CN114237043B - Gas turbine equipment transfer function closed-loop identification method based on deep learning - Google Patents
Gas turbine equipment transfer function closed-loop identification method based on deep learning Download PDFInfo
- Publication number
- CN114237043B CN114237043B CN202111429898.0A CN202111429898A CN114237043B CN 114237043 B CN114237043 B CN 114237043B CN 202111429898 A CN202111429898 A CN 202111429898A CN 114237043 B CN114237043 B CN 114237043B
- Authority
- CN
- China
- Prior art keywords
- transfer function
- loop
- model
- parameter
- closed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000006870 function Effects 0.000 title claims abstract description 39
- 238000012546 transfer Methods 0.000 title claims abstract description 33
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000013135 deep learning Methods 0.000 title claims abstract description 12
- 230000004044 response Effects 0.000 claims abstract description 30
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 6
- 238000004088 simulation Methods 0.000 claims abstract description 4
- 238000002474 experimental method Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 6
- 230000003321 amplification Effects 0.000 claims description 5
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 5
- 238000011176 pooling Methods 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 2
- 239000007789 gas Substances 0.000 description 16
- 238000012545 processing Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 239000003345 natural gas Substances 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a gas turbine equipment transfer function closed-loop identification method based on deep learning, which comprises the following steps: (1) The key equipment of the gas turbine is constructed into a first-order inertial pure hysteresis FOPDT model, a PI controller is adopted as a feedback controller, and an open-loop transfer function model is established; (2) Regularizing model parameters, and carrying out perturbation on each obtained dimensionless parameter; (3) Performing closed loop step response simulation by using a perturbation model generated by random numbers; (4) And constructing a deep learning convolutional neural network by taking the nominal time parameter and the closed loop step response curve as input parameters and taking the dimensionless model parameter as output parameters, and training the convolutional neural network by utilizing a numerical sequence. The closed-loop identification method is not limited by field operation conditions, and an open-loop transfer function is obtained under the closed-loop condition; the training sample number is large enough to fully cover the uncertain change range of the actual object, and the recognition accuracy is high.
Description
Technical Field
The invention relates to the field of automatic control of thermal engineering, in particular to a gas turbine equipment transfer function closed-loop identification method based on deep learning.
Background
At present, gas turbine power generation using natural gas as fuel is an important technology for power supply, and with the wide use of a gas-steam combined cycle generator set, the operation characteristics and control of the gas turbine are increasingly researched. The gas turbine unit cannot perform open-loop disturbance experiments under the limitation of on-site operation conditions, and in order to obtain the dynamic operation characteristics of the gas turbine, the development of closed-loop identification becomes an important technical means for designing a control system.
System identification is a very important part of modern control theory, and is to determine a system equivalent to an identified system based on input and output observations. Classical system identification methods include step response methods, impulse response methods, least squares methods and derivatives thereof, but these methods are difficult to handle time lags and cannot determine process identification of the time lags. Therefore, to solve the deficiencies of the classical identification method, applying new technology to solve the problem is the key point of current research.
Disclosure of Invention
The invention aims to: in view of the above problems, the present invention aims to provide a gas turbine equipment transfer function closed-loop identification method based on deep learning.
The technical scheme is as follows: the invention discloses a gas turbine equipment transfer function closed-loop identification method based on deep learning, which comprises the following steps:
(1) The key equipment of the gas turbine is constructed into a first-order inertial pure hysteresis FOPDT model, a PI controller is adopted as a feedback controller, and an open-loop transfer function model is established;
(2) Regularizing model parameters, and carrying out perturbation on each obtained dimensionless parameter;
(3) Performing closed-loop step response simulation by using a perturbation model generated by random numbers, and recording a numerical sequence of closed-loop step response corresponding to each group of dimensionless parameters;
(4) And constructing a deep learning convolutional neural network by taking the nominal time parameter and the closed loop step response curve as input parameters and taking the dimensionless model parameter as output parameters, and training the convolutional neural network by utilizing a numerical sequence to obtain a neural network model of an identification model.
Further, the first-order inertial pure hysteresis FOPDT model transfer function expression in step (1) is:
Wherein s is complex frequency, K is real object amplification factor, T is real object time constant, L is real object delay time, and parameter K, T, L is parameter to be identified;
Taking the PI controller as a feedback controller, and the corresponding transfer function expression is as follows:
Wherein, K p is a proportionality coefficient, T i is an integration coefficient, and the value of K p、Ti is adjusted according to the object characteristics and the control target;
thus, the open loop transfer function of the gas turbine control system is expressed as:
Setting a parameter K p、Ti in the PI controller according to an internal setting method, wherein the set expression is as follows:
Ti＝T0
where K 0 is the object model magnification factor, T 0 is the object model time constant, and L 0 is the object model delay time.
Further, the parameter regularization in step (2) is defined as follows:
Beta, gamma and alpha are dimensionless parameters with nominal values of 1, theta 0 is a nominal time parameter, and theta is a time parameter to be solved;
Bringing the above into the gas turbine control system, the open loop transfer function is rewritten as:
Further, the perturbation in the step (2) refers to random perturbation of + -30% on dimensionless numbers beta, gamma and alpha under the condition of a known nominal time parameter theta 0, and each dimensionless parameter is uniformly and randomly valued within the range of [0.7,1.3 ].
And (3) carrying out a step response experiment on the closed loop system on the basis of each group of random number pairs, carrying out the closed loop step response experiment according to a corresponding group of beta, gamma and alpha, taking points on the curve by an equal time step delta g after the response curve tends to be stable, and recording each point, a nominal time parameter theta 0 and dimensionless numbers beta, gamma and alpha.
Further, the convolutional neural network comprises 1 input layer, 4 convolutional layers, 4 batch normalization layers, 4 activation function layers, 2 average pooling layers, one discard layer, 1 full connection layer and 1 regression layer.
The beneficial effects are that: compared with the prior art, the invention has the remarkable advantages that: the closed-loop identification method is not limited by field operation conditions, and an open-loop transfer function can be obtained under the closed-loop condition; the training sample number is large enough to fully cover the uncertain change range of the actual object, and the recognition accuracy is high; the problem that the FOPDT of the first-order time lag model is difficult to identify is solved; the parameter identification process is simple and easy to operate.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of an open loop transfer function model;
FIG. 3 is a model diagram of a convolutional neural network;
FIG. 4 is an absolute percentage error plot of the parameters, wherein (a) is an absolute percentage error plot of parameter β; (b) is an absolute percent error profile of parameter gamma; (c) is an absolute percent error profile for parameter α;
fig. 5 is a closed loop step response graph.
Detailed Description
The embodiment of the method for identifying the transfer function closed loop of the gas turbine equipment based on deep learning is shown in fig. 1, and the flow chart comprises the following steps:
(1) The key equipment of the gas turbine is constructed into a first-order inertial pure hysteresis FOPDT model, the key equipment such as a valve, a pipeline, a rotating speed, a sensing and the like, a PI controller is adopted as a feedback controller, an open-loop transfer function model is built, and the model structure is shown in figure 2;
The transfer function expression of the first-order inertial pure hysteresis FOPDT model is as follows:
Wherein s is complex frequency, K is real object amplification factor, T is real object time constant, L is real object delay time, and parameter K, T, L is parameter to be identified;
Taking the PI controller as a feedback controller, and the corresponding transfer function expression is as follows:
Wherein, K p is a proportionality coefficient, T i is an integration coefficient, and the value of K p、Ti is adjusted according to the object characteristics and the control target;
thus, the open loop transfer function of the gas turbine control system is expressed as:
Setting a parameter K p、Ti in the PI controller according to an internal setting method, wherein the set expression is as follows:
Ti＝T0
Wherein K 0 is an object model amplification factor, T 0 is an object model time constant, L 0 is an object model delay time, and K 0,T0,L0 is obtained from model priori knowledge.
(2) Model parameters are regularized, and the definition formula is as follows: ;
Beta, gamma and alpha are dimensionless parameters with nominal values of 1, theta 0 is a nominal time parameter, and theta is a time parameter to be solved;
Bringing the above into the gas turbine control system, the open loop transfer function is rewritten as:
Under the condition of known nominal time parameter theta 0, carrying out random perturbation of +/-30% on dimensionless parameters beta, gamma and alpha, wherein the values of beta, gamma and alpha are uniformly and randomly taken within the range of [0.7,1.3], and the nominal time parameter is theta 0 epsilon [0.5, 10]. The formulas corresponding to each set of dimensionless numbers in the above formula represent a class of similar objects, rather than a particular object.
(3) And carrying out closed loop step response simulation by utilizing a perturbation model generated by random numbers, and recording a numerical sequence of closed loop step response corresponding to each group of dimensionless parameters.
And carrying out a step response experiment on the closed loop system on the basis of each group of random number pairs, carrying out the closed loop step response experiment according to a corresponding group of beta, gamma and alpha, taking points on the curve by an equal time step delta g after the response curve tends to be stable, and recording each point, a nominal time parameter theta 0 and dimensionless numbers beta, gamma and alpha.
(4) And constructing a deep learning convolutional neural network by taking the nominal time parameter and the closed loop step response curve as input parameters and taking the dimensionless model parameter as output parameters, and training the convolutional neural network by utilizing a numerical sequence to obtain a neural network model of an identification model.
The convolutional neural network comprises an input layer, a first convolutional layer, a first batch processing layer, a first activation function layer, a first average pooling layer, a second convolutional layer, a second batch processing layer, a second activation function layer, a second average pooling layer, a third convolutional layer, a third batch processing layer, a third activation function layer, a fourth convolutional layer, a fourth batch processing layer, a fourth activation function layer, a discarding layer, a full connection layer and a regression layer which are sequentially arranged, and the specific structure is shown in figure 3.
The identification method is evaluated through parameter identification performance indexes, absolute percentage errors are used as performance indexes, and the expression is as follows:
Wherein k epsilon [1, n ] and k epsilon Z, n are the group numbers of the test set, m k represents the absolute percentage error of the parameters of the kth group, t k is the parameter true value of the kth group, and p k is the parameter identification value of the kth group.
The total data of the test set is 1000 groups, namely n=1000, the absolute percentage error distribution of each parameter is counted, and the distribution results obtained by beta, gamma and alpha are respectively shown in fig. 4 (a), (b) and (c), wherein the absolute percentage error of the parameters in the 1000 groups is more than 850 groups in the (0, 3%) range.
As shown in fig. 5, on the premise that the nominal time parameter θ 0 = 3.7435, when the dimensionless parameter is true value: When the system is subjected to closed loop step response experiments, a corresponding step response curve is obtained, the response curve and a nominal time parameter theta 0 are used as input parameters of a neural network, and the dimensionless parameters are identified through the convolutional neural network: And the parameters K 0,T0,L0 are obtained by model priori knowledge, so that K, T and L can be obtained, namely parameters to be identified in the open-loop transfer function of important equipment of the gas turbine are obtained according to the closed-loop step response curve and the nominal time parameter of the system, and the open-loop transfer function is finally obtained. And then, performing a closed-loop step response experiment on an open-loop transfer function corresponding to the identification parameter to obtain an identification step response curve of fig. 5, wherein the accuracy of the identification of the neural network established by the method of the embodiment is high according to the coincidence degree of the two response curves of fig. 5.
Claims (4)
1. The gas turbine equipment transfer function closed-loop identification method based on deep learning is characterized by comprising the following steps of:
(1) The key equipment of the gas turbine is constructed into a first-order inertial pure hysteresis FOPDT model, a PI controller is adopted as a feedback controller, and an open-loop transfer function model is established;
(2) Regularizing model parameters, and carrying out perturbation on each obtained dimensionless parameter;
(3) Performing closed-loop step response simulation by using a perturbation model generated by random numbers, and recording a numerical sequence of closed-loop step response corresponding to each group of dimensionless parameters;
(4) The method comprises the steps of constructing a deep learning convolutional neural network by taking a nominal time parameter and a closed loop step response curve as input parameters and taking dimensionless model parameters as output parameters, and training the convolutional neural network by utilizing a numerical sequence to obtain a neural network model of an identification model;
The transfer function expression of the first-order inertial pure hysteresis FOPDT model in the step (1) is as follows:
Wherein s is complex frequency, K is real object amplification factor, T is real object time constant, L is real object delay time, and parameter K, T, L is parameter to be identified;
Taking the PI controller as a feedback controller, and the corresponding transfer function expression is as follows:
Wherein, K p is a proportionality coefficient, T i is an integration coefficient, and the value of K p、Ti is adjusted according to the object characteristics and the control target;
thus, the open loop transfer function of the gas turbine control system is expressed as:
Setting a parameter K p、Ti in the PI controller according to an internal setting method, wherein the set expression is as follows:
Ti＝T0
Wherein K 0 is an object model amplification factor, T 0 is an object model time constant, and L 0 is an object model delay time;
The parameter regularization in step (2) is defined as follows:
Beta, gamma and alpha are dimensionless parameters with nominal values of 1, theta 0 is a nominal time parameter, and theta is a time parameter to be solved;
Bringing the above into the gas turbine control system, the open loop transfer function is rewritten as:
2. the method for closed-loop identification of a transfer function of a gas turbine plant according to claim 1, wherein the perturbation in the step (2) means that the dimensionless numbers β, γ, α are randomly perturbed by ±30% under the condition of a known nominal time parameter θ 0, and each dimensionless parameter is uniformly and randomly valued within the range of [0.7,1.3 ].
3. The method for identifying the closed loop of the transfer function of the gas turbine equipment according to claim 2, wherein step (3) carries out a step response experiment on the closed loop system on the basis of each group of random number pairs, carries out the closed loop step response experiment according to a corresponding group of beta, gamma and alpha, takes a point on the curve by an equal time step delta g after the response curve tends to be stable, and records each point, a nominal time parameter theta 0 and dimensionless numbers beta, gamma and alpha.
4. The gas turbine plant transfer function closed loop identification method of claim 1, wherein the convolutional neural network comprises an input layer, four convolutional layers, four batch normalization layers, four activation function layers, two average pooling layers, a discard layer, a full connection layer, and a regression layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111429898.0A CN114237043B (en) | 2021-11-29 | 2021-11-29 | Gas turbine equipment transfer function closed-loop identification method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111429898.0A CN114237043B (en) | 2021-11-29 | 2021-11-29 | Gas turbine equipment transfer function closed-loop identification method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114237043A CN114237043A (en) | 2022-03-25 |
| CN114237043B true CN114237043B (en) | 2024-04-16 |
Family
ID=80751805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111429898.0A Active CN114237043B (en) | 2021-11-29 | 2021-11-29 | Gas turbine equipment transfer function closed-loop identification method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114237043B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115291509B (en) * | 2022-06-24 | 2024-06-18 | 东南大学溧阳研究院 | Online identification method for dynamic parameters of gas turbine based on hybrid deep learning model |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH04346102A (en) * | 1991-05-23 | 1992-12-02 | Hitachi Ltd | Pid parameter automatic tuning method |
| CN1487268A (en) * | 2003-08-12 | 2004-04-07 | 上海交通大学 | Multi-varible system structural colsed-loop distinguishing method based on step response test |
| CN108333933A (en) * | 2018-01-26 | 2018-07-27 | 东南大学 | A kind of single order pure delay system closed-loop identification method |
| CN110174841A (en) * | 2019-04-30 | 2019-08-27 | 杭州意能电力技术有限公司 | The parallel control method of industrial process large time delay inertia system based on parameter optimization |
| CN110222711A (en) * | 2019-04-30 | 2019-09-10 | 杭州意能电力技术有限公司 | A kind of multistage inertia system Open-loop Identification method of industrial process based on deep learning |
| CN110244557A (en) * | 2019-04-30 | 2019-09-17 | 国网浙江省电力有限公司电力科学研究院 | A kind of closed loop modeling method of the multistage inertia closed-loop system of industrial process |
-
2021
- 2021-11-29 CN CN202111429898.0A patent/CN114237043B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH04346102A (en) * | 1991-05-23 | 1992-12-02 | Hitachi Ltd | Pid parameter automatic tuning method |
| CN1487268A (en) * | 2003-08-12 | 2004-04-07 | 上海交通大学 | Multi-varible system structural colsed-loop distinguishing method based on step response test |
| CN108333933A (en) * | 2018-01-26 | 2018-07-27 | 东南大学 | A kind of single order pure delay system closed-loop identification method |
| CN110174841A (en) * | 2019-04-30 | 2019-08-27 | 杭州意能电力技术有限公司 | The parallel control method of industrial process large time delay inertia system based on parameter optimization |
| CN110222711A (en) * | 2019-04-30 | 2019-09-10 | 杭州意能电力技术有限公司 | A kind of multistage inertia system Open-loop Identification method of industrial process based on deep learning |
| CN110244557A (en) * | 2019-04-30 | 2019-09-17 | 国网浙江省电力有限公司电力科学研究院 | A kind of closed loop modeling method of the multistage inertia closed-loop system of industrial process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114237043A (en) | 2022-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220309122A1 (en) | Iterative algorithm for aero-engine model based on hybrid adaptive differential evolution | |
| WO2019144337A1 (en) | Deep-learning algorithm-based self-adaptive correction method for full-envelope model of aero-engine | |
| CN114237043B (en) | Gas turbine equipment transfer function closed-loop identification method based on deep learning | |
| CN111042928A (en) | Variable cycle engine intelligent control method based on dynamic neural network | |
| CN107272412A (en) | A kind of identifying approach of intermittent wind tunnel flow field control | |
| CN114626573A (en) | Load prediction method for optimizing extreme learning machine based on improved multivariate universe algorithm | |
| CN111079920B (en) | Prediction method for flow non-uniformity coefficient of turbine gas collection cavity outlet | |
| CN115494892A (en) | Decoupling control method for air inlet environment simulation system of high-altitude simulation test bed | |
| CN113641101B (en) | Multi-channel control parameter optimizing algorithm based on numerical simulation | |
| Xu et al. | Short-term wind speed prediction based on GRU | |
| Ma et al. | Prediction of thermal system parameters based on PSO-ELM hybrid algorithm | |
| CN117170236A (en) | Industrial speed regulation system actuator fault intelligent self-healing method based on inverse reinforcement learning | |
| Li et al. | The application of immune genetic algorithm in PID parameter optimization for level control system | |
| CN110985216A (en) | Intelligent multivariable control method for aero-engine with online correction | |
| CN105117530B (en) | Method for identifying parameters of steam turbine and speed regulating system thereof by combination of thickness and thickness regulation | |
| CN114372494A (en) | Air path fault diagnosis method for aviation turboprop engine based on attention mechanism | |
| Zhang et al. | Wind Speed Prediction with LSTM and BPNN Method | |
| Kusumoputro et al. | Quality classification of green pellet nuclear fuels using radial basis function neural networks | |
| Zhang et al. | Optimization of aero-engine H-infinity robust controller based on quantum genetic algorithm | |
| Xu et al. | Design of PID Control based on BP Neural Network for Double-holding Process System | |
| Tang et al. | Computer Prediction Model of Heat Consumption in Thermal System of Coal-Fired Power Station Based on Big Data Analysis and Information Sorting | |
| CN116937540B (en) | Multi-ring medium voltage distribution network line parameter identification method based on non-phase measurement | |
| Wang et al. | Gain scheduling controller of the aero-engine based on LPV model | |
| Ming et al. | Design and Simulation of Adaptive Internal Model Control Algorithm | |
| Chen et al. | Intelligent Modification Method for Individual Difference of Aero-engine based on CNN-GAN |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |