CN107831665A - Absorbing natural gas tower sweetening process control method based on UKF and ADDHP - Google Patents
Absorbing natural gas tower sweetening process control method based on UKF and ADDHP Download PDFInfo
- Publication number
- CN107831665A CN107831665A CN201711115683.5A CN201711115683A CN107831665A CN 107831665 A CN107831665 A CN 107831665A CN 201711115683 A CN201711115683 A CN 201711115683A CN 107831665 A CN107831665 A CN 107831665A
- Authority
- CN
- China
- Prior art keywords
- mrow
- sweetening process
- control
- natural gas
- ukf
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 76
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 title claims abstract description 72
- 239000003345 natural gas Substances 0.000 title claims abstract description 36
- 238000004886 process control Methods 0.000 title claims abstract description 22
- 230000008569 process Effects 0.000 claims abstract description 45
- 238000010521 absorption reaction Methods 0.000 claims abstract description 27
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 239000007789 gas Substances 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- XTHFKEDIFFGKHM-UHFFFAOYSA-N Dimethoxyethane Chemical compound COCCOC XTHFKEDIFFGKHM-UHFFFAOYSA-N 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 claims description 7
- 238000000746 purification Methods 0.000 claims description 6
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 claims description 3
- 239000002253 acid Substances 0.000 claims description 2
- 229910052717 sulfur Inorganic materials 0.000 claims description 2
- 239000011593 sulfur Substances 0.000 claims description 2
- 230000004044 response Effects 0.000 abstract description 3
- 230000008878 coupling Effects 0.000 abstract description 2
- 238000010168 coupling process Methods 0.000 abstract description 2
- 238000005859 coupling reaction Methods 0.000 abstract description 2
- 238000005457 optimization Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 18
- 239000011159 matrix material Substances 0.000 description 18
- 238000010586 diagram Methods 0.000 description 12
- 210000002569 neuron Anatomy 0.000 description 6
- RWSOTUBLDIXVET-UHFFFAOYSA-N Dihydrogen sulfide Chemical compound S RWSOTUBLDIXVET-UHFFFAOYSA-N 0.000 description 4
- 230000009897 systematic effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000006477 desulfuration reaction Methods 0.000 description 2
- 230000023556 desulfurization Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000002994 raw material Substances 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010068052 Mosaicism Diseases 0.000 description 1
- 241000255964 Pieridae Species 0.000 description 1
- 239000005864 Sulphur Substances 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000001627 detrimental effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005530 etching Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 239000003317 industrial substance Substances 0.000 description 1
- 210000002364 input neuron Anatomy 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- CRVGTESFCCXCTH-UHFFFAOYSA-N methyl diethanolamine Chemical compound OCCN(C)CCO CRVGTESFCCXCTH-UHFFFAOYSA-N 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 210000003765 sex chromosome Anatomy 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization 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)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
- Treating Waste Gases (AREA)
- Gas Separation By Absorption (AREA)
Abstract
The invention provides a kind of absorbing natural gas tower sweetening process control method based on UKF and ADDHP.Absorbing natural gas tower sweetening process is modeled using BP neural network and sweetening process is carried out as controlled device using the model and controls emulation experiment, optimization weights are constantly updated according to control error and performance index function, until obtaining optimum control signal, the optimum control of sweetening process is realized.The features such as absorbing natural gas tower sweetening process is complicated, and performance is uncertain, non-linear, strong coupling, dynamic, it is difficult to establish accurate mathematical modeling, control difficulty is larger.It is low for current sweetening process control method control accuracy, the problems such as time lag is big, unstable proposes a kind of absorbing natural gas tower sweetening process control method based on UKF and ADDHP, it not only ensure that the stability and control accuracy of control system, the response time is also reduced, is truly realized the real-time accurate control of absorption tower sweetening process.
Description
Technical field
The present invention relates to absorbing natural gas tower sweetening process control technology, and in particular to one kind is based on Unscented kalman filtering
(UKF) with performing the absorbing natural gas tower sweetening process control method for relying on double heuristic dynamic programmings (ADDHP) and combining.
Background technology
Natural gas is easy to use and possess higher comprehensive warp as a kind of high-quality, cleaning the energy and industrial chemicals
Ji benefit.China possesses abundant natural gas resource, but contains a large amount of element sulphurs in about 30% or so natural gas, wherein
H2Gas reserves of the S contents more than 1% accounts for the 1/4 of gross reserves.H2S presence not only results in the burn into of equipment and pipeline
It is detrimental to health, its combustion product also pollutes the environment.Therefore, during selexol process, H2The control of S contents seems outstanding
To be important.
Selexol process absorption tower is the important component of purifying device for natural gas, directly affects natural gas purification effect
Fruit.Natural gas _ raw material gas is fully contacted and reacted with methyl diethanolamine in tower (MDEA) solution into absorption tower, so as to reach
To the purpose of desulfurization, physical-chemical reaction and phase reaction occur simultaneously for whole process, are related to material conversion and energy transmission, by
The features such as various uncertain factors have a great influence, and performance is uncertain, non-linear, strong coupling, dynamic, it is difficult to establish accurate
Mathematical modeling, so as to bring extreme difficulties to the control of absorption tower sweetening process.
Existing control technology is mostly PID unity loop controls or simple serials control, and control system automaticity is not high
And excessive dependence expertise adjustment control parameter, there is larger hysteresis quality, control accuracy is relatively low, the stabilization of control system
Property be also difficult to ensure that, it is difficult to reach accurate control in real time.
The content of the invention
The application is by providing a kind of absorbing natural gas tower sweetening process control method based on UKF and ADDHP, to solve
The problems such as control accuracy present in absorption tower sweetening process control technology is low at present, and time lag is big, and control system is unstable, ensure
Selexol process effect.
In order to solve the above technical problems, the application is achieved using following technical scheme:
A kind of absorbing natural gas tower sweetening process control method based on UKF and ADDHP, it is characterised in that including following step
Suddenly:
Step 1:By analyzing absorption tower sulfur removal technology process, it is determined that the principal element for influenceing selexol process effect is acid
Property natural gas processing amount and alkanolamine solution internal circulating load, represented respectively with u1 and u2, thus form control variable u=[u1, u2];
Step 2:Determine that sweetening process mode input sample data exports sample data, established using BP neural network natural
Aspiration tower sweetening process model;
Step 3:Set preferable control targe valueUpdate in ADDHP control methods and comment with UKF algorithms
Valency network and perform network weight, and respectively by perform network and evaluation network obtain control signal u (k)=[u1, u2] and
Local derviation of the performance index function to system modeEstablish UKF-ADDHP absorbing natural gas tower sweetening process
Control method;
Step 4:By step 3 gained control signal u (k)=[u1, u2] and current time system mode x (k)=[x1, x2]
As absorption tower sweetening process mode input, so as to obtain system output x (k+1).
Step 5:Control error E (k) is calculated, if being less than anticipation error, terminates training, otherwise return to step 3.
As further explanation, the step 3 specifically follows the steps below:
Step 3-1:According to control error E (k), using UKF algorithms more New Appraisement network and network weight is performed;
Step 3-2:Calculate control signal u (k);
Step 3-3:Calculation Estimation network output λ (k+1).
As further explanation, used in step 3-1 UKF algorithms renewal weights process for:
(1) systematic parameter is initialized;
(2) Sigma dotted states vector is calculated;
(3):Carry out system mode one-step prediction and covariance matrix;
(4):Computing system is observed and covariance matrix;
(5):Calculate kalman gain;
(6):Update system state estimation matrix and covariance matrix.
As further explanation, in step 5, error E (k) calculation formula is controlled to be:
In formula, function U (k) is utility function.
Compared with prior art, the technical scheme that the application provides, the technique effect or advantage having are:Taken off on absorption tower
In sulphur process control, this method control accuracy is high, fast convergence rate, it is possible to increase stability of control system and control accuracy, drop
The low control system response time, ensure selexol process effect.
Brief description of the drawings
Fig. 1 principle of the invention block diagrams;
Fig. 2 absorption towers sweetening process model schematic;
Fig. 3-6 is absorbing natural gas tower sweetening process model test results schematic diagram;
Fig. 3 H2S content prediction schematic diagrames;
Fig. 4 H2S content prediction relative error schematic diagrames;
Fig. 5 CO2Content prediction schematic diagram;
Fig. 6 CO2Content prediction relative error schematic diagram;
Fig. 7 UKF-ADDHP control structure schematic diagrames;
Fig. 8-11 is absorbing natural gas tower sweetening process controlling curve schematic diagram;
H in Fig. 8 natural gas purification gas2S content controlling curve schematic diagrames;
CO in Fig. 9 natural gas purification gas2Content controlling curve schematic diagram;
Figure 10 natural gas _ raw material gas treating capacity controlling curve schematic diagrames;
Figure 11 alkanolamine solution internal circulating load controlling curve schematic diagrames;
Embodiment
The application provides a kind of absorbing natural gas tower sweetening process control method based on UKF and ADDHP, inventive principle frame
Figure is as shown in Figure 1.The technical scheme provided with reference to prior art means, the application, the technique effect or advantage having are:The party
Method is controlled using intelligent algorithm for absorption tower sweetening process, has higher control accuracy, can reduce control system response
Time, can adjust automatically control parameter in real time, improve stability of control system, be really achieved the purpose of real-time control.
In order to be better understood from above-mentioned technical proposal, below in conjunction with Figure of description 2-11 and specific embodiment party
Formula, above-mentioned technical proposal is described in detail.
Initially enter step 1:Choose sour natural gas treating capacity and absorb the alkanolamine solution internal circulating load used in sour gas
Two parameters form control variable u=[u1, u2].
Step 2:With BP neural network, respectively with input1~inputnAnd x1~xnCarried out as input and output sample
Training, examine, so as to establish absorption tower sweetening process model.Wherein, input=[x1, x2, u1, u2], x=[x1, x2], n tables
Show sample size, u1, u2 represent raw natural gas treating capacity and alkanolamine solution internal circulating load in the unit interval, x1, x2 difference respectively
Represent H in natural gas purification gas2S contents (mg/m3) and CO2Content (%).
In the present embodiment, absorption tower sweetening process model as shown in Figure 2 is established, input layer number is 4, defeated
Go out layer neuron number for 2;Rule of thumb, hidden layer node selection be(x is input layer, and y is output
Node layer, a=1,2,9), it is 10 by testing selection modeling measuring accuracy highest hidden layer node successively;It is implicit
Layer transmission function is tansig functions, and output layer transmission function is purelin functions;Anticipation error minimum value is 0.0001, is repaiied
The learning efficiency of positive weights is 0.05.Modeling sample data are puguang gas field actual production data, 500 groups altogether, are randomly selected
80% sample data is used as model training, and the sample of residue 20% is used as model measurement.
If absorption tower sweetening process mode input is P, input neuron number is r, and hidden layer neuron number is s1, right
The activation primitive answered is h1, and hidden layer output is a1;Output layer neuron number is s2, and corresponding activation primitive is h2, output
For a2, target vector T.
Absorption tower sweetening process model is established in step 2 to specifically comprise the following steps:
Step 2-1:Initialization, if iterations g initial values are 0, while W1 is assigned to, W2, B1, mono- (0,1) section of B2
Random value;
Step 2-2:Stochastic inputs sample Pj;
Step 2-3:To input sample Pj, the input and output of every layer of neuron of forward calculation BP neural network;
The output of i-th of neuron of hidden layer is:
The output of s-th of neuron of output layer is:
Step 2-4:According to desired output T and reality output a2 (g), calculation error E (g);
Defining error function is:
Step 2-5:Whether error in judgement E (g), which meets, requires, is such as unsatisfactory for, then into step 2-6, such as meets, then enter
Step 2-8;
Step 2-6:Judge whether iterations g+1 is more than maximum iteration, it is such as larger than, then no into step 2-8
Then, into step 2-7;
Step 2-7:Modified weight amount Δ W is calculated, and corrects weights.
1. output layer weights change
Weights to being input to k-th of output from i-th, have:
Wherein, δki=(tk-a2k) h2 '=ekH2 ', ek=tk-a2k。
2. hidden layer weights change
Weights to being input to i-th of output from j-th, have:
Wherein,
It can similarly obtain:
Δb1i=η δij
In formula, η is learning efficiency;G=g+1 is made, jumps to step 3;
Step 2-8:Judge whether to complete all training samples, if it is, completing modeling, otherwise, continue to jump to
Step 2-2;
By said process, BP neural network prediction effect such as Fig. 3 is can obtain, shown in 5, corresponding prediction error such as Fig. 4,
Shown in 6.By analysis chart 3-6, absorption tower sweetening process model is established in BP neural network training has higher precision, energy
The output of enough accurate forecasting systems, lays the foundation for the research of selexol process course control method for use.
Step 3:Set preferable control targe valueUpdate in ADDHP control methods and comment with UKF algorithms
Valency network and perform network weight, and respectively by perform network and evaluation network obtain control signal u (k)=[u1, u2] and
Local derviation of the performance index function to system modeEstablish UKF-ADDHP absorbing natural gas tower sweetening process
Control method, its control structure are as shown in Figure 7:For Action-UKF to perform network, input and output are respectively system mode x (k)
With control signal u (k);Controlled Object are prototype network, are inputted as system mode x (k) and control signal u (k),
Export as system subsequent time state x (k+1);Critic-UKF is evaluation network, is inputted as x (k+1) and u (k+1), exports and is
Local derviation of the performance index function to system modePerform network and evaluate network training respectively with
Control error E (k) and functionTarget is minimised as, dotted line represents network weight adjusts path.
In the present embodiment, the error calculation formula is controlled to be:
In formula, function U (k) is utility function.
UKF algorithms sampled point and corresponding weight value calculation formula are:
(1) 2n+1 sigma point is calculated
(2) the corresponding weights of sampled point are calculated
Parameter lambda=α2(n+ τ)-n is a scaling parameter, and α selection controls the distribution of sampled point, and τ is
Parameter to be selected, it is usual it is ensured that matrix (n+ λ) P is positive semidefinite matrix although its value is without specific boundary.Parameter σ to be selected
>=0 is a non-negative weight coefficient, and it can merge the moment of higher order term in equation, thus can be the influence bag of higher order term
Including including, reduce error.I-th row of representing matrix root mean square or the i-th row,Respectively square
Battle array, the weighted value of covariance.
As further explanation, the step 3 specifically follows the steps below:
Step 3-1:Right value update
(1) systematic parameter is initialized;
(2) Sigma dotted states vector is calculated;
(3):Carry out system mode one-step prediction and covariance matrix;
(4):Computing system is observed and covariance matrix;
(5):Calculate kalman gain;
(6):Renewal system state estimation matrix and covariance matrix are:
In formula,For the system state estimation matrix at k moment,For kalman gain matrix, G (k+1 | k)
For the systematic observation matrix at k moment,For the systematic observation prediction matrix at k moment;
In formula,For k moment system estimation matrix covariance matrixs,For k when etching system
Observing matrix covariance matrix;
Step 3-2:Calculate control signal u (k)
Network hidden layer weights W will be performeda1With output layer weights Wa2Two groups of parameters are as the state vector X for performing networka:
XaConstantly update and optimize by step 3-1, finally give one group of best initial weights, it is defeated that network hidden layer is performed after renewal
Go out for:
La=x (k) * Wa1
Wherein, x (k) is k moment system modes, i.e. control output, and the output of network output layer is performed after renewal and is:
U (k)=Qa*Wa2
In formula, Qa=(1-exp (- La))./(1+exp(-La)), QaThe function of expression is sigmod functions, as nerve
The threshold function table of network, u (k) are the control signal for performing network output;
Step 3-3:Calculation Estimation network output λ (k+1)
Will evaluation network hidden layer weights Wc1With output layer weights Wc2State vector X of two groups of parameters as evaluation networkc:
XcConstantly update and optimize by step 3-1, finally give one group of best initial weights, it is defeated that network hidden layer is evaluated after renewal
Go out for:
Lc=input*WC1
Wherein, input=[x (k+1), u (k+1)], the output of network output layer is evaluated after renewal is:
Wherein, J (k+1)=Qc*WC2, Qc=(1-exp (- Lc))./(1+exp(-Lc)), QcThe function of expression is sigmod
Function, the threshold function table as neutral net.
Step 4:By step 3 gained control signal u (k)=[u1, u2] and current time system mode x (k)=[x1, x2]
As absorption tower sweetening process mode input, so as to obtain system output x (k+1).
Step 5:Control error E (k) is calculated, if being less than anticipation error, terminates training, otherwise return to step 3.
In the present embodiment, controlled variable desired value x1=1.53 and x2=1.6 are set, by said process, can obtain
H in natural gas purification gas2S and CO2Content controlling curve figure such as Fig. 8, shown in 9, raw natural gas treating capacity and alkanolamine solution circulation
Controlling curve such as Figure 10 is measured, shown in 11.By analyzing 8-11, control of the UKF-ADDHP methods to absorption tower sweetening process
Meet actual process requirement, and there is fast convergence rate, the features such as control accuracy is high.
The invention provides a kind of absorption tower sweetening process control method based on UKF and ADDHP.First, BP god is utilized
Through network training absorption tower desulfurization actual production data, absorption tower sweetening process model is established, so as to get around sweetening process machine
Details sex chromosome mosaicism in reason, solves the problems, such as to model caused by sweetening process complexity difficult, is selexol process process
The research of control method lays the foundation.Then, the control that forecasting system is used for using the model of foundation as controlled device exports, and uses
ADDHP methods are controlled to absorption tower sweetening process and use UKF algorithms renewal optimization ADDHP evaluation networks and perform network
Weights, establish the absorption tower sweetening process control method based on UKF-ADDHP.This method has been broken away from for a long time to expertise
Depend on unduly, solve that control accuracy existing for existing absorption tower sweetening process control technology is low, and time lag is big, control system is not
The problems such as stable, the purpose that absorption tower sweetening process accurately controls in real time is really achieved, has also been asked to solve similar industrial control
Topic provides a kind of new thinking, embodies the power of intelligent algorithm in the industry.
It should be pointed out that it is limitation of the present invention that described above, which is not, the present invention is also not limited to the example above,
What those skilled in the art were made in the essential scope of the present invention changes, is modified, adds or replaces, and also should
Belong to protection scope of the present invention.
Claims (4)
1. the absorbing natural gas tower sweetening process control method based on UKF and ADDHP, it is characterised in that comprise the following steps:
Step 1:By analyzing absorption tower sulfur removal technology process, it is determined that the principal element for influenceing selexol process effect is acid day
Right gas disposal amount and alkanolamine solution internal circulating load, are represented with u1 and u2, thus form control variable u=[u1, u2] respectively;
Step 2:Determine that sweetening process mode input sample data exports sample data, natural aspiration is established using BP neural network
Receive tower sweetening process model;
Step 3:Set control targe valueWith UKF algorithms to the evaluation network in ADDHP control methods and
Perform network weight be updated, and by perform network and evaluation network respectively obtain control signal u (k)=[u1, u2] and
Local derviation of the performance index function to system modeEstablish UKF-ADDHP absorbing natural gas tower sweetening process
Control method;
Step 4:By step 3 gained control signal u (k)=[u1, u2] and current time system mode x (k)=[x1, x2] conduct
Absorption tower sweetening process mode input, so as to obtain system output x (k+1);
Step 5:Control error E (k) is calculated, if being less than anticipation error, terminates training, otherwise return to step 3.
2. the absorbing natural gas tower sweetening process control method according to claim 1 based on UKF and ADDHP, its feature
It is:
When absorption tower sweetening process model is established in step 2, input=[x1, x2, u1, u2] is regard as mode input sample number
According to, while by x=[x1, x2] as model output sample data, x1, x2 represent H in natural gas purification gas respectively2S contents
(mg/m3) and CO2Content (%).
3. the absorbing natural gas tower sweetening process control method according to claim 1 based on UKF and ADDHP, its feature
In step 3, UKF-ADDHP control methods comprise the following steps:
Step 3-1:According to control error E (k), using UKF algorithms more New Appraisement network and network weight is performed;
Step 3-2:Calculate control signal u (k);
Step 3-3:Calculation Estimation network output λ (k+1).
4. the absorbing natural gas tower sweetening process control method according to claim 1 based on UKF and ADDHP, its feature
It is:
Error E (k) calculation formula is controlled to be in step 5:
<mrow>
<mo>|</mo>
<mo>|</mo>
<mi>E</mi>
<mo>|</mo>
<mo>|</mo>
<mo>=</mo>
<munder>
<mo>&Sigma;</mo>
<mi>k</mi>
</munder>
<mi>E</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<munder>
<mo>&Sigma;</mo>
<mi>k</mi>
</munder>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mi>&lambda;</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mfrac>
<mrow>
<mo>&part;</mo>
<mi>U</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mo>&part;</mo>
<mi>x</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mi>&gamma;</mi>
<mfrac>
<mrow>
<mo>&part;</mo>
<mi>J</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>+</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mo>&part;</mo>
<mi>x</mi>
<mrow>
<mo>(</mo>
<mi>k</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&rsqb;</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
In formula, function U (k) is utility function.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711115683.5A CN107831665B (en) | 2017-11-13 | 2017-11-13 | Natural gas absorption tower desulfurization process control method based on UKF and ADDHP |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711115683.5A CN107831665B (en) | 2017-11-13 | 2017-11-13 | Natural gas absorption tower desulfurization process control method based on UKF and ADDHP |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107831665A true CN107831665A (en) | 2018-03-23 |
CN107831665B CN107831665B (en) | 2021-01-29 |
Family
ID=61655167
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711115683.5A Active CN107831665B (en) | 2017-11-13 | 2017-11-13 | Natural gas absorption tower desulfurization process control method based on UKF and ADDHP |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107831665B (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104656441A (en) * | 2014-12-29 | 2015-05-27 | 重庆科技学院 | Natural gas purification process modeling optimization method based on unscented kalman neural network |
CN106777465A (en) * | 2016-11-14 | 2017-05-31 | 重庆科技学院 | High sulfur-containing natural gas purify technique dynamic evolutionary modeling and energy conservation optimizing method |
-
2017
- 2017-11-13 CN CN201711115683.5A patent/CN107831665B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104656441A (en) * | 2014-12-29 | 2015-05-27 | 重庆科技学院 | Natural gas purification process modeling optimization method based on unscented kalman neural network |
CN106777465A (en) * | 2016-11-14 | 2017-05-31 | 重庆科技学院 | High sulfur-containing natural gas purify technique dynamic evolutionary modeling and energy conservation optimizing method |
Non-Patent Citations (5)
Title |
---|
DANIL V.PROKHOROV等: "Adaptive Critic Designs", 《IEEE TRANSACTIONS ON NEURAL NETWORKS》 * |
PAUL H. EATON等: "Neurocontroller Alternatives for "Fuzzy"Ball-and-Beam Systems with Nonuniform Nonlinear Friction", 《IEEE TRANSACTIONS ON NEURAL NETWORKS》 * |
林小峰等: "基于自适应评价设计的一种控制算法", 《PROCEEDINGS OF THE 26TH CHINESE CONTROL CONFERENCE》 * |
罗艳红: "基于神经网络的非线性系统自适应优化控制研究", 《中国博士学位论文全文数据库信息科技辑》 * |
辜小花等: "基于大数据的高含硫天然气脱硫工艺优化", 《集输与加工》 * |
Also Published As
Publication number | Publication date |
---|---|
CN107831665B (en) | 2021-01-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Niu et al. | Model turbine heat rate by fast learning network with tuning based on ameliorated krill herd algorithm | |
CN106227038A (en) | Grain drying tower intelligent control method based on neutral net and fuzzy control | |
CN109143872A (en) | A kind of continuous stirred tank reactor course control method for use based on event triggering GDHP | |
CN107703760A (en) | Absorbing natural gas tower sweetening process control method based on RBF and GDHP | |
CN104765347B (en) | Yield real-time predicting method in a kind of residual oil delayed coking | |
CN104616072B (en) | A kind of method of the raising glutamic acid fermentation production concentration based on range optimization | |
CN103645635A (en) | A ship motion controller based on a simulation annealing-reinforcement learning algorithm | |
CN102393645A (en) | Control method of high-speed electro-hydraulic proportional governing system | |
Wang et al. | Modeling of soft sensor based on DBN-ELM and its application in measurement of nutrient solution composition for soilless culture | |
CN113964884A (en) | Power grid active frequency regulation and control method based on deep reinforcement learning | |
Zhu et al. | Structural safety monitoring of high arch dam using improved ABC-BP model | |
CN107908108A (en) | Absorbing natural gas tower sweetening process control method based on UKF and GDHP | |
CN111756049B (en) | Data-driven reactive power optimization method considering loss of real-time measurement information of power distribution network | |
CN107831666A (en) | Absorbing natural gas tower sweetening process control method based on RBF and ADDHP | |
CN107885084A (en) | Absorbing natural gas tower sweetening process control method based on RBF and ADHDP | |
CN107831665A (en) | Absorbing natural gas tower sweetening process control method based on UKF and ADDHP | |
CN107885083A (en) | Absorbing natural gas tower sweetening process control method based on UKF and ADHDP | |
CN103838205B (en) | BP global optimum propylene polymerization production process optimal soft survey instrument and method | |
CN103761567A (en) | Wavelet neural network weight initialization method based on Bayes estimation | |
CN103838206A (en) | Optimal soft measurement meter and method in optimal BP multi-mode propylene polymerization production process | |
CN115936061A (en) | Thermal power plant flue gas oxygen content soft measurement method and system based on data driving | |
Ao et al. | The application of DQN in thermal process control | |
CN105303056B (en) | A kind of wind speed real-time change speed depicting method | |
CN107944552A (en) | A kind of industrial Internet of Things parameter prediction method based on Elman neutral nets | |
Cui et al. | Study on decoupling control system of temperature and humidity in intelligent plant factory |
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 | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20231115 Address after: 017000 Comprehensive Industrial Park of Etuoqueqian Banner, Ordos, Inner Mongolia Autonomous Region Patentee after: Inner Mongolia Yahai Energy Development Co.,Ltd. Address before: No. 20, East Road, University City, Chongqing, Shapingba District, Chongqing Patentee before: Chongqing University of Science & Technology |