the lower limit by an amount equivalent to the difference (N+1)−N, wherein the upper limit is changed by a value greater than N+1−N; and if the at least one output variable is not improved in the expert controller: not changing the mean of the at least one input variable; not changing the lower limit; and decreasing the lower limit.
Claim 15
Dependent Claim
75 words
The method of controlling a reactor according to claim 12 , wherein when the at least one output variable comprises a plurality, the plurality including product productivity, −P product , and biomass density, −P biomass , further comprising: ascribing a relationship between the product productivity and the biomass density according to P=f(P product )+(1−f)(P biomass ), wherein P is overall process productivity and f is a weighting factor in a range from 0 to 1.
Claim 16
Dependent Claim
127 words
The method of controlling a reactor according to claim 13 , wherein the weighting factor, f, is varied comprising: when a present productivity of a first product is greater than its previous maximum productivity, then the weighting factor is increased; but when the weighting factor, as increased, is greater than 1, then the weighting factor is equal to 1; otherwise the weighting factor is decreased; otherwise when a present productivity of a second product is greater than its previous maximum productivity, then the weighting factor is decreased; otherwise the weighting factor is increased; but when the weighting factor is greater than 1, then the weighting factor is equal to 1; and repeating the method at least once, wherein the weighting factor is a highest retained value thereof.
Claim 17
Independent Claim
229 words
A method of controlling a reactor comprising: establishing in an expert controller a split data distribution for at least one input variable, the split data distribution having a mean of N and a standard deviation, an upper limit, and a lower limit; running the reactor according to the at least one input variable as controlled by a biocontroller and first measuring at a diagnostic system at least one output variable from the reactor; decreasing in the expert controller a value of the at least one input variable from the mean to a new input variable value, N+1; running the reactor according to the value of the at least one input variable as controlled by the biocontroller and second measuring at the diagnostic, system the at least one output variable, and if the at least one output variable is improved, in the expert controller: changing the mean of the at least one input variable to the new input variable value, N+1; changing the lower limit by a first amount; and changing the upper limit by an amount equivalent to a difference (N+1)−N, wherein the lower limit is changed by a value greater than (N+1)−N; and if the at least one output variable is not improved, in the expert controller: not changing the mean of the at least one input variable; not changing the upper limit; and increasing the lower limit.
Claim 18
Independent Claim
229 words
A method of controlling a reactor comprising: establishing in an expert controller a split data distribution for at least one input variable, the split data distribution having a mean of N and a standard deviation, an upper limit, and a lower limit; running the reactor according to the at least one input variable as controlled by a biocontroller and first measuring at a diagnostic system at least one output variable from the reactor; decreasing in the expert controller a value of the at least one input variable from the mean to a new input variable value, N+1; running the reactor according to the value of the at least one input variable as controlled by the biocontroller and second measuring at the diagnostic, system the at least one output variable, and if the at least one output variable is improved, in the expert controller: changing the mean of the at least one input variable to the new input variable value, N+1; changing the lower limit by a first amount; and changing the upper limit by an amount equivalent to a difference (N+1)−N, wherein the lower limit is changed by a value greater than (N+1)−N; and if the at least one output variable is not improved, in the expert controller: not changing the mean of the at least one input variable; not changing the upper limit; and increasing the lower limit.
Claim 19
Dependent Claim
75 words
The method of controlling a reactor according to claim 15 , wherein when the at least one output variable comprises a plurality, the plurality including product productivity, −P product , and biomass density, −P biomass , further comprising: ascribing a relationship between the product productivity and the biomass density according to P=f(P product )+(1−f)(P biomass ), wherein P is overall process productivity and f is a weighting factor in a range from 0 to 1.
Claim 20
Dependent Claim
127 words
The method of controlling a reactor according to claim 16 , wherein the weighting factor, f, is varied comprising: when a present productivity of a first product is greater than its previous maximum productivity, then the weighting factor is increased; but when the weighting factor, as increased, is greater than 1, then the weighting factor is equal to 1; otherwise the weighting factor is decreased; otherwise when a present productivity of a second product is greater than its previous maximum productivity, then the weighting factor is decreased; otherwise the weighting factor is increased; but when the weighting factor is greater than 1, then the weighting factor is equal to 1; and repeating the method at least once, wherein the weighting factor is a highest retained value thereof.
Claim 21
Independent Claim
214 words
A method of controlling an ore treatment process including inoculating the ore with at least one microorganism culture, the at least one microorganism culture being used to make mineral values within the ore more susceptible to extractive metallurgy recovery techniques, the method including adjusting at least one input control variable and monitoring at least one output monitoring variable, comprising: establishing in an expert controller a split data distribution for a selected input parameter having a nominal area of unity, a mean, a standard deviation, an upper limit, and a lower limit; in response to a run of the ore treatment process, first measuring at least one output variable in a diagnostic system; changing in an expert controller a value of the selected input parameter; in response to a further run of the ore treatment process according to the changed input parameter, second measuring the at least one output variable; and if the at least one output variable improves, in the expert controller: adjusting the split data distribution by changing the upper and lower limits, one of the limits being changed more than the other of the limits; but if the at least one output variable does not improve, in the expert controller: adjusting the split data distribution by changing only one of the limits.
Claim 22
Independent Claim
214 words
A method of controlling an ore treatment process including inoculating the ore with at least one microorganism culture, the at least one microorganism culture being used to make mineral values within the ore more susceptible to extractive metallurgy recovery techniques, the method including adjusting at least one input control variable and monitoring at least one output monitoring variable, comprising: establishing in an expert controller a split data distribution for a selected input parameter having a nominal area of unity, a mean, a standard deviation, an upper limit, and a lower limit; in response to a run of the ore treatment process, first measuring at least one output variable in a diagnostic system; changing in an expert controller a value of the selected input parameter; in response to a further run of the ore treatment process according to the changed input parameter, second measuring the at least one output variable; and if the at least one output variable improves, in the expert controller: adjusting the split data distribution by changing the upper and lower limits, one of the limits being changed more than the other of the limits; but if the at least one output variable does not improve, in the expert controller: adjusting the split data distribution by changing only one of the limits.
Claim 23
Dependent Claim
77 words
The method of controlling an ore treatment process according to claim 18 , wherein when the at least one output variable comprises a plurality, the plurality including product productivity, −P product , and biomass density, −P biomass , further comprising: ascribing a relationship between the product productivity and the biomass density according to P=f(P product )+(1−f)(P biomass ), wherein P is overall process productivity and f is a weighting factor in a range from 0 to 1.
Claim 24
Dependent Claim
129 words
The method of controlling an ore treatment process according to claim 19 , wherein the weighting factor, f, is varied comprising: when a present productivity of a first product is greater than its previous maximum productivity, then the weighting factor is increased; but when the weighting factor, as increased, is greater than 1, then the weighting factor is equal to 1; otherwise the weighting factor is decreased; otherwise when a present productivity of a second product is greater than its previous maximum productivity, then the weighting factor is decreased; otherwise the weighting factor is increased; but when the weighting factor is greater than 1, then the weighting factor is equal to 1; and repeating the method at least once, wherein the weighting factor is a highest retained value thereof.
Claim 25
Independent Claim
378 words
A method of controlling a bioreactor with at least one input parameter and at least one output parameter, comprising the steps: establishing in an expert controller a data distribution for the at least one input parameter, each of the data distribution having an upper fixed limit and a lower fixed limit, a mean, and a standard deviation; randomly changing in the expert controller each of the at least one input parameter, updating the standard deviation, and calculating a change in its standard deviation; monitoring in a diagnostic system any change in the at least one output parameter; if the change in the at least one output parameter more closely approximates a desired output parameter: quantifying in the diagnostic system a difference in the at least one input parameter; if the difference in the at least one input parameter is an increase, in the expert controller: updating a positive standard deviation by adding the change in the standard deviation, and updating a negative standard deviation by subtracting the change in the standard deviation; if the difference in the at least one input parameter is a decrease: updating a positive standard deviation by subtracting the change in the standard deviation, and updating a negative standard deviation by adding the change in the standard deviation; if the change in the at least one output parameter does not more closely approximate the desired output parameter: quantifying in the diagnostic system a difference in the at least one input parameter; if the difference in the at least one input parameter is an increase in the expert controller: updating a positive standard deviation by subtracting the change in the standard deviation, and updating a negative standard deviation by adding the change in the standard deviation; if the difference in the at least one input parameter is a decrease: updating a positive standard deviation by adding the change in the standard deviation, and updating a negative standard deviation by subtracting the change in the standard deviation; if the positive standard deviation is less than about 0.001, ascribing in the expert controller a value thereof to about 0.001; if the negative standard deviation is less than about 0.001, ascribing in the expert controller a value thereof to about 0.001; and repeating the steps at least once.
Claim 26
Independent Claim
378 words
A method of controlling a bioreactor with at least one input parameter and at least one output parameter, comprising the steps: establishing in an expert controller a data distribution for the at least one input parameter, each of the data distribution having an upper fixed limit and a lower fixed limit, a mean, and a standard deviation; randomly changing in the expert controller each of the at least one input parameter, updating the standard deviation, and calculating a change in its standard deviation; monitoring in a diagnostic system any change in the at least one output parameter; if the change in the at least one output parameter more closely approximates a desired output parameter: quantifying in the diagnostic system a difference in the at least one input parameter; if the difference in the at least one input parameter is an increase, in the expert controller: updating a positive standard deviation by adding the change in the standard deviation, and updating a negative standard deviation by subtracting the change in the standard deviation; if the difference in the at least one input parameter is a decrease: updating a positive standard deviation by subtracting the change in the standard deviation, and updating a negative standard deviation by adding the change in the standard deviation; if the change in the at least one output parameter does not more closely approximate the desired output parameter: quantifying in the diagnostic system a difference in the at least one input parameter; if the difference in the at least one input parameter is an increase in the expert controller: updating a positive standard deviation by subtracting the change in the standard deviation, and updating a negative standard deviation by adding the change in the standard deviation; if the difference in the at least one input parameter is a decrease: updating a positive standard deviation by adding the change in the standard deviation, and updating a negative standard deviation by subtracting the change in the standard deviation; if the positive standard deviation is less than about 0.001, ascribing in the expert controller a value thereof to about 0.001; if the negative standard deviation is less than about 0.001, ascribing in the expert controller a value thereof to about 0.001; and repeating the steps at least once.
Claim 27
Dependent Claim
70 words
The method of controlling a system according to claim 21 , wherein the system comprises at least one microbial strain and at least one material to be processed, wherein either or each of the at least one microbial strain and the at least one material to be processed is an uncharacterized microbial strain, and wherein the at least one input parameter includes at least one chemical, physical, or biological parameter.
Claim 28
Dependent Claim
98 words
A process controller system including sub-components comprising: a stochastic learning algorithm that uses a split data distribution; a sensor package; at least one set point sub-controller; an actuator package; a productivity measure comprising a set of process goals and/or process outputs; an interconnecting communication system between sub-components; wherein the controller system acquires information from the sensor package, calculates a productivity measure, applies the productivity measure to the stochastic learning algorithm that uses the split data distribution, chooses at least one new set point to issue to the set point sub-controller, drives the actuator package, and optimizes the process.
Claim 29
Dependent Claim
98 words
A process controller system including sub-components comprising: a stochastic learning algorithm that uses a split data distribution; a sensor package; at least one set point sub-controller; an actuator package; a productivity measure comprising a set of process goals and/or process outputs; an interconnecting communication system between sub-components; wherein the controller system acquires information from the sensor package, calculates a productivity measure, applies the productivity measure to the stochastic learning algorithm that uses the split data distribution, chooses at least one new set point to issue to the set point sub-controller, drives the actuator package, and optimizes the process.
Claim 30
Dependent Claim
18 words
The process controller system according to claim 23 , wherein the stochastic learning algorithm comprises an expert system.
References Cited (Backward Citations)
US3986932A
Cited
US5672194A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2007-09-14 | 2008-03-15 | 2008-09-14 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2011-09-14 | 2012-03-15 | 2012-09-14 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2015-09-14 | 2016-03-15 | 2016-09-14 | $7700.00 | Window Closed / Lapsed |