CN109669352B - Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm - Google Patents

Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm Download PDF

Info

Publication number
CN109669352B
CN109669352B CN201710964727.5A CN201710964727A CN109669352B CN 109669352 B CN109669352 B CN 109669352B CN 201710964727 A CN201710964727 A CN 201710964727A CN 109669352 B CN109669352 B CN 109669352B
Authority
CN
China
Prior art keywords
sewage treatment
treatment process
oily sewage
optimization
particle swarm
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
Application number
CN201710964727.5A
Other languages
Chinese (zh)
Other versions
CN109669352A (en
Inventor
宋项宁
郭亚逢
牟桂芹
赵乾斌
隋立华
姚猛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Petroleum and Chemical Corp
Sinopec Safety Engineering Research Institute Co Ltd
Original Assignee
China Petroleum and Chemical Corp
Sinopec Safety Engineering Research Institute Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Petroleum and Chemical Corp, Sinopec Safety Engineering Research Institute Co Ltd filed Critical China Petroleum and Chemical Corp
Priority to CN201710964727.5A priority Critical patent/CN109669352B/en
Publication of CN109669352A publication Critical patent/CN109669352A/en
Application granted granted Critical
Publication of CN109669352B publication Critical patent/CN109669352B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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

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)
  • Activated Sludge Processes (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention relates to an optimal control method for an oily sewage treatment process based on self-adaptive multi-target particle swarm, which mainly solves the problems that the standard particle swarm algorithm in the prior art is easy to generate premature convergence and has low search precision. Firstly, establishing an optimized set value of dissolved oxygen and nitrate nitrogen and a target function of aeration energy consumption and pumping energy consumption in the oily sewage treatment process through a fuzzy neural network; secondly, optimizing an objective function for treating the oily sewage by adopting a self-adaptive multi-objective particle swarm optimization method, and simultaneously obtaining optimized set values of dissolved oxygen and nitrate nitrogen; and finally, the controller is utilized to carry out tracking control on the optimized set values of the dissolved oxygen and the nitrate nitrogen, and the technical scheme of multi-objective optimization control in the oily sewage treatment process is completed, so that the problems are well solved, and the method can be used for optimization control in the oily sewage treatment process.

Description

Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm
Technical Field
The invention relates to an optimal control method for an oily sewage treatment process based on self-adaptive multi-target particle swarm. The invention realizes dissolved oxygen DO and nitrate nitrogen S in the oily sewage treatment process by utilizing an optimization control method based on self-adaptive multi-target particle swarmNOControl of concentration, dissolved oxygen DO and nitrate nitrogen SNOThe concentration of the water is a key control parameter in the oily sewage treatment process, and has important influence on the oily sewage treatment effect, the effluent quality and the energy consumption in the oily sewage treatment process. The optimization control method based on the self-adaptive multi-target particle swarm is applied to the oily sewage treatment processIn the process, dissolved oxygen DO and nitrate nitrogen S are realizedNOThe concentration is optimally controlled, the energy consumption of the oily sewage treatment process is reduced, the investment and the operation cost are saved, the stable and efficient operation of a sewage treatment plant is ensured, and the method belongs to the field of water treatment and also belongs to the field of intelligent control.
Background
With the rapid development of economic society in China, the process of the oil industry is accelerated, the pollution of human activities to water environment is increased continuously, and the oily sewage has extremely serious influence on human beings, animals, plants and even the whole ecological system. Meanwhile, the increase of national economy and the enhancement of public environmental awareness lead the automatic technology of oily sewage treatment to meet the unprecedented development opportunity; how to prevent and treat water pollution and how to effectively treat and reuse oily sewage in time becomes an urgent problem in China; however, the oily sewage treatment process has high electric energy consumption and high operation cost, and the research on the significance of optimizing and controlling the sewage treatment process to realize energy conservation and consumption reduction is significant, and is a necessary development trend of the sewage treatment industry in the future.
The essence of the sewage biochemical treatment process is that the life activity of microorganisms in the sludge is utilized to decompose organic pollutants in the sewage, so that the sewage is purified. In the process of treating the oily sewage, the main control variables are dissolved oxygen DO and nitrate nitrogen SNOAnd (4) concentration. Dissolved oxygen DO and nitrate nitrogen SNOThe change of the concentration can directly influence the nitrification process and the denitrification process, thereby influencing the energy consumption in the oily sewage treatment process. The nitrification reaction is mainly carried out under the aerobic condition, when the dissolved oxygen DO concentration is increased, the concentration of the effluent ammonia nitrogen and the total nitrogen presents a descending trend, when the dissolved oxygen DO concentration is increased to a certain range, the variation amplitude of the effluent ammonia nitrogen begins to be weakened, the total nitrogen is influenced by the nitrate nitrogen, and the total nitrogen concentration is increased while the concentration of the nitrate nitrogen is increased. However, the denitrification reaction is mainly carried out in an anoxic environment in the oily sewage treatment process, and nitrate nitrogen S in an anoxic zoneNOThe concentration is an important index for measuring the denitrification effect, reflects the progress of the denitrification reaction process and converts nitrate nitrogen SNOThe concentration is controlled to be properWithin the range, the potential of denitrification reaction can be improved. Therefore, the dissolved oxygen DO and the nitrate nitrogen S are dynamically and optimally controlled in real timeNOThe concentration is very necessary for ensuring the effluent quality and saving energy and reducing consumption. Because of the characteristics of suspended matters, high chroma content, diversified components of organic matters, nonlinearity, time-varying property, uncertain dynamics and the like of the oily sewage treatment process, the dissolved oxygen DO and the nitrate nitrogen S are increasedNOThe control difficulty of (c); in recent years, some scholars optimize the oily sewage treatment process by adopting a multi-objective genetic algorithm (MOGA) based on BSM1 so as to minimize aeration energy consumption and pumping energy consumption. However, most of the optimization schemes belong to static optimization and steady-state optimization, dynamic real-time adjustment according to the change of the quality and quantity of inlet water is difficult to carry out, and most of the optimization schemes are realized by adopting a genetic algorithm. Compared with a genetic algorithm, the particle swarm algorithm has the advantages of high convergence speed, no need of complex cross variation operation, simple algorithm, few parameters and easiness in implementation, but the standard particle swarm algorithm is easy to have the problems of premature convergence and low search precision, so that the self-adaptive multi-target particle swarm algorithm is designed, the convergence precision of the algorithm is improved, and the dissolved oxygen DO and nitrate nitrogen S can be well realizedNOThe concentration is optimally controlled, the operation cost is reduced, and the method has good practical application value.
Disclosure of Invention
The technical problem to be solved by the invention is that the standard particle swarm algorithm in the prior art is easy to generate premature convergence and has low search precision, and the invention provides a novel optimization control method for the oily sewage treatment process based on self-adaptive multi-target particle swarm. The method has the advantages of no premature convergence and high search precision.
In order to solve the problems, the technical scheme adopted by the invention is as follows: an optimized control method for oily sewage treatment process based on self-adaptive multi-target particle swarm comprises the steps of firstly establishing an optimized set value of dissolved oxygen and nitrate nitrogen and an objective function of aeration energy consumption and pumping energy consumption in the oily sewage treatment process through a fuzzy neural network; secondly, aiming at the defects that the standard particle swarm algorithm is easy to premature convergence and low in convergence precision, a self-adaptive multi-target particle swarm optimization method is adopted to realize the optimization of the oily sewage treatment target function and obtain the optimized set values of dissolved oxygen and nitrate nitrogen; finally, the controller is used for tracking and controlling the optimized set values of the dissolved oxygen and the nitrate nitrogen to complete the multi-objective optimization control of the oily sewage treatment process;
the method specifically comprises the following steps:
1) designing an objective function for optimal control of the oily sewage treatment process;
(2) establishing a relational expression between the optimal set values of dissolved oxygen and nitrate nitrogen and aeration energy consumption and pumping energy consumption by using a fuzzy neural network;
(3) the effluent quality is restrained;
(4) obtaining a Pareto optimal solution by utilizing a self-adaptive multi-target particle swarm optimization objective function, which specifically comprises the following steps:
initializing the velocity v of the particle swarmi(0) Position ai(0) Inertia weight ωi(0) Learning factor c1i(0) And c2i(0) And setting the initial position of each particle as the current historical optimal position pi(0) Simultaneously setting a population scale S, a maximum evolution algebra M and a variable dimension D;
secondly, calculating the fitness value of each particle according to the objective function, and determining the individual optimal solution p of the t iterationi(t) defined by the formula:
Figure BDA0001436067600000031
the non-dominated solution set A (t) is updated by A (t-1), and the formula is as follows:
Figure BDA0001436067600000032
wherein A (t) ═ a1(t),a2(t),…,aQ(t)]Q is the maximum capacity of the knowledge base A (t), K is the number of non-dominant solutions contained in the knowledge base,
Figure BDA0001436067600000033
denotes ai(t-1) and pi(t-1) are independent of each other;
third, determining the global optimal solution gBest (t +1) of the t +1 iteration
Figure BDA0001436067600000034
Wherein, gBest (t +1) is the global optimal solution of the t +1 th iteration, and dgBest (t +1) is the global optimal solution of the t +1 th iteration with better diversity, wherein, the definition formula of the dgBest (t +1) is as follows:
dgBest(t+1)＝a(t),a(t)∈Μtbest. (4)
cgBest (t +1) is a global optimal solution with better convergence in the t +1 th iteration, and the definition formula is as follows:
cgBest(t+1)＝argmaxCDt(ai(t)), (5)
CDt(ai(t)) non-dominant solution ai(t) convergence, i ═ 1,2,3 … K, e (t) is the distribution entropy of the non-dominated solution set for the t th iteration, which is defined by the formula:
Figure BDA0001436067600000035
wherein u isnIs the cell of the non-dominant solution in the knowledge base, N is 1,2,3 …, N is the total number of cells, pt(un) Is the t-th iteration cell unThe expression of the probability distribution function of (1) is:
Figure BDA0001436067600000036
mt(un) Is the t-th iteration cell unNumber of non-dominant solutions, Mt(un) Is the t-th iteration cell unOf all non-dominant solutions contained in (c), wherein the smallest m ist(un) Is marked as mtbestAccordingly, MtbestIs optimal for the t-th iterationThe non-dominant solution set of (2), and the selection of cgBest (t +1) is determined by the degree of convergence that reflects the dominant relationship, thereby determining the degree of convergence CDt(ai(t)) is defined as:
Figure BDA0001436067600000037
wherein,
Figure BDA0001436067600000041
is the jth solution a that can be non-dominatedi(t) solution, DSt(ai(t)) is the t-th iteration non-dominated solution ai(t) a dominant intensity defined by the formula:
Figure BDA0001436067600000042
DSt(ai(t)) is the non-dominant solution a for the t-th iterationi(t) the total number of solutions that can be dominated, A (t) is the set of non-dominated solutions for the t iteration, A (t-1) is the set of non-dominated solutions for the t-1 iteration;
updating the speed and position of each particle;
judging whether the maximum evolution times M is reached, if so, ending, otherwise, returning to the step II;
(5) finding a group of satisfied optimal solutions in the current state from a group of Pareto optimal solutions obtained by a self-adaptive multi-target particle swarm algorithm to serve as an optimal set value of a bottom controller;
(6) executing a bottom layer control strategy, and respectively passing the concentrations of dissolved oxygen and nitrate nitrogen through an oxygen conversion coefficient K of the aeration tankLa5And internal reflux quantity QaAnd (6) carrying out adjustment.
In the above technical solution, preferably, the controller is a PID.
In the above technical solution, preferably, the objective function for optimal control of the oily wastewater treatment process is:
Figure BDA0001436067600000043
Figure BDA0001436067600000044
wherein x (k) ═ x1(k)，x2(k)]TFor the optimization variable at time k, x1(k) The dissolved oxygen DO concentration setpoint at time k, x2(k) Is the nitrate nitrogen concentration set value at the time k, fAE(x) And fPE(x) Respectively are relational expressions g of aeration energy consumption, pumping energy consumption and optimization variables1(x) And g2(x) Respectively are relational expressions of effluent ammonia nitrogen, total nitrogen concentration and optimization variable, C1And C2Respectively the constraint values of the effluent ammonia nitrogen and the total nitrogen, C1∈[0,4],C2∈[0,18]；
Figure BDA0001436067600000045
And
Figure BDA0001436067600000046
respectively representing the lower limit and the upper limit, x, of the optimum set value of dissolved oxygen1∈[0.4,3],
Figure BDA0001436067600000047
And
Figure BDA0001436067600000048
respectively represents the lower limit and the upper limit, x, of the nitrate nitrogen concentration optimization set value2∈[0.5,2]The optimization period is 2 h.
In the above technical solution, preferably, the fuzzy neural network is calculated as follows:
Figure BDA0001436067600000049
Figure BDA00014360676000000410
Figure BDA00014360676000000411
wherein x (k) ═ x1(k)，x2(k)]TFor the fuzzy neural network input at time k, cj＝[c1j，c2j]，σj＝[σ1j，σ2j]Respectively the central vector and the width vector of the jth neuron of the RBF layer,
Figure BDA0001436067600000051
the output of the RBF layer of the j-th neuron at the k-th moment, wherein P is the number of neurons of the RBF layer and the rule layer; v. ofl(k) V (k) is [ v ] for the l-th regular layer output corresponding to the k-th time1(k)，v2(k)…vP(k)]TOutputting a vector for the k-th time rule layer; w is a1＝[w1 1，w2 1…wP 1]And w2＝[w1 2，w2 2…wP 2]Is the weight vector between the output neuron and the rule layer, y (k) is the output of the neural network,
Figure BDA0001436067600000052
the actual physical quantity output of the sewage treatment system is obtained based on BSM1 model data;
setting the target function of network adjustment at the moment k as follows:
Figure BDA0001436067600000053
by adopting a gradient descent algorithm, the weight updating formula is as follows:
Figure BDA0001436067600000054
in the formula, alphaq(k)＝[θq(k)T cq(k)T σq(k)T]TThe learning parameter vector of the network is represented, and the network learning rate eta is 0.01;
in the above technical solution, preferably, the constraint in the optimization model is processed by a penalty function method, where a constraint penalty term is defined as:
fpenalty(x)＝max{g1(x)-4,0}+max{g2(x)-18,0}, (16)
the aeration energy consumption and pumping energy consumption objective function added with penalty items is as follows:
Figure BDA0001436067600000055
and converting the established constraint optimization problem in the oily sewage treatment process into an unconstrained multi-objective optimization problem, wherein epsilon is a penalty factor and is set as a larger positive real number.
In the above technical solution, preferably, the speed and the position of each particle are updated:
vi(t+1)＝ωi(t)vi(t)+c1r1(pi(t)-ai(t))+c2r2(gBestd(t)-ai(t))； (18)
ai(t+1)＝ai(t-1)+vi(t+1)；
wherein r is1And r2Respectively representing the best previous position coefficient and the global optimum position coefficient, r1And r2Take [0,1]Any number of (a).
In the above technical solution, preferably, in the step (6), a bottom layer control strategy is executed, and the concentrations of dissolved oxygen and nitrate nitrogen respectively pass through the 5 th zone oxygen conversion coefficient K of the aeration tankLa5And internal reflux quantity QaAnd (6) carrying out adjustment.
In the above-described aspect, preferably, in step (r), the velocity v of the particle group is initializedi(0) Position ai(0) Inertia weight ωi(0) Learning factor c1i(0) And c2i(0) And setting the initial position of each particle as the current historical optimal position pi(0)，Meanwhile, the population size S is set to be 40, the maximum evolution algebra M is set to be 30, and the variable dimension D is set to be 2.
The invention provides an optimal control method for an oily sewage treatment process based on self-adaptive multi-target particle swarm, which comprises the following steps:
(1) designing an objective function for optimal control of the oily sewage treatment process:
Figure BDA0001436067600000061
Figure BDA0001436067600000062
wherein x (k) ═ x1(k)，x2(k)]TFor the optimization variable at time k, x1(k) The dissolved oxygen DO concentration setpoint at time k, x2(k) Is the nitrate nitrogen concentration set value at the time k, fAE(x) And fPE(x) Respectively are relational expressions g of aeration energy consumption, pumping energy consumption and optimization variables1(x) And g2(x) Respectively are relational expressions of effluent ammonia nitrogen, total nitrogen concentration and optimization variable, C1And C2Respectively the constraint values of the effluent ammonia nitrogen and the total nitrogen, C1∈[0,4],C2∈[0,18]；
Figure BDA0001436067600000063
And
Figure BDA0001436067600000064
respectively representing the lower limit and the upper limit, x, of the optimum set value of dissolved oxygen1∈[0.4,3],
Figure BDA0001436067600000065
And
Figure BDA0001436067600000066
respectively represents the lower limit and the upper limit, x, of the nitrate nitrogen concentration optimization set value2∈[0.5,2]The optimization period is 2 h.
(2) Establishing a relational expression between the optimal set values of the dissolved oxygen and the nitrate nitrogen and the aeration energy consumption and the pumping energy consumption by using a fuzzy neural network, wherein the calculation mode of the fuzzy neural network is as follows:
Figure BDA0001436067600000067
Figure BDA0001436067600000068
Figure BDA0001436067600000069
wherein x (k) ═ x1(k)，x2(k)]TFor the fuzzy neural network input at time k, cj＝[c1j，c2j]，σj＝[σ1j，σ2j]Respectively the central vector and the width vector of the jth neuron of the RBF layer,
Figure BDA00014360676000000610
the output of the RBF layer of the j-th neuron at the k-th moment, wherein P is the number of neurons of the RBF layer and the rule layer; v. ofl(k) V (k) is [ v ] for the l-th regular layer output corresponding to the k-th time1(k)，v2(k)…vP(k)]TOutputting a vector for the k-th time rule layer; w is a1＝[w1 1，w2 1…wP 1]And w2＝[w1 2，w2 2…wP 2]Is the weight vector between the output neuron and the rule layer, y (k) is the output of the neural network,
Figure BDA00014360676000000611
the actual physical quantity output of the sewage treatment system is obtained based on BSM1 model data.
Setting the target function of network adjustment at the moment k as follows:
Figure BDA0001436067600000071
by adopting a gradient descent algorithm, the weight updating formula is as follows:
Figure BDA0001436067600000072
in the formula, alphaq(k)＝[θq(k)T cq(k)T σq(k)T]TThe network learning rate η is 0.01 for the learning parameter vector of the network.
(3) Effluent quality constraint treatment
And processing the constraint in the optimization model by adopting a penalty function method, wherein a constraint penalty term is defined as:
fpenalty(x)＝max{g1(x)-4,0}+max{g2(x)-18,0}, (7)
the aeration energy consumption and pumping energy consumption objective function added with penalty items is as follows:
Figure BDA0001436067600000073
namely, the established constraint optimization problem in the oily sewage treatment process is converted into an unconstrained multi-objective optimization problem. Wherein epsilon is a penalty factor and is set as a large positive real number.
(4) Obtaining a Pareto optimal solution by utilizing a self-adaptive multi-target particle swarm optimization objective function, which specifically comprises the following steps:
initializing the velocity v of the particle swarmi(0) Position ai(0) Inertia weight ωi(0) Learning factor c1i(0) And c2i(0) And setting the initial position of each particle as the current historical optimal position pi(0). Meanwhile, the population scale S is set to be 40, the maximum evolution algebra M is set to be 30, and the variable dimension D is set to be 2;
and secondly, calculating the fitness value of each particle according to the objective function. Determining the t-th iterationIndividual optimal solution pi(t) defined by the formula:
Figure BDA0001436067600000074
the non-dominated solution set A (t) is updated by A (t-1), and the formula is as follows:
Figure BDA0001436067600000075
wherein A (t) ═ a1(t),a2(t),…,aQ(t)]Q is the maximum capacity of the knowledge base A (t), K is the number of non-dominant solutions contained in the knowledge base,
Figure BDA0001436067600000076
denotes ai(t-1) and pi(t-1) are not mutually exclusive.
And thirdly, determining the global optimal solution gBest (t +1) of the t +1 th iteration.
Figure BDA0001436067600000077
Wherein, gBest (t +1) is the global optimal solution of the t +1 th iteration, and dgBest (t +1) is the global optimal solution of the t +1 th iteration with better diversity, wherein, the definition formula of the dgBest (t +1) is as follows:
dgBest(t+1)＝a(t),a(t)∈Μtbest. (12)
cgBest (t +1) is a global optimal solution with better convergence in the t +1 th iteration, and the definition formula is as follows:
cgBest(t+1)＝argmaxCDt(ai(t)), (13)
CDt(ai(t)) non-dominant solution ai(t) convergence, i ═ 1,2,3 … K, e (t) is the distribution entropy of the non-dominated solution set for the t th iteration, which is defined by the formula:
Figure BDA0001436067600000081
wherein u isnIs the cell of the non-dominant solution in the knowledge base, N is 1,2,3 …, N is the total number of cells, pt(un) Is the t-th iteration cell unThe expression of the probability distribution function of (1) is:
Figure BDA0001436067600000082
mt(un) Is the t-th iteration cell unNumber of non-dominant solutions, Mt(un) Is the t-th iteration cell unA solution set of all non-dominant solutions contained in (a). Wherein the smallest mt(un) Is marked as mtbestAccordingly, MtbestIs the optimal non-dominated solution set for the t-th iteration. Further, the selection of cgBest (t +1) is determined by the convergence degree that can reflect the dominant relationship, and the convergence degree CDt(ai(t)) is defined as:
Figure BDA0001436067600000083
wherein,
Figure BDA0001436067600000084
is the jth solution a that can be non-dominatedi(t) solution, DSt(ai(t)) is the t-th iteration non-dominated solution ai(t) a dominant intensity defined by the formula:
Figure BDA0001436067600000085
DSt(ai(t)) is the non-dominant solution a for the t-th iterationi(t) the total number of solutions that can be dominated, A (t) is the set of non-dominated solutions for the t iteration, A (t-1) is the set of non-dominated solutions for the t-1 iteration;
updating the speed and the position of each particle:
vi(t+1)＝ωi(t)vi(t)+c1r1(pi(t)-ai(t))+c2r2(gBestd(t)-ai(t))； (18)
ai(t+1)＝ai(t-1)+vi(t+1)； (19)
wherein r is1And r2Respectively representing the best previous position coefficient and the global optimum position coefficient, r1And r2Take [0,1]Any number of (a);
judging whether the maximum evolution times M is reached, if so, ending, otherwise, returning to the step II.
(5) And finding a group of satisfactory optimal solutions in the current state from a group of Pareto optimal solutions obtained by the self-adaptive multi-target particle swarm algorithm to serve as the optimal set values of the bottom PID controller.
(6) Executing a bottom layer PID control strategy, and respectively passing the concentrations of dissolved oxygen and nitrate nitrogen through the oxygen conversion coefficient K of the 5 th subarea of the aeration tankLa5And internal reflux quantity QaAnd (6) carrying out adjustment.
The invention aims at the complex, dynamic and unstable biochemical reaction process of the current oily sewage treatment process by an activated sludge method and the characteristics of nonlinearity, time-varying property and hysteresis; simultaneously dissolving oxygen DO and nitrate nitrogen SNOThe concentration has strong coupling relation, so as to meet the requirement of reducing the operation energy consumption when the effluent quality reaches the standard and realize the dissolved oxygen DO and the nitrate nitrogen SNOThe multi-target control of the concentration adopts an oily sewage treatment model predictive control method based on multi-target particle swarm to realize dissolved oxygen DO and nitrate nitrogen SNOThe concentration control has the characteristics of high control precision, good stability and the like; the invention adopts an oily sewage treatment model based on self-adaptive multi-target particle swarm to dissolve oxygen DO and nitrate nitrogen S in the sewage treatment processNOThe concentration is optimally controlled, and the optimal control method solves the problem of optimal solution of a plurality of objective functions, so that the controller better meets the change of the current environment, and the dissolved oxygen DO and the nitrate nitrogen S are realizedNOThe concentration is accurately controlled in a real-time closed loop manner, and the condition that a plurality of controllers are required to be designed for the current sewage treatment plant is avoidedThe complex process of line control has the characteristics of strong real-time performance, simple structure and the like, and obtains better technical effect.
Drawings
FIG. 1 is an overall structure diagram of the adaptive multi-objective particle swarm optimization control system of the invention;
FIG. 2 is a graph showing the results of the dissolved oxygen DO concentration in the control system of the present invention;
FIG. 3 is a DO concentration error plot of the dissolved oxygen of the control system of the present invention;
FIG. 4 shows nitrate nitrogen S in the control system of the present inventionNOA concentration result graph;
FIG. 5 shows nitrate nitrogen S in the control system of the present inventionNOConcentration result error plot.
The present invention will be further illustrated by the following examples, but is not limited to these examples.
Detailed Description
[ example 1 ]
An optimized control method for oily sewage treatment process based on self-adaptive multi-target particle swarm is shown in figure 1, and realizes dissolved oxygen DO and nitrate nitrogen S in the oily sewage treatment processNOMulti-objective optimization control of concentration; the control method obtains dissolved oxygen DO and nitrate nitrogen S through online modeling of a fuzzy neural networkNOThe function relation between the optimized set value of the concentration and the aeration energy consumption, the pumping energy consumption and the effluent quality is realized by applying the optimization control method based on the self-adaptive multi-target particle swarm to the oily sewage treatment processNOThe concentration is optimally controlled, the energy consumption of the oily sewage treatment process is reduced, the investment and the operation cost are saved, the stable and efficient operation of a sewage treatment plant is ensured, and the method belongs to the field of water treatment and also belongs to the field of intelligent control.
The invention adopts the following technical scheme and implementation steps:
1. the design of the optimal control method for the oily sewage treatment process based on the self-adaptive multi-target particle swarm comprises the following steps: aiming at the dissolved oxygen DO concentration and nitrate nitrogen S in a sequencing batch intermittent activated sludge systemNOControl is carried out by taking aeration energy consumption and pumping energy consumption as control quantitiesDissolved oxygen DO and nitrate nitrogen SNOThe concentration is controlled quantity, and the whole framework of the self-adaptive multi-target particle swarm optimization control system is as shown in figure 1;
(1) designing an objective function for optimal control of the oily sewage treatment process:
Figure BDA0001436067600000101
Figure BDA0001436067600000102
wherein x (k) ═ x1(k)，x2(k)]TFor the optimization variable at time k, x1(k) Is a dissolved oxygen concentration set value at the time k, x2(k) Is the nitrate nitrogen concentration set value at the time k, fAE(x) And fPE(x) Respectively are relational expressions g of aeration energy consumption, pumping energy consumption and optimization variables1(x) And g2(x) Respectively are relational expressions of effluent ammonia nitrogen, total nitrogen concentration and optimization variable, C1And C2Respectively the constraint values of the effluent ammonia nitrogen and the total nitrogen, C1∈[0,4],C2∈[0,18]；
Figure BDA0001436067600000103
And
Figure BDA0001436067600000104
respectively representing the lower limit and the upper limit, x, of the optimum set value of dissolved oxygen1∈[0.4,3],
Figure BDA0001436067600000105
And
Figure BDA0001436067600000106
respectively represents the lower limit and the upper limit, x, of the nitrate nitrogen concentration optimization set value2∈[0.5,2]The optimization period is 2 h.
(2) Establishing a relational expression between the optimal set values of the dissolved oxygen and the nitrate nitrogen and the aeration energy consumption and the pumping energy consumption by using a fuzzy neural network, wherein the calculation mode of the fuzzy neural network is as follows:
Figure BDA0001436067600000107
Figure BDA0001436067600000108
Figure BDA0001436067600000109
wherein x (k) ═ x1(k)，x2(k)]TFor the fuzzy neural network input at time k, cj＝[c1j，c2j]，σj＝[σ1j，σ2j]Respectively the central vector and the width vector of the jth neuron of the RBF layer,
Figure BDA00014360676000001010
the output of the RBF layer of the j-th neuron at the k-th moment, wherein P is the number of neurons of the RBF layer and the rule layer; v. ofl(k) V (k) is [ v ] for the l-th regular layer output corresponding to the k-th time1(k)，v2(k)…vP(k)]TOutputting a vector for the k-th time rule layer; w is a1＝[w1 1，w2 1…wP 1]And w2＝[w1 2，w2 2…wP 2]Is the weight vector between the output neuron and the rule layer, y (k) is the output of the neural network,
Figure BDA00014360676000001011
the actual physical quantity output of the sewage treatment system is obtained based on BSM1 model data.
Setting the target function of network adjustment at the moment k as follows:
Figure BDA0001436067600000111
by adopting a gradient descent algorithm, the weight updating formula is as follows:
Figure BDA0001436067600000112
in the formula, alphaq(k)＝[θq(k)T cq(k)T σq(k)T]TThe network learning rate η is 0.01 for the learning parameter vector of the network.
(3) Effluent quality constraint treatment
And processing the constraint in the optimization model by adopting a penalty function method, wherein a constraint penalty term is defined as:
fpenalty(x)＝max{g1(x)-4,0}+max{g2(x)-18,0}, (7)
the aeration energy consumption and pumping energy consumption objective function added with penalty items is as follows:
Figure BDA0001436067600000113
namely, the established constraint optimization problem in the oily sewage treatment process is converted into an unconstrained multi-objective optimization problem. Wherein epsilon is a penalty factor and is set as a large positive real number.
(4) Obtaining a Pareto optimal solution by utilizing a self-adaptive multi-target particle swarm optimization objective function, which specifically comprises the following steps:
initializing the velocity v of the particle swarmi(0) Position ai(0) Inertia weight ωi(0) Learning factor c1i(0) And c2i(0) And setting the initial position of each particle as the current historical optimal position pi(0). Meanwhile, the population scale S is set to be 40, the maximum evolution algebra M is set to be 30, and the variable dimension D is set to be 2;
and secondly, calculating the fitness value of each particle according to the objective function. Determining individual optimal solution p of t-th iterationi(t) defined by the formula:
Figure BDA0001436067600000114
the non-dominated solution set A (t) is updated by A (t-1), and the formula is as follows:
Figure BDA0001436067600000115
wherein A (t) ═ a1(t),a2(t),…,aQ(t)]Q is the maximum capacity of the knowledge base A (t), K is the number of non-dominant solutions contained in the knowledge base,
Figure BDA0001436067600000116
denotes ai(t-1) and pi(t-1) are not mutually exclusive.
And thirdly, determining the global optimal solution gBest (t +1) of the t +1 th iteration.
Figure BDA0001436067600000117
Wherein, gBest (t +1) is the global optimal solution of the t +1 th iteration, and dgBest (t +1) is the global optimal solution of the t +1 th iteration with better diversity, wherein, the definition formula of the dgBest (t +1) is as follows:
dgBest(t+1)＝a(t),a(t)∈Μtbest. (12)
cgBest (t +1) is a global optimal solution with better convergence in the t +1 th iteration, and the definition formula is as follows:
cgBest(t+1)＝argmaxCDt(ai(t)), (13)
CDt(ai(t)) non-dominant solution ai(t) convergence, i ═ 1,2,3 … K, e (t) is the distribution entropy of the non-dominated solution set for the t th iteration, which is defined by the formula:
Figure BDA0001436067600000121
wherein u isnIs a cell of the non-dominant solution in the knowledge base,n is 1,2,3 …, N is the total number of cells, pt(un) Is the t-th iteration cell unThe expression of the probability distribution function of (1) is:
Figure BDA0001436067600000122
mt(un) Is the t-th iteration cell unNumber of non-dominant solutions, Mt(un) Is the t-th iteration cell unA solution set of all non-dominant solutions contained in (a). Wherein the smallest mt(un) Is marked as mtbestAccordingly, MtbestIs the optimal non-dominated solution set for the t-th iteration. Further, the selection of cgBest (t +1) is determined by the convergence degree that can reflect the dominant relationship, and the convergence degree CDt(ai(t)) is defined as:
Figure BDA0001436067600000123
wherein,
Figure BDA0001436067600000124
is the jth solution a that can be non-dominatedi(t) solution, DSt(ai(t)) is the t-th iteration non-dominated solution ai(t) a dominant intensity defined by the formula:
Figure BDA0001436067600000125
DSt(ai(t)) is the non-dominant solution a for the t-th iterationi(t) the total number of solutions that can be dominated, A (t) is the set of non-dominated solutions for the t iteration, A (t-1) is the set of non-dominated solutions for the t-1 iteration;
updating the speed and the position of each particle:
vi(t+1)＝ωi(t)vi(t)+c1r1(pi(t)-ai(t))+c2r2(gBestd(t)-ai(t))； (18)
ai(t+1)＝ai(t-1)+vi(t+1)； (19)
wherein r is1And r2Respectively representing the best previous position coefficient and the global optimum position coefficient, r1And r2Take [0,1]Any number of (a); judging whether the maximum evolution times M is reached, if so, ending, otherwise, returning to the step II.
(5) And finding a group of satisfactory optimal solutions in the current state from a group of Pareto optimal solutions obtained by the self-adaptive multi-target particle swarm algorithm to serve as the optimal set values of the bottom PID controller.
(6) And executing a bottom layer PID control strategy, wherein the parameters of the PID controller are set as follows: kP,1＝200，KI,1＝15，K D,12 and KP,2＝20000，KI,2＝5000，KD,2400. The concentrations of dissolved oxygen and nitrate nitrogen respectively pass through the 5 th subarea oxygen conversion coefficient K of the aeration tankLa5And internal reflux quantity QaAnd (6) carrying out adjustment.
(7) Using PID controller to control dissolved oxygen DO and nitrate nitrogen SNOThe concentration optimization set value is tracked and controlled, and the output of the whole control system is dissolved oxygen DO and nitrate nitrogen SNOOptimizing the set value and tracking control value of the concentration value; fig. 2 shows the dissolved oxygen DO concentration optimization setpoint and tracking control values for the system, X-axis: time, in days, Y-axis: the unit of the optimized set value and the tracking control value of the dissolved oxygen DO is milligram/liter, the solid line is the optimized set value of the dissolved oxygen DO concentration, and the dotted line is the actual tracking control value of the dissolved oxygen DO; the error between the optimized set dissolved oxygen DO concentration value and the actual dissolved oxygen DO tracking control concentration value is shown in FIG. 3, X axis: time, in days, Y-axis: dissolved oxygen DO concentration error in milligrams per liter; FIG. 4 shows nitrate nitrogen S of the systemNOOptimization set value and tracking control value, X-axis: time, in days, Y-axis: nitrate nitrogen SNOIn mg/l, and the solid line is nitrate nitrogen SNOConcentration optimization set point, dotted line is nitrate nitrogen SNOTracking control of concentrationA value of the metric; optimized setting of nitrate nitrogen SNOConcentration value and actual nitrate nitrogen SNOError in tracking control concentration values as shown in fig. 5, X-axis: time, in days, Y-axis: nitrate nitrogen SNOThe error in concentration, in mg/l, proved the effectiveness of the method.
Particular attention is paid to: the invention is described for convenience only, and the dissolved oxygen DO and nitrate nitrogen S are adoptedNOThe concentration control and the ammonia nitrogen control in the sewage treatment process can also be applied to the control of the ammonia nitrogen and the like, and the control by adopting the principle of the invention is within the scope of the invention.

Claims (8)

1. An optimized control method for oily sewage treatment process based on self-adaptive multi-target particle swarm comprises the steps of firstly establishing an optimized set value of dissolved oxygen and nitrate nitrogen and an objective function of aeration energy consumption and pumping energy consumption in the oily sewage treatment process through a fuzzy neural network; secondly, aiming at the defects of easy premature convergence and low convergence precision of a standard particle swarm algorithm, a self-adaptive multi-target particle swarm optimization method is adopted to realize the optimization of an oily sewage treatment target function, and simultaneously, optimized set values of dissolved oxygen and nitrate nitrogen are obtained, so that premature convergence is avoided, and the search precision is high; finally, the controller is used for tracking and controlling the optimized set values of the dissolved oxygen and the nitrate nitrogen, so that the complex process that a plurality of controllers are required to be designed for control in the current sewage treatment plant is avoided, and the multi-objective optimized control of the oily sewage treatment process is completed;
the method specifically comprises the following steps:
(1) designing an objective function for optimal control of the oily sewage treatment process;
(2) establishing a relational expression between the optimal set values of dissolved oxygen and nitrate nitrogen and aeration energy consumption and pumping energy consumption by using a fuzzy neural network;
(3) the effluent quality is restrained;
(4) obtaining a Pareto optimal solution by utilizing a self-adaptive multi-target particle swarm optimization objective function, which specifically comprises the following steps:
initializing the velocity v of the particle swarmi(0) Position ai(0) The inertia weightHeavy omegai(0) Learning factor c1i(0) And c2i(0) And setting the initial position of each particle as the current historical optimal position pi(0) Simultaneously setting a population scale S, a maximum evolution algebra M and a variable dimension D;
secondly, calculating the fitness value of each particle according to the objective function, and determining the individual optimal solution p of the t iterationi(t) defined by the formula:
Figure FDA0003454627620000011
the non-dominated solution set A (t) is updated by A (t-1), and the formula is as follows:
Figure FDA0003454627620000013
wherein A (t) ═ a1(t),a2(t),…,aQ(t)]Q is the maximum capacity of the knowledge base A (t), K is the number of non-dominant solutions contained in the knowledge base,
Figure FDA0003454627620000014
denotes ai(t-1) and pi(t-1) are independent of each other;
third, determining global optimum solution gBest (t +1) of the t-th iteration
Figure FDA0003454627620000012
Wherein, gBest (t +1) is the global optimal solution of the t +1 th iteration, and dgBest (t +1) is the global optimal solution of the t +1 th iteration with better diversity, wherein, the definition formula of the dgBest (t +1) is as follows:
dgBest(t+1)＝a(t),a(t)∈Μtbest.
cgBest (t +1) is a global optimal solution with better convergence in the t +1 th iteration, and the definition formula is as follows:
cgBest(t+1)＝arg maxCDt(ai(t)),
CDt(ai(t)) non-dominant solution ai(t) convergence, i ═ 1,2,3 … K, e (t) is the distribution entropy of the non-dominated solution set for the t th iteration, which is defined by the formula:
Figure FDA0003454627620000021
wherein u isnIs the cell of the non-dominant solution in the knowledge base, N is 1,2,3 …, N is the total number of cells, pt(un) Is the t-th iteration cell unThe expression of the probability distribution function of (1) is:
Figure FDA0003454627620000022
mt(un) Is the t-th iteration cell unNumber of non-dominant solutions, Mt(un) Is the t-th iteration cell unOf all non-dominant solutions contained in (c), wherein the smallest m ist(un) Is marked as mtbestAccordingly, MtbestIs a non-dominated solution set optimal for the t-th iteration, and the selection of cgBest (t +1) is determined by the degree of convergence that reflects the dominance relationship, and the degree of convergence CDt(ai(t)) is defined as:
Figure FDA0003454627620000023
wherein,
Figure FDA0003454627620000024
is the jth solution a that can be non-dominatedi(t) solution, DSt(ai(t)) is the t-th iteration non-dominated solution ai(t) a dominant intensity defined by the formula:
Figure FDA0003454627620000025
DSt(ai(t)) is the non-dominant solution a for the t-th iterationi(t) the total number of solutions that can be dominated, A (t) is the set of non-dominated solutions for the t iteration, B (t-1) is the set of non-dominated solutions for the t-1 iteration;
updating the speed and position of each particle;
judging whether the maximum evolution times M is reached, if so, ending, otherwise, returning to the step II;
(5) finding a group of satisfied optimal solutions in the current state from a group of Pareto optimal solutions obtained by a self-adaptive multi-target particle swarm algorithm to serve as an optimal set value of a bottom controller;
(6) executing a bottom layer control strategy, and respectively passing the concentrations of dissolved oxygen and nitrate nitrogen through an oxygen conversion coefficient K of the aeration tankLa5And internal reflux quantity QaAnd (6) carrying out adjustment.
2. The oily sewage treatment process optimization control method based on the adaptive multi-target particle swarm of claim 1, wherein the controller is PID.
3. The oily sewage treatment process optimization control method based on the adaptive multi-objective particle swarm as claimed in claim 1, wherein the objective function for the oily sewage treatment process optimization control is as follows:
Figure FDA0003454627620000031
Figure FDA0003454627620000032
wherein x (k) ═ x1(k)，x2(k)]TFor the optimization variable at time k, x1(k) The dissolved oxygen DO concentration setpoint at time k, x2(k) Nitre at time kNitrogen concentration set value, fAE(x) And fPE(x) Respectively are relational expressions g of aeration energy consumption, pumping energy consumption and optimization variables1(x) And g2(x) Respectively are relational expressions of effluent ammonia nitrogen, total nitrogen concentration and optimization variable, C1And C2Respectively the constraint values of the effluent ammonia nitrogen and the total nitrogen, C1∈[0,4],C2∈[0,18]；
Figure FDA00034546276200000311
And
Figure FDA00034546276200000310
respectively representing the lower limit and the upper limit, x, of the optimum set value of dissolved oxygen1∈[0.4,3],
Figure FDA00034546276200000312
And
Figure FDA00034546276200000313
respectively represents the lower limit and the upper limit, x, of the nitrate nitrogen concentration optimization set value2∈[0.5,2]The optimization period is 2 h.
4. The oily sewage treatment process optimization control method based on the adaptive multi-target particle swarm as claimed in claim 1, wherein the fuzzy neural network is calculated in the following way:
Figure FDA0003454627620000033
Figure FDA0003454627620000034
Figure FDA0003454627620000035
wherein x (k) ═ x1(k)，x2(k)]TFor the fuzzy neural network input at time k, cj＝[c1j，c2j]，σj＝[σ1j，σ2j]Respectively the central vector and the width vector of the jth neuron of the RBF layer,
Figure FDA0003454627620000036
the output of the RBF layer of the j-th neuron at the k-th moment, wherein P is the number of neurons of the RBF layer and the rule layer; v. ofl(k) V (k) is [ v ] for the l-th regular layer output corresponding to the k-th time1(k)，v2(k)…vP(k)]TOutputting a vector for the k-th time rule layer;
Figure FDA0003454627620000037
and
Figure FDA0003454627620000038
is the weight vector between the output neuron and the rule layer, y (k) is the output of the neural network,
Figure FDA0003454627620000039
the actual physical quantity output of the sewage treatment system is obtained based on BSM1 model data;
setting the target function of network adjustment at the moment k as follows:
Figure FDA0003454627620000041
by adopting a gradient descent algorithm, the weight updating formula is as follows:
Figure FDA0003454627620000042
in the formula, alphaq(k)＝[θq(k)T cq(k)T σq(k)T]TAs a networkThe network learning rate η of (1) is 0.01.
5. The oily sewage treatment process optimization control method based on the adaptive multi-target particle swarm is characterized in that constraints in an optimization model are processed by adopting a penalty function method, wherein a constraint penalty term is defined as:
fpenalty(x)＝max{g1(x)-4,0}+max{g2(x)-18,0},
the aeration energy consumption and pumping energy consumption objective function added with penalty items is as follows:
Figure FDA0003454627620000043
and converting the established constraint optimization problem in the oily sewage treatment process into an unconstrained multi-objective optimization problem, wherein epsilon is a penalty factor and is set as a larger positive real number.
6. The oily sewage treatment process optimization control method based on the adaptive multi-target particle swarm of claim 1, which is characterized in that the speed and the position of each particle are updated:
vi(t+1)＝ωi(t)vi(t)+c1r1(pi(t)-ai(t))+c2r2(gBestd(t)-ai(t))；
ai(t+1)＝ai(t-1)+vi(t+1)；
wherein r is1And r2Respectively representing the best previous position coefficient and the global optimum position coefficient, r1And r2Take [0,1]Any number of (a).
7. The method for optimally controlling the oily sewage treatment process based on the adaptive multi-target particle swarm in the claim 1, wherein in the step (6), a bottom layer control strategy is implemented, and the dissolved oxygen and nitrate nitrogen concentrations respectively pass through the 5 th subarea oxygen conversion coefficient K of the aeration tankLa5And internal reflux quantity QaAnd (6) carrying out adjustment.
8. The method for optimally controlling the oily sewage treatment process based on the self-adaptive multi-target particle swarm in the claim 1, wherein in the step (r), the speed v of the particle swarm is initializedi(0) Position ai(0) Inertia weight ωi(0) Learning factor c1i(0) And c2i(0) And setting the initial position of each particle as the current historical optimal position pi(0) Meanwhile, the population size S is set to 40, the maximum evolution algebra M is set to 30, and the variable dimension D is set to 2.
CN201710964727.5A 2017-10-17 2017-10-17 Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm Active CN109669352B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710964727.5A CN109669352B (en) 2017-10-17 2017-10-17 Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710964727.5A CN109669352B (en) 2017-10-17 2017-10-17 Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm

Publications (2)

Publication Number Publication Date
CN109669352A CN109669352A (en) 2019-04-23
CN109669352B true CN109669352B (en) 2022-04-05

Family

ID=66139708

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710964727.5A Active CN109669352B (en) 2017-10-17 2017-10-17 Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm

Country Status (1)

Country Link
CN (1) CN109669352B (en)

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110716432B (en) * 2019-10-14 2022-03-15 北京工业大学 Multi-objective optimization control method for urban sewage treatment process based on self-adaptive selection strategy
CN111399558B (en) * 2020-04-27 2023-09-22 北京工业大学 Knowledge selection-based multi-objective optimization control method for sewage treatment process
CN111474854B (en) * 2020-04-27 2022-05-03 北京工业大学 Sewage treatment process optimization control method based on data-knowledge drive
CN112183719B (en) * 2020-09-15 2024-02-02 北京工业大学 Intelligent detection method for total nitrogen in effluent based on multi-objective optimization-fuzzy neural network
CN112859597A (en) * 2021-01-07 2021-05-28 中南大学 Data-driven optimized control method and system for iron sinking process
CN113589684B (en) * 2021-05-20 2023-11-21 北京工业大学 Sewage treatment process optimization control method based on self-adjusting multi-task particle swarm algorithm
CN113537454A (en) * 2021-06-03 2021-10-22 北京工业大学 Effluent ammonia nitrogen soft measurement method based on ADw-CLPSO radial basis affective neural network
CN113867276B (en) * 2021-08-27 2024-07-09 北京工业大学 Sewage treatment process multitask optimization control method based on self-adaptive knowledge migration strategy
CN114114909B (en) * 2021-11-11 2024-03-22 海南师范大学 Intermittent process 2D output feedback prediction control method based on particle swarm optimization
CN115356930B (en) * 2022-08-24 2024-04-19 东北大学 Multi-objective optimization control system and method in sewage treatment process

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102917441A (en) * 2012-09-29 2013-02-06 北京邮电大学 Target network selection method on basis of particle swarm algorithm for multi-mode terminals
CN103049805A (en) * 2013-01-18 2013-04-17 中国测绘科学研究院 Vehicle route optimization method with time window constraint based on improved particle swarm optimization (PSO)
CN105426954A (en) * 2015-08-20 2016-03-23 武汉科技大学 Particle swarm optimization method based on multi-strategy synergistic function
CN105573115A (en) * 2015-12-09 2016-05-11 中山大学 Sewage treatment process energy-saving optimization control method based on quantum genetic algorithm
CN106354014A (en) * 2016-10-27 2017-01-25 北京工业大学 Sewage disposal optimal control method based on multi-objective differential evolution algorithm

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002250237A (en) * 2001-02-22 2002-09-06 Ishikawajima Harima Heavy Ind Co Ltd Adaptation control method for variation of heating value of sludge gas
WO2007037985A2 (en) * 2005-09-23 2007-04-05 Max Rudolf Junghanns Systems and methods for treating water
CN102122134A (en) * 2011-02-14 2011-07-13 华南理工大学 Method and system for wastewater treatment of dissolved oxygen control based on fuzzy neural network
CN103197544B (en) * 2013-02-25 2015-06-17 北京工业大学 Sewage disposal process multi-purpose control method based on nonlinear model prediction
CN104360035B (en) * 2014-11-02 2016-03-30 北京工业大学 A kind of sewage total phosphorus TP flexible measurement method based on self-organization population-radial base neural net
CN105174417A (en) * 2015-10-13 2015-12-23 中国石油化工股份有限公司 Method for pretreating oil-containing sewage in refined oil depot by catalytic oxidation
CN105512745A (en) * 2015-11-05 2016-04-20 天津大学 Wind power section prediction method based on particle swarm-BP neural network
EP3173880A1 (en) * 2015-11-30 2017-05-31 SUEZ Groupe Method for generating control signals adapted to be sent to actuators in a water drainage network
CN105404151B (en) * 2015-12-12 2017-11-24 北京工业大学 Sewage disposal process dynamic multi-objective optimization control method
CN105372995B (en) * 2015-12-17 2019-01-01 镇江市高等专科学校 Sewage disposal system investigating method
CN106352244A (en) * 2016-08-31 2017-01-25 中国石油化工股份有限公司 Pipeline leakage detection method based on hierarchical neural network
CN106698642B (en) * 2016-12-29 2020-02-11 北京工业大学 Multi-target real-time optimization control method in sewage treatment process

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102917441A (en) * 2012-09-29 2013-02-06 北京邮电大学 Target network selection method on basis of particle swarm algorithm for multi-mode terminals
CN103049805A (en) * 2013-01-18 2013-04-17 中国测绘科学研究院 Vehicle route optimization method with time window constraint based on improved particle swarm optimization (PSO)
CN105426954A (en) * 2015-08-20 2016-03-23 武汉科技大学 Particle swarm optimization method based on multi-strategy synergistic function
CN105573115A (en) * 2015-12-09 2016-05-11 中山大学 Sewage treatment process energy-saving optimization control method based on quantum genetic algorithm
CN106354014A (en) * 2016-10-27 2017-01-25 北京工业大学 Sewage disposal optimal control method based on multi-objective differential evolution algorithm

Also Published As

Publication number Publication date
CN109669352A (en) 2019-04-23

Similar Documents

Publication Publication Date Title
CN109669352B (en) Oily sewage treatment process optimization control method based on self-adaptive multi-target particle swarm
Han et al. Dynamic MOPSO-based optimal control for wastewater treatment process
CN108445757B (en) Sewage treatment process optimization control method based on dynamic multi-target particle swarm algorithm
CN115356930B (en) Multi-objective optimization control system and method in sewage treatment process
Zhou et al. Multiobjective operation optimization of wastewater treatment process based on reinforcement self-learning and knowledge guidance
CN106406094B (en) A kind of sewage treatment dissolved oxygen concentration tracking and controlling method based on two type fuzzy neural network of section
CN109711070A (en) A kind of dissolved oxygen concentration optimization method based on activated sludge water process
Wang et al. Adaptive critic control design with knowledge transfer for wastewater treatment applications
Do et al. A design of higher-level control based genetic algorithms for wastewater treatment plants
CN116360366B (en) Sewage treatment process optimization control method
Vlad et al. Indirect control of substrate concentration for a wastewater treatment process by dissolved oxygen tracking
CN113568311A (en) Intelligent sewage treatment optimization control method based on knowledge information
Du et al. Multi-objective pity beetle algorithm based optimal control of wastewater treatment process
Ye et al. Dissolved oxygen control of the activated sludge wastewater treatment process using adaptive fuzzy PID control
Sulaiman et al. Modeling anaerobic process for wastewater treatment: new trends and methodologies
CN111399558B (en) Knowledge selection-based multi-objective optimization control method for sewage treatment process
CN114879487B (en) Sewage treatment process robust model prediction control method based on interval two-type fuzzy neural network
Xu et al. A novel dissolve oxygen control method based on fuzzy neural network
Qiao et al. Offline data-driven adaptive critic design with variational inference for wastewater treatment process control
Swarnalatha Simulation and Analysis of Ammonium-Based Aeration Control Strategies to Enhance Efficiency in Wastewater Treatment Plant
Du et al. Dissolved oxygen concentration control in wastewater treatment process based on reinforcement learning
Qiao et al. Action-Dependent Heuristic Dynamic Programming With Experience Replay for Wastewater Treatment Processes
Zhang et al. Multi-Objective Optimal Control of Wastewater Treatment Processes Constrainted by Energy-Saving and Effluent Qualities
Bai et al. Bi-Level Optimal Control Based on Similarity Clustering for Wastewater Treatment Process with Greenhouse Gas Emissions
Zhang et al. Dynamic Multiobjective Optimal Control with Knowledge-Decision for Wastewater Treatment Process

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
TA01 Transfer of patent application right

Effective date of registration: 20220316

Address after: Yanan City, Shandong province Qingdao City three road 266071 No. 218

Applicant after: CHINA PETROLEUM & CHEMICAL Corp.

Applicant after: Sinopec Safety Engineering Research Institute Co., Ltd

Address before: Yanan City, Shandong province Qingdao City three road 266071 No. 218

Applicant before: CHINA PETROLEUM & CHEMICAL Corp.

Applicant before: Qingdao Safety Engineering Research Institute of Sinopec

TA01 Transfer of patent application right
GR01 Patent grant
GR01 Patent grant