CN106354018B - Tank dissolved oxygen intelligent control system based on RBF neural - Google Patents
Tank dissolved oxygen intelligent control system based on RBF neural Download PDFInfo
- Publication number
- CN106354018B CN106354018B CN201611021830.8A CN201611021830A CN106354018B CN 106354018 B CN106354018 B CN 106354018B CN 201611021830 A CN201611021830 A CN 201611021830A CN 106354018 B CN106354018 B CN 106354018B
- Authority
- CN
- China
- Prior art keywords
- dissolved oxygen
- rbf neural
- moment
- hidden layer
- neuron
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
- Activated Sludge Processes (AREA)
Abstract
Tank dissolved oxygen intelligent control system based on RBF neural belongs to water treatment field, and belongs to field of intelligent control.A set of complete tank dissolved oxygen intelligent control system is formed by building for control module, and by building hardware platform；It realizes and is difficult to control with the dissolved oxygen DO that accurately controls in real time to comparing in sewage disposal process.By the tank dissolved oxygen intelligent control system application sewage disposal process based on RBF neural, by accurately controlling to realize and accurately control dissolved oxygen DO concentration to air blower electrically operated valve.Dissolved oxygen intelligent control control is aiming at the problem that traditional switch is controlled and be cannot achieve with PID control to having the characteristics that nonlinearity, strong coupling, time-varying, large time delay and uncertain serious sewage disposal process stabilization and accurately controlling, realize the intelligent control to dissolved oxygen DO concentration, the result shows that the system realizes the control of dissolved oxygen DO concentration, improves wastewater treatment efficiency and reduce energy consumption.
Description
Technical field
The tank dissolved oxygen intelligent control system based on RBF neural is designed in the present invention, by designing RBF neural control
Device processed completes building for control module, and forms a set of complete tank dissolved oxygen intelligent control system by building hardware platform；
It realizes and is difficult to control with the dissolved oxygen DO that accurately controls in real time to comparing in sewage disposal process.It will be based on RBF nerve
The tank dissolved oxygen intelligent control system application sewage disposal process of network passes through accurately controlling to real to air blower electrically operated valve
Now dissolved oxygen DO concentration is accurately controlled.Dissolved oxygen intelligent control control belongs to water treatment field, and belongs to intelligent control
Field.
Background technique
In recent years, being gradually increased with China's urbanization, urban population also sharply increases, industry gradually development to
A large amount of sewage is generated, China actively builds sewage treatment facility, the sewage treatment capacity in city and industrial scene is quickly pushed,
Also up to standard to sewage treatment simultaneously to be made that stringent regulation.However, existing sewage treatment plant, which faces, guarantees that sewage treatment reaches
Target serious challenge.Therefore, it is not high to need to be improved emission compliance rate for sewage treatment plant, and pollutant concentration removal not enough etc. is still
Major issue in sewage disposal process, especially to the processing of industrial wastewater.
Since dissolved oxygen plays an important role in sewage disposal process, most of existing control skill of sewage treatment plant
Art, such as switch control, PID control etc. have certain defect, can not be to non-linear, the features such as multivariable, large time delay
Dissolved oxygen DO concentration in sewage treatment is preferably controlled.The metabolic function that sewage treatment mainly passes through microorganism will
Organic matter degradation is inorganic matter, to realize the removal to pollutant.The type of organic matter has very much, but its general character is exactly micro-
It needs to consume the dissolved oxygen in water under biodegrade, equally, the concentration of oxygen in water also directly influences the growth of microorganism.
Suitable dissolved oxygen DO concentration in aerobic tank is kept, is played the role of to wastewater treatment efficiency vital.
Intelligence control system to dissolved oxygen DO concentration control mainly by adjust aerating system in air blower it is electronic
Valve opening is adjusted, and is got rid of sewage sewage treatment plant and is depended on manually adjusting for artificial experience always, increases simultaneously
The reliability of adjusting reduces accidentalia caused by thinking factor.Compared to simple PID control system, time delay,
The higher problem of energy consumption caused by terms of control fluctuation and excessive redundancy, intelligence control system can be avoided effectively.
The present invention relates to the design and research of the intelligence control system based on RBF neural controller, the control systems
System is based primarily upon RBF neural controller, is accurately calculated the control amount of dissolved oxygen and the system by developing will
Control signal is accurately transmitted to executing agency and completes control, solves artificial experience and traditional PID control is insoluble asks
Topic.By take building data acquisition, data transmission and the hardware platforms such as air compressor control and communication network realize data acquisition,
Transmit and control issuing and executing for signal.It is integrated by the exploitation to each functional module, form dissolved oxygen intelligent control
System, improves the stability and reliability of control, while having ensured effluent quality and having reduced consumption and reduce artificial
Interference and operator bring operating cost of the factor to control process.
Summary of the invention
The present invention obtains the intelligence control system based on RBF neural controller, devises the RBF nerve for control
Network controller solves the control problem in sewage disposal process and is carried out modularized encapsulation；Building data acquisition is taken, number
According to processing and storage and the hardware system of control function module, the efficient operation of system ensure that；It is controlled by the system
It makes, dissolved oxygen concentration can reach most preferably in sewage, solve the problems, such as that dissolved oxygen is difficult to control accurately in sewage disposal process；
Meanwhile, it is capable to provide man-machine interface abundant, simplified control adjusts operation；The stability and reality of sewage disposal process are ensured
On-line Control is showed；
Present invention employs the following technical solution and realize step:
Tank dissolved oxygen intelligent control system based on RBF neural
(1) tank dissolved oxygen intelligent control system designs, and this system based on RBF neural network algorithm mainly by being provided in line traffic control
Decision processed, to cope with the non-linear of sewage disposal process, big time-varying, large time delay and close coupling；For the spy of dissolved oxygen DO control
Point devises the hardware system of control system, including detection instrument, electrical equipment, data acquisition, data processing and storage, control
The several main functional modules of function modoularization processed, are implemented as follows:
Field instrument includes dissolved oxygen sensing instrument, temperature measuring set, PH measuring instrument and COD analyzer and NH4- N analysis
Instrument；On-site test instrument is connected with PLC, PLC and data processing and memory module by RS232 and RS485 in fieldbus into
Row communication, data processing are connected with memory module with control function module by communication interface, and electrical equipment is mainly air blower
And electrically operated valve, it is connected between electrical equipment and PLC；Control function module will be controlled after line provides control strategy by PLC
Signal processed is issued to executing agency's electrically operated valve；
It is living for batch-type interval according to the tank dissolved oxygen intelligent control system required described in (1) based on RBF neural
Property sludge system in dissolved oxygen DO concentration controlled, using air blower aeration quantity as control amount, dissolved oxygen DO concentration be controlled volume；
Characterized by comprising the following steps:
(1) the sewage disposal system prediction model based on RBF neural is designed, prediction model RBF neural is divided into
Three layers: input layer, hidden layer and output layer；The input of prediction model RBF neural is u (k)=[u1(k),u2(k)]T, u2(k)
=u1(k-1), u1It (k) is k moment dissolved oxygen DO concentration control amount, u1It (k-1) is the dissolved oxygen DO concentration control amount at k-1 moment, T
For the transposition of matrix；The output of prediction model RBF neural is dissolved oxygen DO concentration prediction value；Its calculation is as follows:
1. initialization prediction model RBF neural: determining the connection type of neural network 2-P-1, i.e. input layer nerve
Member is 2, and hidden layer neuron is P, and P is the positive integer greater than 2；Output layer neuron is 1；Prediction model RBF nerve
The connection weight of network input layer to hidden layer is 1, the connection weight of hidden layer and output interlayer carry out in [0,1] range with
Machine assignment；The output of neural network is expressed as follows:
Wherein, ymIt (k) is the output of k moment prediction model RBF neural, wj(k) for j-th of neuron of hidden layer and
The connection weight of output layer, j=1,2 ..., P；fjIt is the output of prediction model RBF neural j-th of neuron of hidden layer,
Its calculation formula is:
Wherein, μj(k) j-th of neuronal center value of k moment hidden layer, σ are indicatedj(k) j-th of mind of k moment hidden layer is indicated
Center width through member；
2. defining the performance indicator J of prediction model RBF neuralm(k)
em(k)=y (k)-ym(k)
(4)
Wherein, y (k) is the dissolved oxygen DO concentration value of k moment actual measurement, emIt (k) is k moment dissolved oxygen DO concentration value
Error；
3. being updated to the parameter of prediction model RBF neural
wj(k+1)=wj(k)-ηΔwj(k)=wj(k)+ηem(k)fj(u(k))(1-y(k))y(k) (6)
Wherein, Δ wjIt (k) is the correction amount of k moment j-th of hidden layer neuron and output layer neuron connection weight, wj
It (k) is the connection weight of k moment j-th of hidden layer neuron and output layer neuron, wj(k+1) implicit for j-th of the k+1 moment
The connection weight of layer neuron and output layer neuron, μj(k+1) j-th of neuronal center value of k+1 moment hidden layer, σ are indicatedj
(k+1) center width of j-th of neuron of k+1 moment hidden layer is indicated；η is learning rate, η ∈ (0,1]；
4. the size of the objective function of current time dissolved oxygen DO prediction is judged, if Jm(k) > 0.01, then repeatedly step
③；If Jm(k) < 0.01 1. output y that step calculates prediction model RBF neural, is then gone tom(k)；
(2) the RBF neural controller designed for control；X (k)=[x1(k),x2(k)]TFor RBF neural control
The input of device processed, x1It (k) is the error of k moment dissolved oxygen DO concentration set point and actual value, x2(k) dense for k moment dissolved oxygen DO
Spend the change rate of setting value and actual value error；
1. initialization RBF neural controller: determining the connection type of neural network 2-M-1, i.e. input layer
It is 2, hidden layer neuron is M, and M is the positive integer greater than 2；Output layer neuron is 1；RBF neural controller
The connection weight of input layer to hidden layer is 1, and the connection weight of hidden layer and output interlayer is assigned at random in [0,1] range
Value；The output of neural network is expressed as follows:
Wherein, u (k) is the output of k moment RBF neural controller, wi c(k) implicit for RBF neural controller
The connection weight of layer i-th of neuron and output layer, i=1,2 ..., M；fiIt is i-th of neuron of RBF neural hidden layer
Output, its calculation formula is:
Wherein, μi c(k) i-th of neuronal center value of k moment RBF neural controller hidden layer, σ are indicatedi c(k) table
Show the center width of k moment RBF neural controller i-th of neuron of hidden layer；
2. defining the index J of RBF neural controllerc(k)
E (k)=r (k)-y (k)
(12)
Wherein, e (k) is the error of k moment dissolved oxygen DO concentration, and r (k) is k moment dissolved oxygen DO concentration set point；
3. being updated to the parameter of RBF neural controller
Wherein, Δ wi c(k) connect for i-th of neuron of k moment RBF neural controller hidden layer and output layer neuron
Meet the correction amount of weight, wi cIt (k+1) is i-th neuron of RBF neural controller hidden layer at k+1 moment and output layer nerve
The connection weight of member；μi c(k+1) i-th of neuronal center value of k+1 moment RBF neural controller hidden layer, σ are indicatedi c(k
+ 1) center width of k+1 moment RBF neural controller i-th of neuron of hidden layer is indicated；η1For learning rate, η1∈(0,
1]；
4. the size of the objective function of current time dissolved oxygen DO is judged, if Jc(k) > 0.01, then repeatedly step is 3.；Such as
Fruit Jc(k) < 0.01 1. output u (k) that step calculates RBF neural controller, is then gone to；
(3) dissolved oxygen DO is controlled using the u (k) solved, u (k) be aeration quantity i.e. control amount at the k moment, control
The output of system processed is the concentration value of practical dissolved oxygen DO.
Creativeness of the invention is mainly reflected in:
(1) present invention is a mistake with non-linear, close coupling, big time-varying for current sewage disposal process
Journey needs to control dissolved oxygen DO concentration in a reasonable range, however according to the existing control method of sewage treatment plant, it is difficult
Stablize and be accurately controlled to realize；There is very strong adaptive and self-learning capability according to neural network, devise RBF nerve
Network Prediction Model and RBF neural controller, realize the On-line Control of dissolved oxygen, have stability good, real-time is good
And control the features such as precision is high；
(2) present invention devises RBF neural prediction model and RBF neural controller, and control method is preferably
It solves the problems, such as that nonlinear system is difficult to control, realizes the real-time accurate control of dissolved oxygen concentration；Solves complicated dirt
Water treatment procedure, which only relies on, solves artificial experience realization control problem, has the features such as low energy consumption, and structure is simple；
Detailed description of the invention
Fig. 1 is control system architecture figure of the present invention
Fig. 2 is control system model figure of the present invention
Fig. 3 is neural net model establishing of the present invention and controller structure diagram
Fig. 4 is RBF neural network structure of the present invention
Fig. 5 is control system dissolved oxygen DO concentration results figure of the present invention
Fig. 6 is control system dissolved oxygen DO concentration error figure of the present invention
Specific embodiment
The present invention obtains the tank dissolved oxygen intelligent control system based on RBF neural, realizes molten in sewage disposal process
Solve the accurate control of oxygen DO concentration；It is controlled by the system, dissolved oxygen concentration can reach most preferably in sewage, solve dirt
The problem of dissolved oxygen is difficult to control accurately in water treatment procedure improves the precision of dissolved oxygen DO concentration control；Meanwhile it simplifying
The operating process of control realizes automatic on-line control；
Present invention employs the following technical solution and realize step:
(1) tank dissolved oxygen intelligent control system designs, and this system based on RBF neural network algorithm mainly by being provided in line traffic control
Decision processed devises the hardware system of control system, as Fig. 1 gives the structure chart of control system, including it is detection instrument, electrical
Equipment, data acquisition, data processing and storage, the several main functional modules of control function modularization；
Field instrument includes dissolved oxygen sensing instrument, temperature measuring set, PH measuring instrument and COD analyzer and NH4- N analysis
Instrument；On-site test instrument is connected with PLC, and PLC and data processing and memory module are carried out by fieldbus RS232 and RS485
Communication, data processing is connected with memory module with control function module by communication interface, electrical equipment be mainly air blower with
And electrically operated valve, it is connected between electrical equipment and PLC；Control function module will be controlled after line provides control strategy by PLC
Signal is issued to executing agency's electrically operated valve；
(2) in Control System Design, design RBF neural prediction model and design RBF neural controller are embedding
Enter in control module, control strategy is provided online, the model of control system is given in Fig. 2, simply describes control system
The basic function having.
Present invention obtains a kind of neural network dissolved oxygen DO concentration control method based on gradient descent algorithm, realizes
The accurate control of dissolved oxygen DO concentration in sewage disposal process；This method is by the method based on data-driven and gradient decline
Solve the control problem in sewage disposal process；After being controlled by this method, dissolved oxygen concentration can reach most in sewage
It is good, it solves the problems, such as that dissolved oxygen is difficult to control accurately in sewage disposal process, improves the precision of dissolved oxygen DO concentration control；
Meanwhile it ensureing the stability of sewage disposal process and having realized On-line Control；
Present invention employs the following technical solution and realize step:
A kind of dissolved oxygen accuracy control method based on RBF neural,
It is controlled for dissolved oxygen DO concentration in batch-type interval activated Sludge System, is control with air blower aeration quantity
Amount, dissolved oxygen DO concentration are controlled volume, control structure figure such as Fig. 3；
(1) the sewage disposal system prediction model based on RBF neural is designed, prediction model RBF neural is divided into
Three layers: input layer, hidden layer and output layer；Predict that the input of mould RBF neural is u (k)=[u1(k),u2(k)]T, u2(k)=
u1(k-1), u1It (k) is k moment dissolved oxygen DO concentration control amount, u1It (k-1) is the dissolved oxygen DO concentration control amount at k-1 moment, T is
The transposition of matrix；The output of prediction model RBF neural is dissolved oxygen DO concentration prediction value；Its calculation is as follows:
1. initialization prediction model RBF neural: determining the connection type of neural network 2-P-1, i.e. input layer nerve
Member is 2, and hidden layer neuron is that P is 15；Output layer neuron is 1；Prediction model RBF neural input layer is to hidden
Connection weight containing layer is 1, and the connection weight of hidden layer and output interlayer carries out random assignment in [0,1] range；Nerve net
The output of network is expressed as follows:
Wherein, ymIt (k) is the output of k moment prediction model RBF neural, wj(k) for j-th of neuron of hidden layer and
The connection weight of output layer, j=1,2 ..., P；fjIt is the output of prediction model RBF neural j-th of neuron of hidden layer,
Its calculation formula is:
Wherein, μj(k) j-th of neuronal center value of k moment hidden layer, σ are indicatedj(k) j-th of mind of k moment hidden layer is indicated
Center width through member；
2. defining the performance indicator J of prediction model RBF neuralm(k)
em(k)=y (k)-ym(k)
(20)
Wherein, y (k) is the dissolved oxygen DO concentration value of k moment actual measurement, emIt (k) is k moment dissolved oxygen DO concentration value
Error；
3. being updated to the parameter of prediction model RBF neural
wj(k+1)=wj(k)-ηΔwj(k)=wj(k)+ηem(k)fj(u(k))(1-y(k))y(k) (22)
Wherein, Δ wjIt (k) is the correction amount of k moment j-th of hidden layer neuron and output layer neuron connection weight, wj
It (k) is the connection weight of k moment j-th of hidden layer neuron and output layer neuron, wj(k+1) implicit for j-th of the k+1 moment
The connection weight of layer neuron and output layer neuron, μj(k+1) j-th of neuronal center value of k+1 moment hidden layer, σ are indicatedj
(k+1) center width of j-th of neuron of k+1 moment hidden layer is indicated；Learning rate η=0.1；
4. the size of the objective function of current time dissolved oxygen DO prediction is judged, if Jm(k) > 0.01, then repeatedly step
③；If Jm(k) < 0.01 1. output y that step calculates prediction model RBF neural, is then gone tom(k)；
(2) the RBF neural controller designed for control；X (k)=[x1(k),x2(k)]TFor RBF neural control
The input of device processed, x1It (k) is the error of k moment dissolved oxygen DO concentration set point and actual value, x2(k) dense for k moment dissolved oxygen DO
Spend the change rate of setting value and actual value error；
1. initialization RBF neural controller: determining the connection type of neural network 2-M-1, i.e. input layer
It is 2, hidden layer neuron is that M is 17；Output layer neuron is 1；RBF neural controller input layer is to hidden layer
Connection weight be 1, the connection weight of hidden layer and output interlayer carries out random assignment in [0,1] range；Neural network
Output is expressed as follows:
Wherein, u (k) is the output of k moment RBF neural controller, wi c(k) implicit for RBF neural controller
The connection weight of layer i-th of neuron and output layer, i=1,2 ..., M；fiIt is i-th of neuron of RBF neural hidden layer
Output, its calculation formula is:
Wherein, μi c(k) i-th of neuronal center value of k moment RBF neural controller hidden layer, σ are indicatedi c(k) table
Show the center width of k moment RBF neural controller i-th of neuron of hidden layer；
2. defining the index J of RBF neural controllerc(k)
E (k)=r (k)-y (k)
(28)
Wherein, e (k) is the error of k moment dissolved oxygen DO concentration, and r (k) is k moment dissolved oxygen DO concentration set point；
3. being updated to the parameter of RBF neural controller
Wherein, Δ wi c(k) connect for i-th of neuron of k moment RBF neural controller hidden layer and output layer neuron
Meet the correction amount of weight, wi cIt (k+1) is i-th neuron of RBF neural controller hidden layer at k+1 moment and output layer nerve
The connection weight of member；μi c(k+1) i-th of neuronal center value of k+1 moment RBF neural controller hidden layer, σ are indicatedi c(k
+ 1) center width of k+1 moment RBF neural controller i-th of neuron of hidden layer is indicated；Learning rate, η1=0.1；
4. the size of the objective function of current time dissolved oxygen DO is judged, if Jc(k) > 0.01, then repeatedly step is 3.；Such as
Fruit Jc(k) < 0.01 1. output u (k) that step calculates RBF neural controller, is then gone to；
(3) dissolved oxygen DO is controlled using the u (k) solved, u (k) be aeration quantity i.e. control amount at the k moment, control
The output of system processed is the concentration value of practical dissolved oxygen DO；The dissolved oxygen DO concentration value of Fig. 5 display system, X-axis: time, unit
It is 15 minutes/sample, Y-axis: dissolved oxygen DO concentration, unit are mg/litres, and solid line is desired dissolved oxygen DO concentration value, and dotted line is
Practical dissolved oxygen DO exports concentration value；Error such as Fig. 6 of reality output dissolved oxygen DO concentration and desired dissolved oxygen DO concentration, X-axis:
Time, unit are 15 minutes/samples, and Y-axis: dissolved oxygen DO concentration error value, unit is mg/litre, as a result prove this method
Validity.
Claims (1)
1. the tank dissolved oxygen intelligent control system based on RBF neural, it is characterised in that:
Hardware includes detecting instrument, electrical equipment, data acquisition, data processing and memory module, control function module, specific real
It is now as follows:
Detecting instrument includes dissolved oxygen sensing instrument, temperature measuring set, PH measuring instrument and COD analyzer and NH4- N analyzer；Inspection
It surveys instrument to be connected with PLC, PLC is communicated with memory module by RS232 in fieldbus and RS485 with data processing, number
It is connected with memory module with control function module by communication interface according to processing, electrical equipment includes air blower and motor-driven valve
Door, connects between electrical equipment and PLC；Control function module issues after line provides control strategy, through PLC by signal is controlled
To executing agency's electrically operated valve；
Dissolved oxygen DO concentration in batch-type interval activated Sludge System is controlled, it is molten using air blower aeration quantity as control amount
Solution oxygen DO concentration is controlled volume；
The following steps are included:
(2) the sewage disposal system prediction model based on RBF neural is designed, prediction model RBF neural is divided into three layers:
Input layer, hidden layer and output layer；The input of prediction model RBF neural is u (k)=[u1(k),u2(k)]T, u2(k)=u1
(k-1), u1It (k) is k moment dissolved oxygen DO concentration control amount, u1It (k-1) is the dissolved oxygen DO concentration control amount at k-1 moment, T is square
The transposition of battle array；The output of prediction model RBF neural is dissolved oxygen DO concentration prediction value；Its calculation is as follows:
1. initialization prediction model RBF neural: determining the connection type of neural network 2-P-1, i.e. input layer is 2
A, hidden layer neuron is P, and P is the positive integer greater than 2；Output layer neuron is 1；Prediction model RBF neural is defeated
The connection weight for entering layer to hidden layer is 1, and the connection weight of hidden layer and output interlayer is assigned at random in [0,1] range
Value；The output of neural network is expressed as follows:
Wherein, ymIt (k) is the output of k moment prediction model RBF neural, wjIt (k) is j-th of neuron of hidden layer and output
The connection weight of layer, j=1,2 ..., P；fjIt is the output of prediction model RBF neural j-th of neuron of hidden layer, meter
Calculate formula are as follows:
Wherein, μj(k) j-th of neuronal center value of k moment hidden layer, σ are indicatedj(k) j-th of neuron of k moment hidden layer is indicated
Center width；
2. defining the performance indicator J of prediction model RBF neuralm(k)
em(k)=y (k)-ym(k) (4)
Wherein, y (k) is the dissolved oxygen DO concentration value of k moment actual measurement, emIt (k) is the error of k moment dissolved oxygen DO concentration value；
3. being updated to the parameter of prediction model RBF neural
wj(k+1)=wj(k)-ηΔwj(k)=wj(k)+ηem(k)fj(u(k))(1-y(k))y(k) (6)
Wherein, Δ wjIt (k) is the correction amount of k moment j-th of hidden layer neuron and output layer neuron connection weight, wj(k) it is
The connection weight of k moment j-th of hidden layer neuron and output layer neuron, wjIt (k+1) is j-th of hidden layer mind of k+1 moment
Connection weight through member with output layer neuron, μj(k+1) j-th of neuronal center value of k+1 moment hidden layer, σ are indicatedj(k+1)
Indicate the center width of j-th of neuron of k+1 moment hidden layer；η is learning rate, η ∈ (0,1]；
4. the size of the objective function of current time dissolved oxygen DO prediction is judged, if Jm(k) > 0.01, then repeatedly step is 3.；Such as
Fruit Jm(k) < 0.01 1. output y that step calculates prediction model RBF neural, is then gone tom(k)；
(3) the RBF neural controller designed for control；X (k)=[x1(k),x2(k)]TFor RBF neural controller
Input, x1It (k) is the error of k moment dissolved oxygen DO concentration set point and actual value, x2(k) it is set for k moment dissolved oxygen DO concentration
The change rate of definite value and actual value error；
1. initialization RBF neural controller: determining the connection type of neural network 2-M-1, i.e. input layer is 2
A, hidden layer neuron is M, and M is the positive integer greater than 2；Output layer neuron is 1；The input of RBF neural controller
The connection weight of layer to hidden layer is 1, and the connection weight of hidden layer and output interlayer carries out random assignment in [0,1] range；
The output of neural network is expressed as follows:
Wherein, u (k) is the output of k moment RBF neural controller, wi cIt (k) is RBF neural controller hidden layer i-th
The connection weight of a neuron and output layer, i=1,2 ..., M；fiIt is the defeated of i-th of neuron of RBF neural hidden layer
Out, its calculation formula is:
Wherein, μi c(k) i-th of neuronal center value of k moment RBF neural controller hidden layer, σ are indicatedi c(k) when indicating k
Carve the center width of i-th of neuron of RBF neural controller hidden layer；
2. defining the index J of RBF neural controllerc(k)
E (k)=r (k)-y (k) (12)
Wherein, e (k) is the error of k moment dissolved oxygen DO concentration, and r (k) is k moment dissolved oxygen DO concentration set point；
3. being updated to the parameter of RBF neural controller
Wherein, Δ wi cIt (k) is i-th of neuron of k moment RBF neural controller hidden layer and output layer neuron connection weight
The correction amount of value, wi cIt (k+1) is RBF neural controller hidden layer i-th neuron at k+1 moment and output layer neuron
Connection weight；μi c(k+1) i-th of neuronal center value of k+1 moment RBF neural controller hidden layer, σ are indicatedi c(k+1)
Indicate the center width of k+1 moment RBF neural controller i-th of neuron of hidden layer；η1For learning rate, η1∈(0,1]；
4. the size of the objective function of current time dissolved oxygen DO is judged, if Jc(k) > 0.01, then repeatedly step is 3.；If Jc
(k) < 0.01 1. output u (k) that step calculates RBF neural controller, is then gone to；
(4) dissolved oxygen DO is controlled using the u (k) solved, u (k) be aeration quantity i.e. control amount at the k moment, control system
The output of system is the concentration value of practical dissolved oxygen DO.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611021830.8A CN106354018B (en) | 2016-11-21 | 2016-11-21 | Tank dissolved oxygen intelligent control system based on RBF neural |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611021830.8A CN106354018B (en) | 2016-11-21 | 2016-11-21 | Tank dissolved oxygen intelligent control system based on RBF neural |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106354018A CN106354018A (en) | 2017-01-25 |
| CN106354018B true CN106354018B (en) | 2019-03-22 |
Family
ID=57862094
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611021830.8A Expired - Fee Related CN106354018B (en) | 2016-11-21 | 2016-11-21 | Tank dissolved oxygen intelligent control system based on RBF neural |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106354018B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107390546A (en) * | 2017-07-31 | 2017-11-24 | 广东工业大学 | Piezoelectric Driving locating platform modeling method, control method and system based on EOS ELM |
| CN107505840A (en) * | 2017-07-31 | 2017-12-22 | 广东工业大学 | Piezoelectric Driving FTS modeling methods, control method and system based on FReOS ELM |
| CN109978024B (en) * | 2019-03-11 | 2020-10-27 | 北京工业大学 | Effluent BOD prediction method based on interconnected modular neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1376276A1 (en) * | 2002-06-21 | 2004-01-02 | H2L Co., Ltd | An AI based control system and method for treating sewage/waste water by means of a neural network and a back-propagation algorithm |
| CN101576734A (en) * | 2009-06-12 | 2009-11-11 | 北京工业大学 | Dissolved oxygen control method based on dynamic radial basis function neural network |
| CN102411308A (en) * | 2011-12-24 | 2012-04-11 | 北京工业大学 | Self-adaptive control method of dissolved oxygen based on recurrent neural network model |
| CN103064290A (en) * | 2013-01-01 | 2013-04-24 | 北京工业大学 | Dissolved oxygen model prediction control method based on self-organization radial basis function neural network |
| CN103499982A (en) * | 2013-09-30 | 2014-01-08 | 北京工业大学 | Self-organization control method of sewage treatment process |
-
2016
- 2016-11-21 CN CN201611021830.8A patent/CN106354018B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1376276A1 (en) * | 2002-06-21 | 2004-01-02 | H2L Co., Ltd | An AI based control system and method for treating sewage/waste water by means of a neural network and a back-propagation algorithm |
| CN101576734A (en) * | 2009-06-12 | 2009-11-11 | 北京工业大学 | Dissolved oxygen control method based on dynamic radial basis function neural network |
| CN102411308A (en) * | 2011-12-24 | 2012-04-11 | 北京工业大学 | Self-adaptive control method of dissolved oxygen based on recurrent neural network model |
| CN103064290A (en) * | 2013-01-01 | 2013-04-24 | 北京工业大学 | Dissolved oxygen model prediction control method based on self-organization radial basis function neural network |
| CN103499982A (en) * | 2013-09-30 | 2014-01-08 | 北京工业大学 | Self-organization control method of sewage treatment process |
Non-Patent Citations (1)
| Title |
|---|
| 基于RBF神经网络的出水氨氮预测研究;乔俊飞 等;《控制工程》;20160930;第23卷(第9期);第1301-1305页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106354018A (en) | 2017-01-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2021101438A4 (en) | Adaptive control method and system for aeration process | |
| US11709463B2 (en) | Control method based on adaptive neural network model for dissolved oxygen of aeration system | |
| CN104360035B (en) | A kind of sewage total phosphorus TP flexible measurement method based on self-organization population-radial base neural net | |
| CN101833314B (en) | Sewage treatment control system and sewage treatment control method | |
| CN106354018B (en) | Tank dissolved oxygen intelligent control system based on RBF neural | |
| CN102411308B (en) | Adaptive control method of dissolved oxygen (DO) based on recurrent neural network (RNN) model | |
| CN102262147A (en) | Soft measurement method and system for effluent chemical oxygen demand (COD) of waste water treatment system | |
| CN109133351A (en) | Membrane bioreactor-MBR fouling membrane intelligent early-warning method | |
| CN102122134A (en) | Method and system for wastewater treatment of dissolved oxygen control based on fuzzy neural network | |
| CN103728431A (en) | Industrial sewage COD (chemical oxygen demand) online soft measurement method based on ELM (extreme learning machine) | |
| CN110378533A (en) | A kind of intelligence aeration management method based on big data analysis | |
| CN108898215A (en) | A kind of sludge bulking INTELLIGENT IDENTIFICATION method based on two type fuzzy neural networks | |
| CN107402586A (en) | Dissolved Oxygen concentration Control method and system based on deep neural network | |
| CN106682316A (en) | Real-time effluent total-phosphorus monitoring system based on peak radial basis function neural network | |
| KR102311657B1 (en) | Smart management system for wastewater treatment | |
| CN103632032A (en) | Effluent index online soft measurement prediction method in urban sewage treatment process | |
| CN106354019B (en) | A kind of dissolved oxygen accuracy control method based on RBF neural | |
| CN106096730A (en) | A kind of intelligent detecting method of MBR film permeability rate based on Recurrent RBF Neural Networks | |
| CN113837364A (en) | Sewage treatment soft measurement method and system based on residual error network and attention mechanism | |
| CN118210237B (en) | Intelligent dosing control system of integrated sewage treatment equipment | |
| CN103809436A (en) | Method for intelligent modeling sewage disposal process using activated sludge process | |
| Wu et al. | Data-driven intelligent warning method for membrane fouling | |
| CN106706491B (en) | Intelligent detection method for membrane bioreactor-MBR water permeability | |
| CN113448245A (en) | Deep learning-based dissolved oxygen control method and system in sewage treatment process | |
| CN106769748B (en) | Intelligent detection system for water permeability of membrane bioreactor-MBR (Membrane bioreactor) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190322 Termination date: 20211121 |