CN105888971A - Active load reducing control system and method for large wind turbine blade - Google Patents

Abstract

The invention relates to a large-scale wind turbine blade active load reduction control system and method. The fiber optic strain sensor measures the stress value at the root of the blade and transmits it to the control unit. The control unit includes a PID controller that uses teaching and learning algorithms for parameter optimization and a fuzzy controller that processes the strain signal at the root of the blade. Through the fuzzy controller The switching control of the flap swing angle with the PID controller combines the respective advantages of fuzzy control and PID control; at the same time, the front wind field signal sensing device measures the front wind speed of the wind turbine and transmits it to the control unit, and the feedforward in the control unit The controller calculates the control amount required to reduce the uneven load caused by random wind or turbulent wind by monitoring the wind speed in front of the aircraft in real time; the control unit couples the two parts of the control amount to complete the control of the flap swing angle. The invention effectively reduces the strain force at the root of the blade, prolongs the service life of the blade, and reduces the use cost of the wind power unit.

Description

A large-scale wind turbine blade active load reduction control system and method

technical field

The invention belongs to the technical field of wind power generation, and in particular relates to a large-scale wind turbine blade active load reduction control system and method.

Background technique

With the rapid development of wind power generation, wind turbines are also gradually developing towards offshore and large-scale trends, which puts forward higher requirements for the core components of wind turbines - fan blades. The large-scale fan means increasing the load of the fan and the quality of the system, which will increase the fatigue load and ultimate load sharply, seriously reduce the service life of the unit, and also increase the hardware and engineering costs and post-maintenance costs.

In order to reduce the fatigue load and ultimate load of the wind turbine and improve the life of the wind turbine, the stability of the power generation and output, a large wind turbine blade with active trailing edge flaps has been produced in recent years. Its hardware structure and control method are different from traditional blades. not exactly. Due to the complexity of the blade structure, it is very difficult to establish an accurate mathematical model, and the impact of random wind and turbulent wind on the load of the fan blades cannot be ignored, so the conventional single-stage PID control is difficult to meet the load reduction requirements of the blades.

Contents of the invention

Aiming at the deficiencies in the prior art, the present invention provides an active load reduction control system and method for large wind turbine blades.

A large-scale wind turbine blade active load reduction control system, the system includes a signal acquisition module, a control module and an electrical servo module;

The signal acquisition module includes an optical fiber strain sensor, a wind field signal sensing device in front of the machine, and an optical fiber strain signal processing device; the control module includes a 5-way low-pass filter, a 5-way analog-to-digital converter, a 4-way control unit, and a PLC , 4-way digital-to-analog converters and 4-way signal isolators; the electrical servo module includes 4-way flap actuator drive circuits and 4-way flap actuators;

The optical fiber strain sensor is installed at the root of the wind turbine blade, and is connected with the optical fiber strain signal processing equipment; 1 signal output of the optical fiber strain signal processing equipment and 4 signal outputs of the front wind field signal sensing equipment corresponding to 4 flaps Corresponding to 1 channel of low-pass filter and 1 channel of analog-to-digital converter are connected sequentially; 1 channel of analog-to-digital converter corresponding to the optical fiber strain signal processing equipment is respectively connected to 4 channels of control units, and connected to the wind field signal sensing equipment in front of the machine The corresponding 4-way analog-to-digital converters are connected to 4-way control units one by one, and each control unit corresponds to 1-way digital-to-analog converter, 1-way signal isolator, 1-way flap actuator drive circuit and 1-way control unit through PLC. Road flap actuators are connected sequentially;

The optical fiber strain sensor is used to collect the strain force signal of the blade root, the optical fiber strain signal processing device is used to convert the signal collected by the optical fiber strain sensor into a voltage signal; the wind field signal sensing device in front of the machine is used to measure the wind force The wind speed in front of the machine; the optical fiber strain signal processing equipment and the wind field signal sensing equipment in front of the machine respectively transmit the signal to the corresponding low-pass filter, and the low-pass filter is used to filter out high-frequency interference signals; the analog-to-digital converter uses The control unit is used to convert the analog signal into a digital signal; the control unit includes a feed-forward controller for processing the wind field signal in front of the machine, a PID controller for parameter optimization using a teaching and learning algorithm, and a signal for strain force at the root of the blade. The processed fuzzy controller is used to reduce the control operation of the blade root stress; the PLC is used to switch the output signal of the PID controller and the fuzzy controller, and couples the control signal from the feedforward controller; the digital-to-analog conversion The device is used to convert digital signals into analog signals; the signal isolator is used to isolate the output signal of the control system from the electrical servo module; the flap actuator drive circuit generates electrical signals for driving the flap actuator; the The flap actuator adjusts the flap to produce different swing angles according to the output signal of the flap actuator drive circuit.

The feed-forward controller calculates the control amount required to reduce the uneven load caused by random wind or turbulent wind through real-time monitoring of wind speed changes in front of the wind turbine when the wind power generating set is running.

The above-mentioned control method for the active load reduction control system of large-scale wind turbine blades specifically includes the following steps:

Step 1: Initialize fuzzy controller FC _i , feedforward controller FB _i and PID controller PID _i , i=1,2,3,4;

Step 2: Read the current wind speed and the stress value of the blade root, and calculate the difference between the obtained blade root stress value y(k) and the blade root rated stress value r(k), and obtain the stress deviation e(k) and deviation change Rate ec(k), where the blade rated stress value r(k) is measured by the fan before leaving the factory;

Step 21: use the stress deviation e(k) and deviation change rate ec(k) obtained in step 2 as input variables of the fuzzy controller FC _i ;

Step 22: Select the membership function for fuzzification, and obtain the control quantity of the flap according to the fuzzy rules, and obtain the output variable of the fuzzy controller FC _i after defuzzification, and this output is the desired flap control angle θ1 _i ;

Step 23: Use the stress deviation e(k) of step 21 as the input signal of the PID controller PID _i , and use the teaching and learning optimization algorithm to carry out online self-tuning of the PID _i parameters KP _i , KI _i , KD _i , the PID The output variable of the controller PID _i is the flap control desired angle θ2 _i ;

Step 24: Collect the wind speed v(t) in front of the wind turbine, use it as the independent variable, and use the flap swing angle _θi as the dependent variable, fit the flap angle-wind speed, and establish the flap angle-wind speed front Feed-forward controller FB _i model: θ _i (v)＝a ₀ +a ₁ v+a ₂ v ² +L+a _n v ⁿ , the coefficients are determined by the least square method; the wind speed is used as the feed-forward controller FB _i The input signal of the feedforward controller FB _i is the output variable of the flap control desired angle θ3 _i ;

Step 3: Send the desired flap angle control signals obtained in Step 22 and Step 23 to the PLC for processing, and set the switching algorithm in the PLC: when the blade root stress error is greater than the set value, use the signal from the fuzzy controller FC _i θ1 _i , the control signal θ2 _i from the PID controller PID _i will not work; when the blade root stress error is less than the set value, use the control signal θ2 _i from the PID controller PID _i , the signal from the fuzzy controller FC _i θ1 _i will not work; finally, the PLC couples the control signals θ3 _i from the feedforward controller FB _i , and transmits these signals to the corresponding flap actuator drive circuit;

Step 4: The four flap actuators receive signals from the four flap actuator drive circuits respectively, and perform the flap swing action to reduce the stress on the blade root;

The above steps 2-4 are run repeatedly until the control task is completed.

The teaching and learning optimization algorithm for PID parameters includes the following steps:

Step 1): Set the initial parameters, the search area is defined as X=(x ¹ ,x ² ,...,x ^d )∈[L,U], L=(L ¹ ,L ² ,...,L ^d ) is the lower bound of the space, U=(U ¹ , U ² ,...,U ^d ) is the upper bound of the space, d is the dimension of the optimization problem, and the i-th student in the d-dimensional space is defined as The size of the trainees is N, and the maximum number of iterations is maxgen;

Step 2): Teacher's Teaching Phase:

Step 21): Calculate the fitness value of each student, select the best individual as the teacher X _teacher , and calculate the individual average Then learn according to the difference between the students and the average level of the individual, as follows:

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; h</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; TF</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

TF _i =2-gen/max gen (2)

In the formula: and represent the values of the i-th student before and after learning respectively; w1=1-gen/max gen is the adaptive weight coefficient; r _i is a random number between 0-1; TF _i is a certain value between 1-2 The number, its value changes with the number of iterations; r _i and TF _i are used to adjust the learning rate; gen and max gen are the current number of iterations and the maximum number of iterations, respectively;

Step 22): Student update:

if The fitness value ratio The fitness value is good, then use replace Otherwise, continue using

Step 3): Mutual learning phase among students:

Step 31): Each student X _i randomly selects a learning object X _j (j≠i) in the class, and X _i adjusts learning by analyzing the difference between itself and the student X _j , as follows:

If X _i is better than X _j ,

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

If X _j is better than X _i ,

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

In the formula: w2=1-gen/max gen is the adaptive weight coefficient; r _i is a random number between 0-1;

Step 32): Student update:

if The fitness value ratio The fitness value is good, then use replace Otherwise, continue using

Step 4): Calculate the fitness value of each student according to the fitness function, the formula is as follows:

$\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; \&infin;</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; d</span>}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

In the formula, e(t) is the system error, and the global optimal value of the students is updated according to the fitness function. When the calculated optimal value reaches the set value or the algorithm reaches the maximum number of iterations, the teaching and learning optimization algorithm is exited. Otherwise return to step 2).

The beneficial effects of the present invention are:

1. Compared with the traditional single-stage PID control system, the introduction of a feed-forward controller in this system can effectively reduce the impact of the random fluctuation of the wind on the stress of the blade root. 2. Use fuzzy control and PID control to switch control at the set value, which not only takes advantage of the advantages of the fuzzy controller that does not need to establish an accurate mathematical model and can adapt to the nonlinearity and time-varying nature of the controlled object, but also uses the PID controller algorithm Simple and stable features. 3. Use the teaching and learning algorithm to optimize the speed of optimization and solve the characteristics of high precision, and optimize the three parameters of the PID controller online, thereby effectively reducing the strain force on the root of the blade, prolonging the service life of the blade, and reducing the wind turbine. The cost. 4. The present invention is simple in design, convenient in application, more accurate and reliable in control, and is very suitable for modeling and control of large-scale wind turbine blades.

Description of drawings

Figure 1 is a structural block diagram of an active load reduction control system for large wind turbine blades.

Fig. 2 is a control principle block diagram of a large-scale wind turbine blade active load reduction control system.

Fig. 3 is a flow chart of teaching and learning optimization algorithm for online self-tuning of PID controller parameters.

Fig. 4 is a comparison result between an active load reduction control method for large wind turbine blades of the present invention and the existing method.

detailed description

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. It should be emphasized that the following description is only exemplary and not intended to limit the scope of the invention and its application.

As shown in Figure 1, a large-scale wind turbine blade active load reduction control system, the system includes a signal acquisition module, a control module and an electrical servo module;

The signal acquisition module includes an optical fiber strain sensor, a wind field signal sensing device in front of the machine, and an optical fiber strain signal processing device; the control module includes a 5-way low-pass filter, a 5-way analog-to-digital converter, a 4-way control unit, and a PLC , 4-way digital-to-analog converters and 4-way signal isolators; the electrical servo module includes 4-way flap actuator drive circuits and 4-way flap actuators;

The optical fiber strain sensor is installed at the root of the wind turbine blade, and is connected with the optical fiber strain signal processing equipment; 1 signal output of the optical fiber strain signal processing equipment and 4 signal outputs of the front wind field signal sensing equipment corresponding to 4 flaps Corresponding to 1 channel of low-pass filter and 1 channel of analog-to-digital converter are connected sequentially; 1 channel of analog-to-digital converter corresponding to the optical fiber strain signal processing equipment is respectively connected to 4 channels of control units, and connected to the wind field signal sensing equipment in front of the machine The corresponding 4-way analog-to-digital converters are connected to 4-way control units one by one, and each control unit corresponds to 1-way digital-to-analog converter, 1-way signal isolator, 1-way flap actuator drive circuit and 1-way control unit through PLC. Road flap actuators are connected sequentially;

The optical fiber strain sensor is used to collect the strain force signal of the blade root, the optical fiber strain signal processing device is used to convert the signal collected by the optical fiber strain sensor into a voltage signal; the wind field signal sensing device in front of the machine is used to measure the wind force The wind speed in front of the machine; the optical fiber strain signal processing equipment and the wind field signal sensing equipment in front of the machine respectively transmit the signal to the corresponding low-pass filter, and the low-pass filter is used to filter out high-frequency interference signals; the analog-to-digital converter uses The control unit is used to convert the analog signal into a digital signal; the control unit is used to reduce the stress of the blade root. It includes a feed-forward controller for processing the wind field signal in front of the machine. When the unit is running, through real-time monitoring of wind speed changes in front of the machine, the control amount required to reduce the uneven load caused by random wind or turbulent wind is calculated; the PID controller using the teaching and learning algorithm for parameter optimization; and the root of the blade A fuzzy controller for processing the strain signal; the PLC is used to switch the output signal of the PID controller and the fuzzy controller, and couples the control signal from the feedforward controller; the digital-to-analog converter is used to convert the digital signal into an analog signal; the signal isolator is used to isolate the output signal of the control system from the electrical servo module; the flap actuator drive circuit generates an electrical signal to drive the flap actuator; the flap actuator is based on The output signal of the flap actuator drive circuit adjusts the flap to produce different swing angles.

As shown in Figure 2, a large-scale wind turbine blade active load reduction control method specifically includes the following steps:

Step 1: Initialize fuzzy controller FC _i , feedforward controller FB _i and PID controller PID _i , i=1,2,3,4.

Step 2: Read the current wind speed and the stress value of the blade root, and calculate the difference between the obtained blade root stress value y(k) and the blade root rated stress value r(k), and obtain the stress deviation e(k) and deviation change Rate ec(k), where the blade rated stress value r(k) is measured by the fan before leaving the factory;

Step 21: use the stress deviation e(k) and deviation change rate ec(k) obtained in step 2 as input variables of the fuzzy controller FC _i ;

Step 22: Select the membership function for fuzzification, and obtain the control quantity of the flap according to the fuzzy rules, and obtain the output variable of the fuzzy controller FC _i after defuzzification, and this output is the desired flap control angle θ1 _i ;

Step 23: Use the stress deviation e(k) of step 21 as the input signal of the PID controller PID _i , and use the teaching and learning optimization algorithm to carry out online self-tuning of the PID _i parameters KP _i , KI _i , KD _i , the PID The output variable of the controller PID _i is the flap control desired angle θ2 _i ;

As shown in Figure 3, the teaching and learning optimization algorithm for PID parameters includes the following steps:

Step 1): Set the initial parameters, the search area is defined as X=(x ¹ ,x ² ,...,x ^d )∈[L,U], L=(L ¹ ,L ² ,...,L ^d ) is the lower bound of the space, U=(U ¹ , U ² ,...,U ^d ) is the upper bound of the space, d is the dimension of the optimization problem, and the i-th student in the d-dimensional space is defined as The size of the trainees is N, and the maximum number of iterations is max gen;

Step 2): Teacher's Teaching Phase:

Step 21): Calculate the fitness value of each student, select the best individual as the teacher X _teacher , and calculate the individual average Then learn according to the difference between the students and the average level of the individual, as follows:

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; h</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; TF</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

TF _i =2-gen/max gen (2)

In the formula: and represent the values of the i-th student before and after learning respectively; w1=1-gen/max gen is the adaptive weight coefficient; r _i is a random number between 0-1; TF _i is a certain value between 1-2 The number, its value changes with the number of iterations; r _i and TF _i are used to adjust the learning rate; gen and max gen are the current number of iterations and the maximum number of iterations, respectively;

Step 22): Student update:

if The fitness value ratio The fitness value is good, then use replace Otherwise, continue using

Step 3): Mutual learning phase among students:

Step 31): Each student X _i randomly selects a learning object X _j (j≠i) in the class, and X _i adjusts learning by analyzing the difference between itself and the student X _j , as follows:

If X _i is better than X _j ,

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

If X _j is better than X _i ,

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

In the formula: w2=1-gen/max gen is the adaptive weight coefficient; r _i is a random number between 0-1;

Step 32): Student update:

if The fitness value ratio The fitness value is good, then use replace Otherwise, continue using

Step 4): Calculate the fitness value of each student according to the fitness function, the formula is as follows:

$\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; \&infin;</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; d</span>}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

In the formula, e(t) is the system error, and the global optimal value of the students is updated according to the fitness function. When the calculated optimal value reaches the set value or the algorithm reaches the maximum number of iterations, the teaching and learning optimization algorithm is exited. Otherwise return to step 2);

Step 24: Collect the wind speed v(t) in front of the wind turbine, use it as the independent variable, and use the flap swing angle _θi as the dependent variable, fit the flap angle-wind speed, and establish the flap angle-wind speed front Feed-forward controller FB _i model: θ _i (v)＝a ₀ +a ₁ v+a ₂ v ² +L+a _n v ⁿ , the coefficients are determined by the least square method; the wind speed is used as the feed-forward controller FB _i The input signal of the feedforward controller FB _i is the output variable of the flap control desired angle θ3 _i .

Step 3: Send the desired flap angle control signals obtained in Step 22 and Step 23 to the PLC for processing, and set the switching algorithm in the PLC: when the blade root stress error is greater than the set value, use the signal from the fuzzy controller FC _i θ1 _i , the control signal θ2 _i from the PID controller PID _i will not work; when the blade root stress error is less than the set value, use the control signal θ2 _i from the PID controller PID _i , the signal from the fuzzy controller FC _i θ1 _i will be inactive; finally, the PLC couples the control signals θ3 _i from the feed-forward controller FB _i and transmits these signals to the corresponding flap actuator drive circuit.

Step 4: The four flap actuators respectively receive the signals from the four flap actuator drive circuits, and perform the flap swing action to reduce the stress on the blade root.

The above steps 2-4 are run repeatedly until the control task is completed.

The 5WM reference wind turbine is under the turbulent wind whose wind condition is 11.4m/s, and uses the root bending moment of the blade as the load reduction target of the wind turbine, and adopts a kind of active load reduction control method for large wind turbine blades of the present invention and the existing method The comparison results are shown in Fig. 4, it can be seen that the method of the present invention effectively reduces the strain force at the root of the blade.

The above are only preferred embodiments of the present invention, and are not intended to limit the present invention in any form. Any person skilled in the art can easily think of simple modifications and equivalent changes within the technical scope disclosed in the present invention. , should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims (4)

1. A large-scale wind turbine blade active load reduction control system, characterized in that the system includes a signal acquisition module, a control module and an electrical servo module; The signal acquisition module includes an optical fiber strain sensor, a wind field signal sensing device in front of the machine, and an optical fiber strain signal processing device; the control module includes a 5-way low-pass filter, a 5-way analog-to-digital converter, a 4-way control unit, and a PLC , 4-way digital-to-analog converters and 4-way signal isolators; the electrical servo module includes 4-way flap actuator drive circuits and 4-way flap actuators; The optical fiber strain sensor is installed at the root of the wind turbine blade, and is connected with the optical fiber strain signal processing equipment; 1 signal output of the optical fiber strain signal processing equipment and 4 signal outputs of the front wind field signal sensing equipment corresponding to 4 flaps Corresponding to 1 channel of low-pass filter and 1 channel of analog-to-digital converter are connected sequentially; 1 channel of analog-to-digital converter corresponding to the optical fiber strain signal processing equipment is respectively connected to 4 channels of control units, and connected to the wind field signal sensing equipment in front of the machine The corresponding 4-way analog-to-digital converters are connected to 4-way control units one by one, and each control unit corresponds to 1-way digital-to-analog converter, 1-way signal isolator, 1-way flap actuator drive circuit and 1-way control unit through PLC. Road flap actuators are connected sequentially; The optical fiber strain sensor is used to collect the strain force signal of the blade root, the optical fiber strain signal processing device is used to convert the signal collected by the optical fiber strain sensor into a voltage signal; the wind field signal sensing device in front of the machine is used to measure the wind force The wind speed in front of the machine; the optical fiber strain signal processing equipment and the wind field signal sensing equipment in front of the machine respectively transmit the signal to the corresponding low-pass filter, and the low-pass filter is used to filter out high-frequency interference signals; the analog-to-digital converter uses The control unit is used to convert the analog signal into a digital signal; the control unit includes a feed-forward controller for processing the wind field signal in front of the machine, a PID controller for parameter optimization using a teaching and learning algorithm, and a signal for strain force at the root of the blade. The processed fuzzy controller is used to reduce the control operation of the blade root stress; the PLC is used to switch the output signal of the PID controller and the fuzzy controller, and couples the control signal from the feedforward controller; the digital-to-analog conversion The device is used to convert digital signals into analog signals; the signal isolator is used to isolate the output signal of the control system from the electrical servo module; the flap actuator drive circuit generates electrical signals for driving the flap actuator; the The flap actuator adjusts the flap to produce different swing angles according to the output signal of the flap actuator drive circuit. 2. A kind of active load reduction control system for large wind turbine blades according to claim 1, characterized in that, the feedforward controller calculates the reason for the reduction by real-time monitoring of wind speed changes in front of the wind turbine when the wind turbine is running. The amount of control required for uneven loads caused by random or turbulent winds. 3. The control method of a large-scale wind turbine blade active load reduction control system according to any one of claims 1-2, characterized in that it specifically comprises the following steps: Step 1: Initialize fuzzy controller FC _i , feedforward controller FB _i and PID controller PID _i , i=1,2,3,4; Step 2: Read the current wind speed and the stress value of the blade root, and calculate the difference between the obtained blade root stress value y(k) and the blade root rated stress value r(k), and obtain the stress deviation e(k) and deviation change Rate ec(k), where the blade rated stress value r(k) is measured by the fan before leaving the factory; Step 21: use the stress deviation e(k) and deviation change rate ec(k) obtained in step 2 as input variables of the fuzzy controller FC _i ; Step 22: Select the membership function for fuzzification, and obtain the control quantity of the flap according to the fuzzy rules, and obtain the output variable of the fuzzy controller FC _i after defuzzification, and this output is the desired flap control angle θ1 _i ; Step 23: Use the stress deviation e(k) of step 21 as the input signal of the PID controller PID _i , and use the teaching and learning optimization algorithm to carry out online self-tuning of the PID _i parameters KP _i , KI _i , KD _i , the PID The output variable of the controller PID _i is the flap control desired angle θ2 _i ; Step 24: Collect the wind speed v(t) in front of the wind turbine, use it as the independent variable, and use the flap swing angle _θi as the dependent variable, fit the flap angle-wind speed, and establish the flap angle-wind speed front Feed-forward controller FB _i model: θ _i (v)＝a ₀ +a ₁ v+a ₂ v ² +L+a _n v ⁿ , the coefficients are determined by the least square method; the wind speed is used as the feed-forward controller FB _i The input signal of the feedforward controller FB _i is the output variable of the flap control desired angle θ3 _i ; Step 3: Send the desired flap angle control signals obtained in Step 22 and Step 23 to the PLC for processing, and set the switching algorithm in the PLC: when the blade root stress error is greater than the set value, use the signal from the fuzzy controller FC _i θ1 _i , the control signal θ2 _i from the PID controller PID _i will not work; when the blade root stress error is less than the set value, use the control signal θ2 _i from the PID controller PID _i , the signal from the fuzzy controller FC _i θ1 _i will not work; finally, the PLC couples the control signals θ3 _i from the feedforward controller FB _i , and transmits these signals to the corresponding flap actuator drive circuit; Step 4: The four flap actuators receive signals from the four flap actuator drive circuits respectively, and perform the flap swing action to reduce the stress on the blade root; The above steps 2-4 are run repeatedly until the control task is completed. 4. a kind of control method according to claim 3 is characterized in that, the teaching and learning optimization algorithm for PID parameter comprises the following steps: Step 1): Set the initial parameters, the search area is defined as X=(x ¹ ,x ² ,...,x ^d )∈[L,U], L=(L ¹ ,L ² ,...,L ^d ) is the lower bound of the space, U=(U ¹ , U ² ,...,U ^d ) is the upper bound of the space, d is the dimension of the optimization problem, and the i-th student in the d-dimensional space is defined as The size of the trainees is N, and the maximum number of iterations is maxgen; Step 2): Teacher's Teaching Phase: Step 21): Calculate the fitness value of each student, select the best individual as the teacher X _teacher , and calculate the individual average Then learn according to the difference between the students and the average level of the individual, as follows: ${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; h</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; TF</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ TF _i =2-gen/maxgen (2) In the formula: and respectively represent the value of the i-th student before and after learning; w1=1-gen/maxgen is the adaptive weight coefficient; r _i is a random number between 0-1; TF _i is a certain value between 1-2 number, its value changes with the number of iterations; r _i and TF _i are used to adjust the learning rate; gen and max gen are the current number of iterations and the maximum number of iterations, respectively; Step 22): Student update: if The fitness value ratio The fitness value is good, then use replace Otherwise, continue using Step 3): Mutual learning phase among students: Step 31): Each student X _i randomly selects a learning object X _j (j≠i) in the class, and X _i adjusts learning by analyzing the difference between itself and the student X _j , as follows: If X _i is better than X _j , ${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$ If X _j is better than X _i , ${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; w</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; l</span>}\mathrm{<span\; class="notranslate">\; d</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$ In the formula: w2=1-gen/maxgen is the adaptive weight coefficient; r _i is a random number between 0-1; Step 32): Student update: if The fitness value ratio The fitness value is good, then use replace Otherwise, continue using Step 4): Calculate the fitness value of each student according to the fitness function, the formula is as follows: $\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; \&infin;</span>}}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; d</span>}\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$ In the formula, e(t) is the system error, and the global optimal value of the students is updated according to the fitness function. When the calculated optimal value reaches the set value or the algorithm reaches the maximum number of iterations, the teaching and learning optimization algorithm is exited. Otherwise return to step 2).