CN104612898B - A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method - Google Patents

A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method Download PDF

Info

Publication number
CN104612898B
CN104612898B CN201410706827.4A CN201410706827A CN104612898B CN 104612898 B CN104612898 B CN 104612898B CN 201410706827 A CN201410706827 A CN 201410706827A CN 104612898 B CN104612898 B CN 104612898B
Authority
CN
China
Prior art keywords
mrow
msub
msup
output
neural network
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
Application number
CN201410706827.4A
Other languages
Chinese (zh)
Other versions
CN104612898A (en
Inventor
李泰�
侯小燕
石铭霄
潘庭龙
吴定会
朱志宇
王媛媛
张福特
于唯楚
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 E Tech Ningbo Maritime Electronics Research Institute Co ltd
Original Assignee
Jiangsu University of Science and Technology
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 Jiangsu University of Science and Technology filed Critical Jiangsu University of Science and Technology
Priority to CN201410706827.4A priority Critical patent/CN104612898B/en
Publication of CN104612898A publication Critical patent/CN104612898A/en
Application granted granted Critical
Publication of CN104612898B publication Critical patent/CN104612898B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03D—WIND MOTORS
    • F03D7/00—Controlling wind motors 
    • F03D7/02—Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor
    • F03D7/022—Adjusting aerodynamic properties of the blades
    • F03D7/0224—Adjusting blade pitch
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03D—WIND MOTORS
    • F03D7/00—Controlling wind motors 
    • F03D7/02—Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor
    • F03D7/04—Automatic control; Regulation
    • F03D7/042—Automatic control; Regulation by means of an electrical or electronic controller
    • F03D7/043—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic
    • F03D7/044—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic with PID control
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03D—WIND MOTORS
    • F03D7/00—Controlling wind motors 
    • F03D7/02—Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor
    • F03D7/04—Automatic control; Regulation
    • F03D7/042—Automatic control; Regulation by means of an electrical or electronic controller
    • F03D7/043—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic
    • F03D7/046—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic with learning or adaptive control, e.g. self-tuning, fuzzy logic or neural network
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/70—Type of control algorithm
    • F05B2270/707—Type of control algorithm fuzzy logic
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/70—Type of control algorithm
    • F05B2270/709—Type of control algorithm with neural networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/70—Wind energy
    • Y02E10/72—Wind turbines with rotation axis in wind direction

Landscapes

  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Sustainable Development (AREA)
  • Sustainable Energy (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Mechanical Engineering (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Fluid Mechanics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Control Of Eletrric Generators (AREA)
  • Wind Motors (AREA)

Abstract

The present invention relates to a kind of wind electricity change paddle away from Multivariable Fuzzy NN-PID Control Method, it comprises the following steps：Module is adjusted using fuzzy parameter, and pre-tuning is carried out to the weights of PID neural network modules；Error between wind-driven generator speed reference and actual speed output is obtained into wind-driven generator torque reference output quantity after the calculating of PID computing modules；The error and error rate of wind-driven generator power output value and value and power reference are obtained to the pre-tuning parameter of PID neural network module weights after fuzzy parameter adjusts the adjusting of module；By the weights of the negative gradient Algorithm for Training PID neural network modules with factor of momentum, the output of regulation wind-driven generator torque reference value and the output of propeller pitch angle reference value.The present invention can realize that wind driven generator output power is stable near rated value, it is ensured that the safety of blower fan.

Description

Wind power variable pitch multivariable fuzzy neural network PID control method
Technical Field
The invention relates to a wind power variable pitch multivariable fuzzy neural network PID control method, in particular to a wind power variable pitch multivariable fuzzy neural network PID control method which is suitable for a control method of a double-fed wind turbine generator variable pitch and belongs to the technical field of wind power control.
Background
The wind energy is green renewable energy, the proportion of the wind energy in the green energy is increased year by year, and the development and the utilization of the wind energy have wide commercial prospect. When the wind speed is higher than the rated value, great attention is paid to how to effectively control the pitch system so as to reduce the power fluctuation and the mechanical fatigue of the wind turbine. The more commonly used methods include PI control and LQG control. The PI control adjusts the pitch angle through the error between the actual rotating speed value and the reference rotating speed value of the wind driven generator, however, the PI control method needs a large number of parameters for off-line training, and the control precision of the system is greatly reduced. The LQG control cannot guarantee the global gradual stabilization of the system, and the calculation amount of the control system is increased.
In recent years, neural networks have unique advantages when approaching to be responsible for nonlinear systems, and are increasingly widely applied in the field of industrial control. Common neural networks mainly include: BP neural network, RBF neural network, Hopefield neural network, and neural network have been successfully applied to the field of wind power systems. The parallelism of the neural network and the independence of the neural network independent of the mathematical model provide an effective scheme for solving the control problem in the wind power field. How to realize effective control of wind power variable pitch is still a technical problem of wind power systems.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a wind power variable pitch multivariable fuzzy neural network PID control method which is simple and convenient to control, can realize effective control of wind power variable pitch, reduces the cost and increases the running safety of a fan.
According to the technical scheme provided by the invention, the wind power pitch-variable multivariable fuzzy neural network PID control method is characterized in that the wind power pitch-variable multivariable fuzzy neural network comprises a PID calculation module and a fuzzy neural network PID module, and the fuzzy neural network PID module comprises a fuzzy parameter setting module and a PID neural network module;
the control method comprises the following steps:
a. pre-setting the weight of the PID neural network module by using a fuzzy parameter setting module;
b. calculating an error between a rotating speed reference value and actual rotating speed output of the wind driven generator by a PID (proportion integration differentiation) calculation module to obtain a torque reference output quantity of the wind driven generator;
c. the error and the error change rate of the power output value and the power reference value of the wind driven generator are set by a fuzzy parameter setting module to obtain a pre-setting parameter of the weight of the PID neural network module; and training the weight of the PID neural network module through a negative gradient algorithm with momentum factors, and adjusting the torque reference value output and the pitch angle reference value output of the wind driven generator.
The fuzzy parameter setting module adopts a fuzzy controller.
In the step b, the torque of the wind driven generator is output according to the referenceGrAdopting PID closed-loop control:
Gr＝e(kp+ki/s+kds) (1)
wherein,Gris the wind power generator torque reference output, e is the error between the wind power generator rotating speed reference value and the rotating speed actual output, kpIs a proportionality coefficient, kiIs a differential coefficient, kdIs an integral coefficient. s is a differential operator. In the step a, the PID neural network module includes two input layer neurons, three hidden layer neurons, and one output layer neuron;
the two inputs of the PID neural network input layer are r (k) and y (k), wherein r (k) is a wind power generator power reference value PrAnd y (k) is the actual value P of the power of the wind driven generator.
The state of the input layer neurons is:
u1 j(k)＝net1 j(3)
input layer output of yj 1＝f1(net1 j) Wherein j is 1, 2.
The PID neural network hidden layer comprises three neurons of a proportional element, an integral element and a differential element, and the input weighted sum of each neuron of the hidden layer isi＝1,2,3，wij 2The input weight of the ith neuron of the hidden layer.
The state of the proportional element is as follows:
u2 1(k)＝net2 1(4)
the states of the integrator are:
u2 2(k)＝u2 2(k-1)+net2 2(5)
the states of the differential elements are:
u2 3(k)＝net2 3(k)-net2 3(k-1) (6)
the output of each neuron of the hidden layer is:
wherein i is 1,2, 3.
The output layer of the PID neural network comprises a neuron, and the input weighted sum of the neuron of the output layer isl＝1，wli 3Is the input weight of the ith neuron of the output layer.
The state function of the output layer neurons is:
u3 l(k)＝net3 l(8)
the output of the output layer neurons is yl 3＝f3(net3 l) Where l is 1, the output of the PID neural network is equal to the output layer nerveOutput of the element yl 3。
The input layer output function f1Output layer output function f as tan sig function3Is a purelin function. Selecting an initial weight: initial weight w from input layer to hidden layeri1 2＝+1，wi2 2-1; the initial weight from the hidden layer to the output layer is the proportion k output by the fuzzy controllerfPIntegral kfIDifferential kfDAnd (4) parameters.
In the step c, the initial weight of the PID neural network module is adjusted through an error back propagation algorithm, and the training target of the PID neural network module is as follows:
ytpdesired output for neural network, i.e. wind turbine power reference value Pr；yopAnd (4) the actual output of the neural network, namely the actual power value P of the wind driven generator, wherein m is the number of samples.
And (3) adjusting the weight of the error back propagation process by utilizing a negative gradient algorithm with momentum factors: the learning step length is eta, the learning step is n, the formula (10) is the weight correction from the hidden layer to the output layer,
wherein,
yi 2(k) the output values of the neurons of the hidden layer, m is the sample point, η is the learning step size, where,
equation (13) is the input layer to hidden layer weight correction:
wherein,
yj 1(k) is the output value of each neuron of the input layer, m is the sample point, η is the learning step size, wherein,
wherein sgn is a sign function, wjlNet being the weight from hidden layer to output layeri 2(k) Weighted sum of inputs for hidden layer neurons, u1 i(k) The output of the hidden layer neurons.
The invention has the advantages that:
1. the fuzzy neural network PID control method is used for controlling the wind power variable pitch servo system, so that the pitch angle is efficiently adjusted, the rated output of power is kept, and the running safety of the fan is improved.
2. The weight of the PID neural network is pre-set by using a fuzzy rule, so that the convergence speed of the controller is improved.
3. By adopting the online training of the PID neural network weight, the parameter self-tuning of the PID controller is effectively realized, and the network is prevented from falling into a local extreme value.
Drawings
FIG. 1 is a structural diagram of a wind power pitch controller based on a fuzzy neural network PID of the invention;
FIG. 2 is a diagram of a PID neural network architecture of the present invention;
FIG. 3 is a flow chart of a fuzzy neural network PID controller of the invention;
fig. 4 is a structural diagram of the double-fed wind power system after the double-fed wind power system is connected.
Detailed Description
The invention is further illustrated by the following specific figures and examples.
As shown in fig. 1 to 4: in order to realize effective control of the wind power variable pitch system, the invention realizes the required control on one FPGA chip. The FPGA chip is divided into a PID module and a fuzzy neural network PID module, and specifically, the control method comprises the following steps:
as shown in fig. 1, according to the technical scheme provided by the present invention, a wind power pitch-variable multivariable fuzzy neural network PID control method, the wind power pitch-variable multivariable fuzzy neural network comprises a PID calculation module and a fuzzy neural network PID module, the fuzzy neural network PID module comprises a fuzzy parameter setting module and a PID neural network module;
the control method comprises the following steps:
a. pre-setting the weight of the PID neural network module by using a fuzzy parameter setting module;
b. calculating an error between a rotating speed reference value and actual rotating speed output of the wind driven generator by a PID (proportion integration differentiation) calculation module to obtain a torque reference output quantity of the wind driven generator;
c. the error and the error change rate of the power output value and the power reference value of the wind driven generator are set by a fuzzy parameter setting module to obtain a pre-setting parameter of the weight of the PID neural network module; and training the weight of the PID neural network module through a negative gradient algorithm with momentum factors, and adjusting the torque reference value output and the pitch angle reference value output of the wind driven generator.
The PID neural network module comprises two input layer neurons, three hidden layer neurons and one output layer neuron. The fuzzy parameter setting module comprises a fuzzy controller;
in step a, as shown in FIG. 1, the wind turbine torque is output according to the referenceGrAdopting PID closed-loop control:
Gr＝e(kp+ki/s+kds) (1)
wherein,Gris the wind power generator torque reference output, e is the error between the wind power generator rotating speed reference value and the rotating speed actual output, kpIs a proportionality coefficient, kiIs a differential coefficient, kdIs the integral coefficient and s is the differential operator.
As shown in fig. 2, the input layer of the PID neural network:
the two inputs of the PID neural network input layer are r (k) and y (k), wherein r (k) is a wind power generator power reference value PrAnd y (k) is the actual value P of the power of the wind driven generator.
The state of the input layer neurons is:
u1 j(k)＝net1 j(3)
input layer output of yj 1＝f1(net1 j) Wherein j is 1, 2.
The PID neural network hidden layer comprises three neurons of a proportional element, an integral element and a differential element, and the input weighted sum of each neuron of the hidden layer isi＝1,2,3，wij 2The input weight of the ith neuron of the hidden layer.
The state of the proportional element is as follows:
u2 1(k)＝net2 1(4)
the states of the integrator are:
u2 2(k)＝u2 2(k-1)+net2 2(5)
the states of the differential elements are:
u2 3(k)＝net2 3(k)-net2 3(k-1) (6)
the output of each neuron of the hidden layer is:
wherein i is 1,2, 3.
The output layer of the PID neural network comprises a neuron, and the input weighted sum of the neuron of the output layer isl＝1，wli 3Is the input weight of the ith neuron of the output layer.
The state function of the output layer neurons is
u3 l(k)＝net3 l(8)
The output of the output layer neurons is yl 3＝f3(net3 l) Where l is 1, the output v of the PID neural network is equal to the output y of the output layer neuronsl 3。
The input layer output function f1Output layer output function f as tan sig function3Is a purelin function.
Selecting an initial weight of the PID neural network: initial weight w from input layer to hidden layeri1 2＝+1，wi2 2Is-1. As shown in FIG. 1, the initial weight from the hidden layer to the output layer is the ratio k of the fuzzy controller outputfPIntegral kfIDifferential kfDAnd (4) parameters.
The input of the PID calculation module is the error between the reference value of the rotating speed of the wind driven generator and the actual rotating speed output, and the output of the PID calculation module is the reference output quantity of the torque of the wind driven generator;
in the embodiment of the invention, the fuzzy setting parameter module is used for pre-setting the PID neural network weight through a fuzzy rule.
Referring to FIG. 3, the error and the error change rate of the wind turbine power output value and the power reference value are used as the input of the fuzzy parameter setting module, and the output of the fuzzy parameter setting module, i.e. the ratio kfPIntegral kfIDifferential kfDAnd the parameter is used as a pre-setting parameter of the weight of the PID neural network module. The wind power variable pitch multivariable fuzzy neural network PID controller is realized on an FPGA chip, the weight of a PID neural network module is trained by a gradient descent method with momentum factors, and the torque reference value of the wind driven generator is adjustedarOutput and Pitch Angle reference βrAnd (6) outputting.
In the embodiment of the invention, the error back-propagation algorithm adjusts the weight of the PID neural network, and the network training target is as follows:
ytpdesired output for neural network, i.e. wind turbine power reference value Pr。yopAnd the actual output of the neural network is the actual value P of the power of the wind driven generator. Wherein m is the number of samples.
In the embodiment of the invention, the weight of the error back propagation process of the PID neural network is adjusted by a negative gradient algorithm with momentum factors: the learning step length is eta, the learning step is n, the formula (10) is the weight correction from the hidden layer to the output layer,
wherein,
yi 2(k) the output values of the neurons of the hidden layer, m is the sample point, η is the learning step size, where,
equation (13) is the input layer to hidden layer weight correction:
wherein,
yj 1(k) is the output value of each neuron of the input layer, m is the sample point, η is the learning step size, wherein,
wherein sgn is a sign function, wjlNet being the weight from hidden layer to output layeri 2(k) Weighted sum of inputs for hidden layer neurons, u1 i(k) The output of the hidden layer neurons.
The wind power variable pitch multivariable fuzzy neural network PID control method is characterized by comprising the following steps: input wind driven generator rotating speed omega of wind power variable pitch multivariable fuzzy neural network PID controllerhPower P, output wind power generator torque control quantityGrControl amount of Pitch Angle βrAll realized on the same FPGA chip EP1C6T144C8, the wind driven generator torque control quantity output by the FPGA chipGrPitch angle control βrAnd controlling the wind turbine generator.
The grid-connected double-fed wind power system shown in fig. 4 is composed of a wind wheel, a gear box, a double-fed motor, a rotor side converter, a grid side converter, a capacitor, a transformer, a variable pitch servo system and a power grid. The controller of the grid-connected double-fed wind power system is realized by an F2812DSP chip of TI, the chip comprises an IO port, a QEP3-4 orthogonal pulse coding port (connected with a variable pitch servo system and a rotary encoder), a CAP1-2 capturing port, an A/D port (connected with a voltage and current processing circuit), a PWM port (connected with a driving isolation circuit) and the like, and the vector control of the converter and the calculation of active power and rotating speed are realized.
Input of a wind power variable pitch multivariable fuzzy neural network PID controller, namely the rotating speed omega of the wind power generatorhAnd the power P of the wind driven generator is connected with a digital quantity output port DO of the F2812 DSP.
Wind power generation is carried out by the output of the wind power variable pitch multivariable fuzzy neural network PID controllerAmount of control of machine torqueGrAnd pitch angle control βrAnd is connected to digital input DI of F2812.

Claims (8)

1. A wind power variable pitch multivariable fuzzy neural network PID control method is characterized in that the wind power variable pitch multivariable fuzzy neural network comprises a PID calculation module and a fuzzy neural network PID module, and the fuzzy neural network PID module comprises a fuzzy parameter setting module and a PID neural network module;
the control method comprises the following steps:
a. the fuzzy parameter setting module is used for pre-setting the weight of the PID neural network module, namely, the error and the error change rate of the power output value and the power reference value of the wind driven generator are subjected to setting by the fuzzy parameter setting module to obtain the pre-setting parameter of the weight of the PID neural network module;
b. calculating an error between a rotating speed reference value and actual rotating speed output of the wind driven generator by a PID (proportion integration differentiation) calculation module to obtain a torque reference output quantity of the wind driven generator;
c. and training the weight of the PID neural network module through a negative gradient algorithm with momentum factors, and adjusting the reference value of the pitch angle to output.
2. The wind power pitch-variable multivariable fuzzy neural network PID control method according to claim 1, characterized in that: the fuzzy parameter setting module adopts a fuzzy controller.
3. The wind power pitch-variable multivariable fuzzy neural network PID control method according to claim 1, characterized in that: in the step b, the torque of the wind driven generator is output according to the referenceGrAdopting PID closed-loop control:
Gr＝e(kp+ki/s+kds) (1)
wherein,Gris the wind power generator torque reference output, e is the error between the wind power generator rotating speed reference value and the rotating speed actual output, kpIs a proportionality coefficient, kiIs a differential coefficient, kdIs the integral coefficient and s is the differential operator.
4. The wind power pitch-variable multivariable fuzzy neural network PID control method according to claim 1, characterized in that: in the step a, the PID neural network module includes two input layer neurons, three hidden layer neurons, and one output layer neuron;
the two inputs of the PID neural network input layer are r (k) and y (k), wherein r (k) is a wind power generator power reference value PrY (k) is the actual value P of the power of the wind driven generator;
<mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>net</mi> <mn>1</mn> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>r</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>net</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>y</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow>
the state of the input layer neurons is:
u1 j(k)＝net1 j(3)
the input layer outputs areWherein j is 1, 2;
the PID neural network hidden layer comprises three neurons of a proportional element, an integral element and a differential element, and the input weighted sum of each neuron of the hidden layer iswij 2The input weight of the ith neuron of the hidden layer;
the state of the proportional element is as follows:
u2 1(k)＝net2 1(4)
the states of the integrator are:
u2 2(k)＝u2 2(k-1)+net2 2(5)
the states of the differential elements are:
u2 3(k)＝net2 3(k)-net2 3(k-1) (6)
the output of each neuron of the hidden layer is:
<mrow> <msubsup> <mi>y</mi> <mi>i</mi> <mn>2</mn> </msubsup> <mo>=</mo> <mo>=</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mn>1</mn> </mtd> <mtd> <mrow> <msub> <msup> <mi>u</mi> <mn>1</mn> </msup> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>&gt;</mo> <mn>1</mn> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <msup> <mi>u</mi> <mn>1</mn> </msup> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mn>1</mn> <mo>&amp;le;</mo> <msub> <msup> <mi>u</mi> <mn>1</mn> </msup> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>&amp;le;</mo> <mn>1</mn> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </mtd> <mtd> <mrow> <msub> <msup> <mi>u</mi> <mn>1</mn> </msup> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>&lt;</mo> <mo>-</mo> <mn>1</mn> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>
wherein i is 1,2, 3;
the PID neural network output layer comprises a neuron, and the input weighted sum of the neuron of the output layer is expressed as The input weight of the first neuron of the output layer;
the state function of the output layer neurons is:
u3 l(k)＝net3 l(8)
the output of the output layer neurons is yl 3＝f3(net3 l) Where l is 1, the output of the PID neural network is equal to the output of the output layer neurons
5. The wind power pitch multivariable fuzzy neural network PID control method according to claim 4, characterized in that: the input layer output function f1Output layer output function f as tan sig function3Is a purelin function.
6. The wind of claim 4The PID control method of the electric variable pitch multivariable fuzzy neural network is characterized by comprising the following steps: selecting an initial weight: initial weight w from input layer to hidden layeri1 2＝+1，wi2 2-1; the initial weight from the hidden layer to the output layer is the proportion k output by the fuzzy controllerfPIntegral kfIDifferential kfDAnd (4) parameters.
7. The wind power pitch-variable multivariable fuzzy neural network PID control method according to claim 1, characterized in that: in the step c, the initial weight of the PID neural network module is adjusted through an error back propagation algorithm, and the training target of the PID neural network module is as follows:
<mrow> <msub> <mi>E</mi> <mi>t</mi> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>)</mo> </mrow> <mo>=</mo> <mn>1</mn> <mo>/</mo> <mi>m</mi> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>p</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <msup> <mrow> <mo>&amp;lsqb;</mo> <msub> <mi>y</mi> <mrow> <mi>t</mi> <mi>p</mi> </mrow> </msub> <mo>-</mo> <msub> <mi>y</mi> <mrow> <mi>o</mi> <mi>p</mi> </mrow> </msub> <mo>&amp;rsqb;</mo> </mrow> <mn>2</mn> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>9</mn> <mo>)</mo> </mrow> </mrow>
ytpdesired output for neural network, i.e. wind turbine power reference value Pr；yopAnd (4) the actual output of the neural network, namely the actual power value P of the wind driven generator, wherein m is the number of samples.
8. The wind power pitch-variable multivariable fuzzy neural network PID control method according to claim 6, characterized in that: adjusting the weight of the error back propagation process by using a negative gradient algorithm with momentum factors, wherein the learning step length is eta, and after n-1 learning steps:
<mrow> <msup> <msub> <mi>w</mi> <mrow> <mi>l</mi> <mi>i</mi> </mrow> </msub> <mn>3</mn> </msup> <mrow> <mo>(</mo> <mi>n</mi> <mo>)</mo> </mrow> <mo>=</mo> <msup> <msub> <mi>w</mi> <mrow> <mi>l</mi> <mi>i</mi> </mrow> </msub> <mn>3</mn> </msup> <mrow> <mo>(</mo> <mi>n</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mi>&amp;eta;</mi> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>E</mi> </mrow> <mrow> <mo>&amp;part;</mo> <msup> <msub> <mi>w</mi> <mrow> <mi>l</mi> <mi>i</mi> </mrow> </msub> <mn>3</mn> </msup> </mrow> </mfrac> <mo>+</mo> <msup> <msub> <mi>&amp;alpha;&amp;Delta;w</mi> <mrow> <mi>l</mi> <mi>i</mi> </mrow> </msub> <mn>3</mn> </msup> <mrow> <mo>(</mo> <mi>n</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>10</mn> <mo>)</mo> </mrow> </mrow>
the formula (10) is the weight correction from hidden layer to output layer, alpha is the momentum factor, wherein,
<mrow> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>E</mi> </mrow> <mrow> <mo>&amp;part;</mo> <msup> <msub> <mi>w</mi> <mrow> <mi>l</mi> <mi>i</mi> </mrow> </msub> <mn>3</mn> </msup> </mrow> </mfrac> <mo>=</mo> <mo>-</mo> <mfrac> <mn>1</mn> <mi>m</mi> </mfrac> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <msup> <mi>&amp;delta;</mi> <mo>&amp;prime;</mo> </msup> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <msup> <msub> <mi>y</mi> <mi>i</mi> </msub> <mn>2</mn> </msup> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>11</mn> <mo>)</mo> </mrow> </mrow>
wherein,the output values of the neurons of the hidden layer, m is the sample point, η is the learning step size, where,
<mrow> <msup> <mi>&amp;delta;</mi> <mo>&amp;prime;</mo> </msup> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <mn>2</mn> <mo>&amp;lsqb;</mo> <msub> <mi>y</mi> <mrow> <mi>t</mi> <mi>p</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>y</mi> <mrow> <mi>o</mi> <mi>p</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> <mi>sgn</mi> <mfrac> <mrow> <msub> <mi>y</mi> <mrow> <mi>o</mi> <mi>p</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>y</mi> <mrow> <mi>o</mi> <mi>p</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mi>v</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <mi>v</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>12</mn> <mo>)</mo> </mrow> </mrow>
equation (13) is the input layer to hidden layer weight correction:
<mrow> <msup> <msub> <mi>w</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mn>2</mn> </msup> <mrow> <mo>(</mo> <mi>n</mi> <mo>)</mo> </mrow> <mo>=</mo> <msup> <msub> <mi>w</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mn>2</mn> </msup> <mrow> <mo>(</mo> <mi>n</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mi>&amp;eta;</mi> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>E</mi> </mrow> <mrow> <mo>&amp;part;</mo> <msup> <msub> <mi>w</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mn>2</mn> </msup> </mrow> </mfrac> <mo>+</mo> <msup> <msub> <mi>&amp;alpha;&amp;Delta;w</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mn>2</mn> </msup> <mrow> <mo>(</mo> <mi>n</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>13</mn> <mo>)</mo> </mrow> </mrow>
wherein,
<mrow> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>E</mi> </mrow> <mrow> <mo>&amp;part;</mo> <msup> <msub> <mi>w</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mn>2</mn> </msup> </mrow> </mfrac> <mo>=</mo> <mo>-</mo> <mfrac> <mn>1</mn> <mi>m</mi> </mfrac> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <mi>&amp;delta;</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <msup> <msub> <mi>y</mi> <mi>j</mi> </msub> <mn>1</mn> </msup> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>14</mn> <mo>)</mo> </mrow> </mrow>
yj 1(k) for the output values of the neurons of the input layer, α is a momentum factor, m is a sample point, η is a learning step size, where,
<mrow> <mi>&amp;delta;</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <msup> <mi>&amp;delta;</mi> <mo>&amp;prime;</mo> </msup> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <msup> <msub> <mi>w</mi> <mrow> <mi>j</mi> <mi>l</mi> </mrow> </msub> <mn>2</mn> </msup> <mi>sgn</mi> <mfrac> <mrow> <msub> <msup> <mi>u</mi> <mn>1</mn> </msup> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <msub> <msup> <mi>u</mi> <mn>1</mn> </msup> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> <mrow> <msubsup> <mi>net</mi> <mi>i</mi> <mn>2</mn> </msubsup> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <msubsup> <mi>net</mi> <mi>i</mi> <mn>2</mn> </msubsup> <mrow> <mo>(</mo> <mi>k</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>15</mn> <mo>)</mo> </mrow> </mrow>
sgn is a sign function, wjlThe weights for the hidden layer to the output layer,weighted sum of inputs for hidden layer neurons, u1 i(k) The output of the hidden layer neurons.
CN201410706827.4A 2014-11-27 2014-11-27 A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method Expired - Fee Related CN104612898B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410706827.4A CN104612898B (en) 2014-11-27 2014-11-27 A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410706827.4A CN104612898B (en) 2014-11-27 2014-11-27 A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method

Publications (2)

Publication Number Publication Date
CN104612898A CN104612898A (en) 2015-05-13
CN104612898B true CN104612898B (en) 2017-09-08

Family

ID=53147504

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410706827.4A Expired - Fee Related CN104612898B (en) 2014-11-27 2014-11-27 A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method

Country Status (1)

Country Link
CN (1) CN104612898B (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106321352B (en) * 2015-06-30 2018-11-27 中国船舶重工集团海装风电股份有限公司 A kind of equivalent variable pitch plusderivative control method and device
CN111353589B (en) * 2016-01-20 2024-03-01 中科寒武纪科技股份有限公司 Apparatus and method for performing artificial neural network forward operations
CN105759607B (en) * 2016-02-26 2018-08-14 北京工业大学 The design method of PAC controllers based on intelligent control algorithm
CN109162867B (en) * 2017-06-29 2019-11-26 北京金风科创风电设备有限公司 Wind power generating set PID control method and device
CN107707141A (en) * 2017-10-28 2018-02-16 安徽工程大学 A kind of circuit control device and method using artificial intelligence process device
CN108223274B (en) * 2017-12-04 2019-11-12 重庆邮电大学 Large Scale Variable Pitch Wind Turbine System discrimination method based on optimization RBF neural
CN108809167A (en) * 2018-06-26 2018-11-13 长春工业大学 A kind of BP neural network PID speed regulating control algorithms based on fuzzy control
CN109063907B (en) * 2018-07-27 2020-06-05 中南大学 Intelligent traversal large-step-length prediction method for maximum wind speed along high-speed railway
CN109034478B (en) * 2018-07-27 2021-09-24 中南大学 High-precision prediction method for high-wind iterative competition along high-speed railway
CN110531614B (en) * 2019-09-06 2022-05-06 长春工业大学 Novel brushless DC motor fuzzy neural network PI controller
CN110529335B (en) * 2019-09-29 2020-10-13 三一重能有限公司 Control method for inhibiting overspeed of fan and wind generating set
CN112782499B (en) * 2019-11-11 2022-04-15 株洲中车时代电气股份有限公司 Multi-information fusion-based converter state evaluation method and device
CN113780543B (en) * 2021-09-08 2024-02-02 福州大学 FPGA implementation method for PID closed-loop control of neural network
CN114407679A (en) * 2022-03-04 2022-04-29 浙江理工大学 Expansion bracket butt joint control method and mobile charging method
CN115009278B (en) * 2022-08-08 2022-11-29 潍柴动力股份有限公司 Cruise control method, device, equipment and storage medium

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1959130A2 (en) * 2007-02-02 2008-08-20 General Electric Company Method for optimizing the operation of a wind turbine
CN101706335A (en) * 2009-11-11 2010-05-12 华南理工大学 Wind power forecasting method based on genetic algorithm optimization BP neural network
CN101976044A (en) * 2010-10-22 2011-02-16 天津理工大学 Wind power system modeling and DSP (Digital Signal Processor) realizing method based on neural network
CN102518560A (en) * 2011-12-30 2012-06-27 北京国电思达科技有限公司 Method for regulating active power of wind power field
CN103061972A (en) * 2011-10-20 2013-04-24 华锐风电科技(集团)股份有限公司 Power control method and system of wind generating set
KR101287319B1 (en) * 2011-12-15 2013-07-22 주식회사 웰케이원 Grid connected wind power system and sensorless maximum power point tracking control method thereof using neural network
CN103410662A (en) * 2013-08-06 2013-11-27 江苏科技大学 Neural network compensation control method for capturing maximum wind energy in wind power system
CN103701378A (en) * 2013-12-27 2014-04-02 内蒙古久和能源科技有限公司 Output power control method for wind driven generator
CN103903087A (en) * 2014-03-06 2014-07-02 东南大学 Steam-driven induced draft fan all-working-condition online monitoring method based on BP neural network
CN104091209A (en) * 2014-06-26 2014-10-08 沈阳工业大学 Wind driven generator set power characteristic assessment method based on BP neural network

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1959130A2 (en) * 2007-02-02 2008-08-20 General Electric Company Method for optimizing the operation of a wind turbine
CN101706335A (en) * 2009-11-11 2010-05-12 华南理工大学 Wind power forecasting method based on genetic algorithm optimization BP neural network
CN101976044A (en) * 2010-10-22 2011-02-16 天津理工大学 Wind power system modeling and DSP (Digital Signal Processor) realizing method based on neural network
CN103061972A (en) * 2011-10-20 2013-04-24 华锐风电科技(集团)股份有限公司 Power control method and system of wind generating set
KR101287319B1 (en) * 2011-12-15 2013-07-22 주식회사 웰케이원 Grid connected wind power system and sensorless maximum power point tracking control method thereof using neural network
CN102518560A (en) * 2011-12-30 2012-06-27 北京国电思达科技有限公司 Method for regulating active power of wind power field
CN103410662A (en) * 2013-08-06 2013-11-27 江苏科技大学 Neural network compensation control method for capturing maximum wind energy in wind power system
CN103701378A (en) * 2013-12-27 2014-04-02 内蒙古久和能源科技有限公司 Output power control method for wind driven generator
CN103903087A (en) * 2014-03-06 2014-07-02 东南大学 Steam-driven induced draft fan all-working-condition online monitoring method based on BP neural network
CN104091209A (en) * 2014-06-26 2014-10-08 沈阳工业大学 Wind driven generator set power characteristic assessment method based on BP neural network

Also Published As

Publication number Publication date
CN104612898A (en) 2015-05-13

Similar Documents

Publication Publication Date Title
CN104612898B (en) A kind of wind electricity change paddle is away from Multivariable Fuzzy NN-PID Control Method
Abdelbaky et al. Design and implementation of partial offline fuzzy model-predictive pitch controller for large-scale wind-turbines
JP5562366B2 (en) Hybrid intelligent power control system and method
CN103595046B (en) Wind bavin hybrid power system LOAD FREQUENCY control method based on diesel engine side
Tiwari et al. Comparative analysis of pitch angle controller strategies for PMSG based wind energy conversion system
CN104329224A (en) Variable-rate pitching system and variable-rate pitching method for direct-drive permanent magnet wind generating set
CN103410662A (en) Neural network compensation control method for capturing maximum wind energy in wind power system
CN102900603B (en) Variable pitch controller design method based on finite time non-crisp/guaranteed-cost stable wind turbine generator set
CN103746628B (en) Method for controlling rotor-side converter of doubly fed induction generator (DFIG)
Chatri et al. Improved high-order integral fast terminal sliding mode-based disturbance-observer for the tracking problem of PMSG in WECS
Arezki et al. Study and regulation of dc bus voltages of wind-photovoltaic system
Baburajan Improving the efficiency of a wind turbine system using a fuzzy-pid controller
Qi et al. PID control in adjustable-pitch wind turbine system based on fuzzy control
CN103595045B (en) Blower fan participates in the wind bavin hybrid power system LOAD FREQUENCY control method for coordinating of frequency modulation
Yao et al. RBF neural network based self-tuning PID pitch control strategy for wind power generation system
CN204239155U (en) A kind of direct-driving permanent-magnetic wind-driven group variable Rate pitch-controlled system
CN103248292B (en) A kind of control method for mechanical energy storage system
CN102900605A (en) Wind turbine generator set variable pitch controller design method based on finite time stabilization
CN102900613A (en) Wind turbine generator set variable pitch controller design method based on finite time robustness or guaranteed cost stabilization
CN102890449B (en) Based on the Wind turbines Variable-pitch Controller method for designing of finite time robust stability
Fard et al. Fuzzy logic based pitch angle controller for variable speed wind turbine
Chang et al. Combined predictive feedforward and feedback control for blade pitch of wind turbine
Roshandel et al. A Control Strategy for DFIG wind turbine using SLS-TLBO and fuzzy logic
Qin et al. H-infinity Variable-Pitch Control for Wind Turbines Based on Takagi-Sugeno Fuzzy Theory
Shengsheng et al. H-infinity Variable-Pitch Control for Wind Turbines Based on Takagi-Sugeno Fuzzy Theory.

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20210119

Address after: Room 17-1, building 033, building 2, No.15, Lane 587, Juxian Road, high tech Zone, Ningbo, Zhejiang 315000

Patentee after: CHINA E-TECH (NINGBO) MARITIME ELECTRONICS RESEARCH INSTITUTE Co.,Ltd.

Address before: Meng Xi Road 212003 Zhenjiang city of Jiangsu province Jingkou District No. 2

Patentee before: JIANGSU University OF SCIENCE AND TECHNOLOGY

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: 20170908