CN103281022B - Double-efficiency fuzzy optimization control method for doubly-fed wind generator - Google Patents

Abstract

A double-fed wind generator dual-efficiency fuzzy optimization control method, the improvement of which is: after the start of grid-connected power generation and before the wind generator reaches the maximum allowable speed, the maximum wind energy tracking control is used to control the wind generator. In the tracking control process, the control mode is switched to the optimal reactive power search control in good time, thereby realizing the double optimization of the operating efficiency of the generator; the beneficial technical effect of the present invention is: in the maximum wind energy tracking process, by changing the fuzzy reasoning rules The quantitative factor of the input variable of the table improves the search accuracy of the fuzzy reasoning rule table, avoids the adjustment dead zone, and improves the utilization rate of wind energy by the wind turbine; in the process of optimal reactive power tracking, it can effectively suppress flutter and improve the search accuracy. Accuracy, reducing the self-loss of the wind turbine and increasing the active output.

Description

Fuzzy optimization control method for double efficiency of doubly-fed wind turbine

technical field

The invention relates to a wind power generator control technology, in particular to a double-fed wind power generator double-efficiency fuzzy optimization control method.

Background technique

The dual-efficiency fuzzy optimal control of doubly-fed wind turbines can simultaneously achieve the dual control objectives of maximum wind energy tracking and generator minimum loss operation, which is of great significance for increasing the power generation of the unit. It is of great significance to conduct further in-depth research on the operation level.

Generally, by dynamically setting the active power reference of the generator stator for indirect speed control, the maximum wind energy tracking without wind speed measurement can be realized. The setting of the active power of the generator stator is related to the characteristics of the fan. In practice, parameters such as fan parameters and air density will affect fan characteristics. The generator stator active power setting during unit operation will always deviate from the optimal value, which will affect the tracking effect of the maximum wind energy. Therefore, it is a better method to use fuzzy logic to search for the optimal speed of the wind turbine to realize the maximum wind energy tracking without wind speed measurement; principle, due to the The calculation expression involves many parameters such as generator resistance and inductance. The measurement is limited by accuracy, and the resistance value is significantly affected by temperature. Therefore, it is difficult for this strategy to achieve the lowest generator loss in actual operation. The use of fuzzy logic to search for the optimal reactive power of the generator does not depend on the accuracy of the generator parameters and has good adaptability. Therefore, the fuzzy logic strategy can be used to optimize the generator speed and reactive power setting respectively, so as to achieve the dual goals of maximum wind energy tracking and minimum loss operation of the generator without wind speed measurement, so as to avoid the optimal reference value setting from affecting the accuracy of the object parameters. Dependency, which is the advantage of using artificial intelligence decision-making.

On the one hand, in the process of using fuzzy logic to realize the maximum wind energy tracking and search for the optimal speed reference value, when approaching the optimal point with a certain search step size, there is still an adjustment dead zone in the small deviation range. As a result, the speed search forms a limit cycle oscillation near the optimal point, and at the same time, the output power will also oscillate near the maximum power point, which limits the high-precision search and positioning of the optimal speed reference value and maximum power by fuzzy logic. It will cause long-term vibration of the transmission shaft, which will cause great damage to the mechanical parts of the fan. This is the defect of fuzzy logic search, so it is necessary to take corresponding measures to improve the accuracy of the search; at the same time, here, the double-fed wind turbine to achieve maximum wind energy tracking and minimum loss operation is the dual control goal, and it is necessary to solve the problem of coordination between the two. The problem. Because only after the maximum wind energy tracking search is completed, the search for the next optimal reactive power reference value can be continued. Therefore, the measures taken to improve the search accuracy should also be simple and fast. After quickly completing the search for the maximum wind energy, the next step of searching for the optimal reactive power reference value can be quickly entered.

On the other hand, in the process of searching for optimal reactive power, the single fuzzy logic search currently used also has the problem of search oscillation. How to quickly search for the optimal reactive power value with high precision, reduce the search oscillation, minimize the loss of the generator itself, and further improve the operating efficiency of the generator is also a problem to be solved by the single fuzzy logic search.

Contents of the invention

Aiming at the problems in the background technology, the present invention proposes a double-fed wind generator double-efficiency fuzzy optimization control method, including a wind generator controlled by a double-fed control system; when the wind speed is greater than or equal to the cut-in wind speed, the wind generator Grid-connected power generation, with the increase of wind speed, when the wind turbine reaches the allowable maximum speed, it enters the constant speed power generation state, which is characterized in that: after starting grid-connected power generation and before the wind turbine reaches the allowable maximum speed, the following control method is adopted To control the wind turbine:

When the wind turbine is first connected to the grid for power generation, the maximum wind energy tracking control should be carried out as follows:

1) When the wind speed is greater than or equal to the cut-in wind speed, the wind generator is connected to the grid to generate electricity, and the speed and active power of the wind generator are continuously sampled, and the change value of the speed and the change value of the active power are calculated in each sampling period ; Suppose the rotational speed change value recorded in a single sampling period is Δω _ri , the active power change value recorded in a single sampling period is Δp _ei , i is the number of samples, i=1, 2, 3, 4...n; set Δω _{Both r1} and Δp _e1 are positive;

2) In the second sampling period, Lp _e1 and Δω _r1 are used as the two input variables of the first fuzzy inference rule table, and the theoretical speed change value is obtained according to the first fuzzy inference rule table, and the wind turbine generator is calculated according to the theoretical speed change value The optimized rotational speed value, the double-fed control system adjusts the rotational speed of the wind turbine according to the optimized rotational speed value; enter step 3); the second sampling period also forms the first control period;

Among them, Lp _e1 is the active power input variable corresponding to k _p and Δp _e1 , Lp _e1 = k _p Δp _e1 , k _p is the quantization factor of the input variable corresponding to Δp _ei ;

3) In the subsequent control cycle, the current Lp _ei and the theoretical speed change value obtained in the previous control cycle are used as the two input variables of the first fuzzy inference rule table, and the theoretical speed change value is obtained according to the first fuzzy inference rule table, Calculate the optimal speed value of the wind turbine according to the theoretical speed change value; the double-fed control system continuously and dynamically adjusts the speed of the wind generator according to the corresponding optimal speed value; enter step 4);

Among them, Lp _ei =k _p ·Δp _ei ; Lp _ei is the active power input variable corresponding to k _p and Δp _ei ;

4) In the control process of step 3), compare the absolute value of Δω _ri |Δω _ri | corresponding to the current sampling period with Δω _r1 /k ₁ in real time: if |Δω _ri | is greater than Δω _r1 /k ₁ , return Step 3); if |Δω _ri | is less than or equal to Δω _r1 /k ₁ , go to step 5);

Among them, k ₁ is an adjustment factor obtained from empirical data, and k ₁ takes a value between 2.5 and 3.5;

5) Take the current Lp _ei ^* and the theoretical rotational speed change value obtained in the previous control cycle as the two input variables of the first fuzzy inference rule table, obtain the theoretical rotational speed change value according to the first fuzzy inference rule table, and obtain the theoretical rotational speed change value according to the theoretical rotational speed change value to calculate the optimal speed value of the wind turbine; the double-fed control system continuously and dynamically adjusts the speed of the wind turbine according to the corresponding optimal speed value; enter step 6);

Among them, Lp _ei ^* = Δp _ei · k _y , Lp _ei ^* is the active power input variable corresponding to ky and Δp _ei , k _y = k _p · k _{1 ;} k _y is the quantization of the input variable corresponding to Δp _ei after adjustment factor;

6) Compare |Δω _ri | corresponding to the current _sampling period with ε _wmin in real time: if |Δω _ri | is greater than or equal to ε _wmin , return to step 5); if |Δω _ri | The control is switched to optimal reactive power search control;

Among them, _εwmin is the critical switching value corresponding to the maximum wind energy tracking control;

After entering the optimal reactive power search control, control according to the following steps:

1] Continuously sample the reactive power and active power of the wind turbine, and calculate the change value of reactive power and the change value of active power in each reactive power sampling cycle; set a single reactive power sampling cycle The recorded reactive power change value is ΔQ _sv , and the recorded active power change value in a single reactive power sampling period is Δp _v , where v is the number of samples, v=1, 2, 3, 4...n;

2] In the second reactive power sampling cycle, ΔQ _s1 and Δf ₁ are used as the two input variables of the second fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the second fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the theoretical reactive power change value Calculate the optimal reactive power value of the wind-driven generator, and the doubly-fed control system adjusts the reactive power of the wind-driven generator according to the optimized reactive power value; enter step 3]; the second reactive power sampling cycle also forms the first Reactive power control period;

Among them, Δf ₁ is the active power input variable corresponding to Δp ₁ , Δf ₁ =-Δp ₁ k _f , and k _f is the quantization factor of the input variable corresponding to Δp _v ;

3] In the subsequent reactive power control cycle, the current Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle are used as the two input variables of the second fuzzy inference rule table, and according to the second fuzzy inference rule table, the theoretical Reactive power change value, calculate the optimal reactive power value of the wind turbine according to the theoretical reactive power change value, and the double-fed control system continuously and dynamically adjusts the reactive power of the wind turbine according to the optimal reactive power value; enter step 4 ];

Among them, Δf _v is the active power input variable corresponding to Δp _v , Δf _v =-Δp _v k _f ;

4] In the control process of step 3], the absolute value of current ΔQ _sv |ΔQ _sv | is compared with k ₂ |ΔQ _s1 | in real time in each reactive power control cycle:

If |ΔQ _sv | is greater than k ₂ |ΔQ _s1 |, continue to compare the absolute value of Δp _v |Δp _v | with ε _w1 : if |Δp _v | is greater than ε _w1 , stop the optimal reactive power search control, At the same time, the doubly-fed control system adjusts the reactive power of the wind turbine according to the set reactive power rating, and switches to the maximum wind energy tracking control; if |Δp _v |≤ε _w1 , return to step 3];

If |ΔQ _sv | is less than or equal to k ₂ ·|ΔQ _s1 |, enter step 5];

Among them, k ₂ is an adjustment factor obtained from empirical data, and k ₂ takes a value between 0.3 and 0.4;

5] The current Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle are used as the two input variables of the third fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the third fuzzy inference rule table, according to the theoretical The reactive power change value calculates the optimized reactive power value of the wind-driven generator, and the doubly-fed control system continuously and dynamically adjusts the reactive power of the wind-driven generator according to the optimized reactive power value; enter step 6];

6] In the control process of step 5], the real-time comparison between |Δp _v | and ε _w1 is performed in each reactive power control cycle:

If the condition of |Δp _v |>ε _w1 is satisfied, the optimal reactive power search control is stopped, and at the same time, the doubly-fed control system adjusts the reactive power of the wind turbine according to the set reactive power rating, and switches to Maximum wind energy tracking control; if the condition of |Δp _v |≤ε _w1 is met, return to step 5];

Among them, ε _w1 is the critical switching value corresponding to the optimal reactive power search control.

The present invention is different from the existing fuzzy efficiency optimization method in that:

On the one hand, in the process of realizing maximum wind energy tracking, when the optimal rotational speed value found through fuzzy search is gradually approaching the optimal rotational speed reference value, the present invention can only pass Increase the input variable quantization factor k _p to improve the resolution of the control system to the total active input increment Δpe _ei of the generator stator and rotor, so that the control system can respond to the small Δpe _ei , increase its control effect, and reduce the search Oscillation, the optimal speed reference value with higher accuracy is obtained; at the same time, the solution is simple, and after quickly completing the processing of the maximum wind energy tracking, it quickly enters the processing stage of the optimal reactive power search, which can better solve the maximum wind energy A coordination problem between tracking and optimal var search.

On the other hand, in the process of realizing the optimal reactive power search, in order to solve the problem of search oscillation when using a single fuzzy logic search, the present invention adopts a two-stage fuzzy logic search method: use coarse-tuning fuzzy logic (that is, the second fuzzy reasoning in the preceding text) table) to speed up the search, fine-tune the fuzzy logic (that is, the third fuzzy inference table in the previous article) to improve the accuracy of the search, and solve the problem of contradiction between the speed and accuracy in the search using a single fuzzy logic, The copper loss of the generator's own stator and rotor is minimized, which is more conducive to the further improvement of the active power output of the generator.

The beneficial technical effects of the present invention are: in the process of tracking the maximum wind energy, by changing the quantization factor of the input variable of the fuzzy reasoning rule table, the search accuracy of the fuzzy reasoning rule table can be improved, the adjustment dead zone can be avoided, and the wind power generator's ability to wind energy can be improved. In the optimal reactive power tracking process, it can effectively suppress flutter, improve search accuracy, reduce the self-loss of wind turbines, and increase active power output.

Description of drawings

Figure 1. Schematic diagram of the doubly-fed wind turbine control system;

Figure 2. Schematic diagram of the global segment control of doubly-fed wind turbines;

Figure 3. The relationship between the output mechanical power and the rotational speed of the wind turbine at different wind speeds;

Figure 4. Evaluation function f vs. relationship curve;

Fig. 5, realize the schematic diagram of the fuzzy controller principle of the first fuzzy inference rule table;

Fig. 6, the fuzzy membership function of Lp _ei ;

Fig. 7, realize the fuzzy controller schematic diagram of the second fuzzy inference rule table and the third fuzzy inference rule table;

Fig. 8, the fuzzy membership function of the input variable Δf _v of the second fuzzy inference rule table;

Fig. 9, the fuzzy membership function of the input variable ΔQ _sv of the second fuzzy inference rule table;

Fig. 10, the fuzzy membership function of the output variable ΔQ _s of the second fuzzy inference rule table;

The fuzzy membership function of the input variable Δf _v of Fig. 11, the 3rd fuzzy inference rule table;

Fig. 12, the fuzzy membership function of the input variable ΔQ _sv of the third fuzzy inference rule table;

Fig. 13, the fuzzy membership function of the output variable ΔQ _s of the third fuzzy inference rule table;

Fig. 14, the logical block diagram of the present invention.

Detailed ways

The system used for doubly-fed wind turbine control is shown in Figure 1. The system uses a vector control system based on stator flux orientation to realize the decoupling between the generator electromagnetic torque and rotor excitation, and then removes it through feed-forward compensation. After the cross-coupling term caused by the back electromotive force, the generator rotor flux linkage and electromagnetic torque can be controlled respectively by adjusting the d and q axis components of the rotor voltage.

In the actual environment, the speed of the blades changes with the wind speed, and the operating state of the wind turbine will change with the changes of the speed of the blades; When the wind speed is cut, the wind turbines are connected to the grid for power generation; in the maximum wind energy tracking area, the speed of the wind turbines changes accordingly with the wind speed to ensure that the wind energy utilization coefficient of the wind turbines is always maintained at the maximum value. At this time, the wind turbine control subsystem implements constant pitch control, and the generator control subsystem adjusts the speed of the unit through the output power of the generator to achieve variable speed and constant frequency operation; when in the constant speed area, the wind turbine has reached the maximum speed. However, the output power of the wind turbine has not yet reached the rated working state. In order to protect the unit from overload, the maximum wind energy is no longer tracked, but the pitch angle is adjusted through the pitch control of the wind turbine control subsystem to ensure that the maximum speed is allowed The constant speed power generation operation on the wind turbine; as the wind speed increases, the output mechanical power of the wind turbine continues to increase, and the generator reaches its power limit. At this time, it is necessary to control the output power of the unit so that it does not exceed the rated value, and the wind turbine is in a constant speed and constant power operation state.

The control section targeted by the method of the present invention is the maximum wind energy tracking section after the wind power generator starts grid-connected power generation and before entering the constant speed power generation state. The control strategy of the present invention (that is, the dual efficiency fuzzy optimal control in Figure 2) The principle of the global segment control system composed of the constant speed control strategy and the constant power control strategy is shown in Figure 2. The control strategies corresponding to different operating segments of the wind turbine can be switched through an integrated coordination controller; for constant power Both the control strategy and the constant speed control strategy have mature schemes in the prior art, and the present invention is only aimed at the double-efficiency fuzzy optimization control area in Fig. 2; the core idea of the present invention is: in Fig. 1 The value of (generator optimal speed reference given value) adopts the optimal speed value in the maximum wind energy tracking control of the present invention to realize the tracking of the maximum wind energy, as shown in Fig. 1 The value of (optimum reactive power given value) adopts the optimized reactive power value in the optimal reactive power search control of the present invention to realize the operation of the wind turbine with the lowest loss, thereby realizing the control objective of dual efficiency fuzzy optimization.

In order to enable those skilled in the art to better understand the solution of the present invention, the principles of maximum wind energy tracking and optimal reactive power tracking in the present invention are now explained separately:

1) The principle of maximum wind energy tracking: see Figure 3, which shows the relationship between the output mechanical power of the wind turbine and the blade speed under different wind speed conditions; where, P _m is the mechanical power output by the fan, and ω _w is the blade Angular velocity of rotation, v ₁ , v ₂ , and v ₃ are all wind speeds, and v ₁ >v ₂ >v ₃ ; it can be seen from the figure that under the same wind speed, when the pitch angle of the blade is constant, the There are differences in wind energy utilization coefficients corresponding to different speeds, and this difference ultimately makes the mechanical power output by the fan different. However, under each wind speed condition, there is a highest power point P _max corresponding to the optimal blade speed. Here Under the condition of optimal blade speed, the blade can reach the optimal tip speed ratio, thereby capturing the maximum energy at this wind speed. Since the wind generator is driven by the fan, if the output power of the wind generator can correspond to the optimal The rotation speed of the blades can maximize the utilization rate of wind energy. Therefore, the process of maximum wind energy tracking can be understood as effectively controlling the wind turbine speed The process of controlling the output power of the wind generator to match the optimal blade speed by adjusting the speed of the wind generator improves the utilization effect of wind energy.

2) The principle of optimal reactive power tracking: The calculation principle is to select the reactive power value that can make a certain performance evaluation function f reach the optimal value within the allowable operating range of the wind turbine, and the evaluation function f and can be shown in Figure 4, the wind turbine loss characteristics of double-fed control in the case of ignoring the frequency converter loss, the loss related to the stator reactive power Q _s is mainly the generator stator copper loss P _cus and rotor copper loss P _cur , so the evaluation function f can be defined as:

f＝P _cus +P _cur

expands to:

① $f=i_{\mathrm{the\; s}}^2{r}_{\mathrm{the\; s}}+i_r^2{r}_r=(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2)r_{\mathrm{the\; s}}+(i_{\mathrm{rd}}^2+i_{\mathrm{rq}}^2)r_r$

Among them, r _s and r _r are equivalent resistances of stator and rotor respectively; i _sd and i _sq are stator d and q axis currents respectively; i _rd and i _rq are rotor d and q axis currents respectively;

According to the mathematical model of the double-fed motor in the dq synchronous rotating coordinate system, the d-axis coincides with the stator flux ψ _s of the double-fed motor, and there is a flux equation:

②ψ _sd ＝-L _s i _sd +L _m i _rd ＝ψ _s and ③ψ _sq ＝-L _s i _sq +L _m i _rq ＝0

Among them, ψ _sd , ψ _sq are stator d-axis and q-axis flux linkage respectively; L _s , L _m are stator equivalent self-inductance and mutual inductance respectively.

The active power P _s and reactive power Q _s in the dq synchronous rotating coordinate system are respectively:

④P _s ＝U _s i _sq and ⑤Q _s ＝U _s i _sd

Among them, U _s is the stator effective voltage; substituting ②, ③, ④, ⑤ into ① to get:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; a</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; b</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; c</span>}$

Among them, the coefficients a, b, and c are respectively:

$\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}{{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; ,</span>$ $\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}{\mathrm{<span\; class="notranslate">\; \&psi;</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}{{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; ,</span>$ $\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; P</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; P</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&psi;</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}{{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

Obviously, a, b, c are all greater than zero, at synchronous angular velocity ω ₁ , when (Formula ⑥), there are: $f_{\min }=\frac{4\mathrm{ac}-b^2}{4a};$

Due to the optimal reactive power value is always negative, which satisfies the reactive power condition for the stable operation of the doubly-fed generator. In the prior art, it is generally set directly according to formula ⑥ is the reactive power reference; however, due to the accuracy limitation of the generator parameter measurement and the resistance value being significantly affected by the temperature, it is difficult for this strategy to achieve the lowest motor loss in actual operation. Aiming at this problem, the present invention proposes an optimal reactive power search control based on the following ideas: It can be seen from formula ⑥ that although the shape of f will be changed due to the perturbation of the motor parameters, when reducing the loss of the wind power generator optimal as an option When the principle of , since the quadratic curve relationship with the minimum value between f and Q _s is always established, technical means can be used to find the condition that satisfies the minimum loss Thereby reducing the loss of wind turbines;

The "maximum wind energy tracking control" and "optimum reactive power search control" recorded in the previous "Invention Summary" are the solutions proposed based on the aforementioned analysis, which involves the first fuzzy inference rule table, the second fuzzy inference Rule table and the 3rd fuzzy inference rule table are technical tools for realizing the object of the present invention, and these fuzzy inference rules are combined with the present invention like this:

(1) The first fuzzy inference rule table

Referring to Fig. 5, Fig. 5 shows the schematic diagram of the fuzzy inference for realizing the first fuzzy inference rule table: when the blade speed changes, the shaft mechanical loss variation is negligible compared to the absorbed wind energy variation Δp _m ; and p Since _m cannot be directly measured, Δp _m can be replaced by the total active power increment of the wind turbine stator and rotor Δp _e ; Z ^-1 in Figure 5 represents a one-step lag (time delay) link; where k _p is the quantization factor of the input variable ( That is, the "quantization factor of the input variable corresponding to Δp _ei " in the corresponding scheme) to realize the conversion of the total active power increment Δp _ei of the stator and rotor of the wind turbine from the basic domain of discourse to the language variable domain of discourse, and the definition of the domain of discourse of language variables is [-1, 1]; k _w is the scale factor of the output variable; the optimal speed reference fuzzy logic control module in the figure is the core part (i.e. realizes fuzzy inference), which synthesizes the power increment and speed change in each beat sampling A new speed reference is given, which does not depend on parameters such as fan parameters and air density, so this maximum wind energy tracking strategy has strong adaptability to the environment (the existing technology generally uses the dynamic setting of the active power reference of the generator stator to indirect The speed control can realize the maximum wind energy tracking without wind speed measurement, and the setting of the active power of the generator stator is related to the characteristics of the fan. In practice, as the meteorological and geographical conditions change, the air density will change, and the unit will age. Conditions such as dust deposition on blades will also affect the characteristics of the fan, so the active power setting of the generator stator will always deviate from the optimum during the operation of the unit, resulting in poor tracking of the maximum wind energy).

The inference rules of the first fuzzy inference rule table can be set according to the following ideas:

1. In the last control cycle, if both Δω _ri and Δp _ei are positive, it means that the operating point is approaching the extreme point (that is, the optimal speed value), and the new generator speed increment Δω _r (that is, the theoretical speed change value ) should also be positive, and a forward search is required;

2. In the last control cycle, if Δω _ri is positive and Δp _ei is negative, it means that the operating point is moving away from the extreme point. At this time, the new generator speed increment Δω _r should be negative, and a reverse search is required;

In addition to the inverse process of the conditions set by the above-mentioned 1 and 2 inference rules, a fuzzy rule with Lp _ei and Δω _ri as input variables and the theoretical rotational speed change value Δω _r as output variables can be established, as shown in Table 1 Shown:

Table 1

The above table is the first fuzzy inference rule table, which adopts the double-input-single-output mode. The input variable Lp _ei fuzzy discourse contains 7 fuzzy subsets, and the language value on the discourse is {NB, NM, NS, ZE, PS, PM, PB}, namely {negative big, negative middle, negative small, zero, positive small, positive middle, positive big}. Accompanying drawing 6 is the fuzzy membership function of the input variable _Lpei . The membership function uses a triangular function with uneven distribution. The purpose is that when the variable is close to zero, the sensitivity of the membership function increases, so that when the search is approaching the optimal point, the search step size can be adjusted immediately to improve the search efficiency. The Δω _ri fuzzy universe of input variables contains three fuzzy subsets, and the language values on the universe are {N, ZE, P}, namely {negative, zero, positive}. The linguistic value of the output variable Δω _r on the fuzzy universe is the same as that of the input variable Lp _ei .

In steps 4) and 5) of the maximum wind energy tracking control scheme, it involves the adjustment of the input variable quantization factor k _p , the principle is: although in the design of the membership function of the input variables Lp _ei and Δω _ri , the selection of different Uniformly distributed triangular function to improve the sensitivity of the membership function when the variable is close to zero, but the fuzzy controller itself does not have an integral link, so the search accuracy is not high, and there is a static residual error, which is the defect of fuzzy logic search. When approaching the optimal point with a certain search step, there is still an adjustment dead zone in the small deviation range, which causes the search to form a limit cycle oscillation near the optimal point, which will cause long-term vibration of the transmission shaft. This will cause great damage to the mechanical parts of the wind turbine, so further measures need to be taken to improve the search accuracy; in the prior art, when solving the adjustment dead zone in fuzzy reasoning, generally by dividing the files of the fuzzy rule table However, this processing method will also increase the number of rules and the amount of calculation of the system, resulting in a significant increase in system overhead and delay time, affecting the real-time performance of control, and will lead to subsequent control failures. delay. Therefore, when solving the problem of adjusting the dead zone in fuzzy reasoning, a simpler and faster search method should be adopted. The present invention adopts the solutions in steps 4) and 5) to realize simple and faster search:

The schemes in steps 4) and 5) make the fuzzy search of the output variable Δω _r form a variable step-size search in the present invention, that is, the value of |Δω _r | in the early stage is large, and the value of |Δω _r | in the later stage is small. When the fuzzy search is gradually approaching the optimal speed reference value, the fuzzy division on the input domain with a larger given range at first appears rough, and the control accuracy is not high; The search function of the small range of the speed reference increment, the solution starts from the quantization factor k _p of the input variable, which is different from the method in which the quantization factor of the input variable selected by the fuzzy controller is a fixed value, and the process of approaching the optimal point in the later stage of the search Increasing k _p in the middle is equivalent to narrowing the basic domain of the input variable Lp _ei , improving the resolution of the fuzzy controller to the input variable Lp _ei , so that the fuzzy controller can respond to the small Lp _ei , increasing the It controls the action, thereby being able to reduce search oscillations and improve search performance.

In the process of maximum wind energy tracking control, the above scheme can improve the search accuracy and solve the problem of adjusting the dead zone without increasing system overhead and delay; the complete maximum wind energy tracking control scheme is as follows:

1) When the wind speed is greater than or equal to the cut-in wind speed, the wind turbine is started to connect to the grid for power generation, and the speed and active power of the wind turbine are continuously sampled. In each sampling period, the change value of the speed and the change value of active power are analyzed. Calculation; assuming that the rotational speed change value recorded in a single sampling period is Δω _ri , the active power change value recorded in a single sampling period is Δp _ei , i is the number of sampling times, i=1, 2, 3, 4...n; set Both Δω _r1 and Δp _e1 are positive;

2) In the second sampling period, Lp _e1 and Δω _r1 are used as the two input variables of the first fuzzy inference rule table (Δω _r1 is the rotational speed change value collected in the first sampling period, which is also the largest rotational speed change value value), according to the first fuzzy inference rule table to obtain the theoretical speed change value, calculate the optimal speed value of the wind turbine according to the theoretical speed change value, and the double-fed control system adjusts the wind turbine speed according to the optimal speed value; enter step 3 ); the second sampling period also forms the first control period;

Among them, Lp _e1 is the active power input variable corresponding to k _p and Δp _e1 , Lp _e1 = k _p Δp _e1 , k _p is the quantization factor of the input variable corresponding to Δp _ei ;

3) In the subsequent control cycle, the current Lp _ei and the theoretical speed change value obtained in the previous control cycle are used as the two input variables of the first fuzzy inference rule table, and the theoretical speed change value is obtained according to the first fuzzy inference rule table, Calculate the optimal speed value of the wind turbine according to the theoretical speed change value; the double-fed control system continuously and dynamically adjusts the speed of the wind generator according to the corresponding optimal speed value; enter step 4);

Among them, Lp _ei =k _p ·Δp _ei ; Lp _ei is the active power input variable corresponding to k _p and Δp _ei ;

4) In the control process of step 3), compare the absolute value of Δω _ri |Δω _ri | corresponding to the current sampling period with Δω _r1 /k ₁ in real time: if |Δω _ri | is greater than Δω _r1 /k ₁ , it means that the wind force There is still a large margin for the increase of the generator speed, no need to adjust k _p , return to step 3); if |Δω _ri | is less than or equal to Δω _r1 /k ₁ , it means that the increase margin of the wind turbine speed is already small, K _p should be adjusted, go to step 5);

Among them, k ₁ is an adjustment factor obtained from empirical data, and k ₁ takes a value between 2.5 and 3.5;

5) Take the current Lp _ei ^* and the theoretical rotational speed change value obtained in the previous control cycle as the two input variables of the first fuzzy inference rule table, obtain the theoretical rotational speed change value according to the first fuzzy inference rule table, and obtain the theoretical rotational speed change value according to the theoretical rotational speed change value to calculate the optimal speed value of the wind turbine; the double-fed control system continuously and dynamically adjusts the speed of the wind turbine according to the corresponding optimal speed value; enter step 6);

Among them, Lp _ei ^* = Δp _ei · k _y , Lp _ei ^* is the active power input variable corresponding to ky and Δp _ei , k _y = k _p · k _{1 ;} k _y is the quantization of the input variable corresponding to Δp _ei after adjustment factor;

6) When the wind speed enters a steady state, the speed basically fluctuates around the optimal power point, and it is of little significance to continue the maximum wind energy tracking control, and attention should be shifted to reducing the loss of the wind turbine, so by setting a ε _wmin with a smaller value is used to switch the control strategy to the optimal reactive power search control in a timely manner: compare |Δω _ri | corresponding to the current sampling period with ε _wmin in real time: if |Δω _ri | is greater than or equal to ε _wmin , Then return to step 5); if |Δω _ri | is less than ε _wmin , switch from maximum wind energy tracking control to optimal reactive power search control;

Among them, _εwmin is the critical switching value corresponding to the maximum wind energy tracking control, and its value can be determined through experiments or empirical data;

(2) The second fuzzy inference rule table and the third fuzzy inference rule table

Referring to Figure 4, the basic design idea of the fuzzy inference rule table used for optimal reactive power search control is: through fuzzy search, the generator reactive power increment is changed in real time, and at the same time, the change of input loss increment and the reactive power The change of the reference increment, so as to determine the new reactive reference increment, through such a search, the working point can finally be stabilized near the minimum point f _min of the evaluation function.

When the fuzzy controller used for optimal reactive power tracking approaches the optimal point with a certain search step size, there will also be the same adjustment dead zone problem as in the prior art. Therefore, in the design for optimal reactive power search When controlling the fuzzy controller, in order to achieve better results, it should be considered how to take into account the problem of fast search in the early stage and precise positioning in the later stage;

It can be seen from the switching conditions of the maximum wind energy tracking control and the optimal reactive power search control above that the section where the optimal reactive power search control plays a controlling role corresponds to the operating state of the wind turbine when the wind force is relatively stable. In this case, the complexity of the change of the operating state of the wind turbine is simpler than that of the maximum wind energy tracking control stage, and the number of fuzzy inference rules required at this time is relatively small. Therefore, even if the gears of the fuzzy controller are further refined The number of fuzzy inference rules increased accordingly is relatively small, and the negative impact on system overhead and delay is very limited; therefore, in the optimal reactive power search control, the present invention draws lessons from the prior art to deal with the adjustment dead area, two fuzzy controllers (that is, two tables of fuzzy inference rules) are designed to correspond to the two time domains respectively, which are used to meet the rapidity of the early search and the precise positioning when the later stage is close to the optimal reactive value point. requirements.

The system structures of the two fuzzy controllers both adopt the structure shown in Figure 7. Z ^-1 in Figure 7 represents a one-step lag (time delay) link, and the two input variables Δf _v and ΔQ _sv in the figure are the evaluation function f change and reactive power change value, k _f is the quantization factor of the input variable, and the range of discourse of the language variable is defined as [-1, 1]; k _q is the proportional factor of the output variable, and the optimal reactive power reference fuzzy logic control is The core part, it inputs Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle, and outputs a new theoretical reactive power change value. This method does not depend on the accuracy of generator parameters and has good adaptability.

Since the direct calculation of the variation of the evaluation function f requires the use of multiple parameter values such as stator and rotor resistance and inductance, this will complicate the problem. Considering that the wind turbine is already working in the wind speed stable stage, the evaluation function The variation can be replaced by the inverse value of the variation of active power, namely -Δp _v , so Δf _v = -Δp _v k _f ;

The inference rules of the second fuzzy inference rule table can be set according to the following ideas:

1. When both ΔQ _sv and Δf _v are positive, it means that the current operating point is far away from the optimal reactive power value point, then the new ΔQ _s (that is, the theoretical reactive power change value) should be negative, and reverse search is required;

2. When ΔQ _sv is positive and Δf _v is negative, it means that the current working point is approaching the optimal reactive value point, and you should continue to search according to the current search direction;

Coupled with the inverse process of the conditions set by the aforementioned inference rules 1 and 2, it is possible to establish a second model with ΔQ _sv and Δf _v as input variables and the theoretical reactive power change value (i.e. ΔQ _s in Figure 7) as output variable. The table of fuzzy inference rules is shown in Table 2 below:

Table 2,

The second fuzzy inference rule table adopts a double-input-single-output mode. Figures 8, 9, and 10 are the fuzzy membership functions of Δf _v , ΔQ _sv and output variables, respectively. The fuzzy domain of Δf _v contains 5 fuzzy subsets, and the linguistic values on the domain of discourse are {NB, NS, ZE, PS, PB}, that is, {negative large, negative small, zero, positive small, positive large}. The membership function uses a triangular function with uneven distribution. The purpose is that when the variable is close to zero, the sensitivity of the membership function increases, so that when the search is approaching the optimal point, the search step can be adjusted in time to improve the search efficiency. The fuzzy universe of ΔQ _sv contains two fuzzy subsets, and the linguistic value on the universe is {N, P}, that is, {negative, positive}. The linguistic value of the output variable on the fuzzy universe is the same as that of the input variable Δf _v .

The third fuzzy inference rule table is shown in Table 3 below.

table 3

The third fuzzy inference rule table adopts the double-input-single-output mode, and the accompanying drawings 11, 12, and 13 are the fuzzy membership functions of Δf _v , ΔQ _sv and output variables, respectively. The fuzzy domain of Δf _v has 4 more fuzzy subsets than Table 2, and it is 9 fuzzy subsets, and the language value on the domain of discourse is {NVB, NB, NM, NS, ZE, PS, PM, PB, PVB} , that is, {negative large, negative large, negative medium, negative small, zero, positive small, positive medium, positive large, positive large}. The membership function selects a triangular function with uneven distribution, and its purpose is the same as before. The fuzzy universe of ΔQ _sv contains two fuzzy subsets, and the linguistic value on the universe is {N, P}, that is, {negative, positive}. The linguistic value on the fuzzy universe of the output variable is the same as Δf _v .

The complete optimal reactive search control scheme is as follows:

1] Continuously sample the reactive power and active power of the wind turbine, and calculate the change value of reactive power and the change value of active power in each reactive power sampling cycle; set a single reactive power sampling cycle The recorded reactive power change value is ΔQ _sv , and the recorded active power change value in a single reactive power sampling period is Δp _v , where v is the number of samples, v=1, 2, 3, 4...n;

2] In the second reactive power sampling cycle, ΔQ _s1 and Δf ₁ are used as the two input variables of the second fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the second fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the theoretical reactive power change value Calculate the optimal reactive power value of the wind-driven generator, and the doubly-fed control system adjusts the reactive power of the wind-driven generator according to the optimized reactive power value; enter step 3]; the second reactive power sampling cycle also forms the first Reactive power control cycle;

Among them, Δf ₁ is the active power input variable corresponding to Δp ₁ , Δf ₁ =-Δp ₁ k _f , and k _f is the quantization factor of the input variable corresponding to Δp _v ;

3] In the subsequent reactive power control cycle, the current Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle are used as the two input variables of the second fuzzy inference rule table, and according to the second fuzzy inference rule table, the theoretical Reactive power change value, calculate the optimal reactive power value of the wind turbine according to the theoretical reactive power change value, and the double-fed control system continuously and dynamically adjusts the reactive power of the wind turbine according to the optimal reactive power value; enter step 4 ];

Among them, Δf _v is the active power input variable corresponding to Δp _v , Δf _v =-Δp _v k _f ;

4] In the control process of step 3], the absolute value of current ΔQ _sv |ΔQ _sv | is compared with k ₂ |ΔQ _s1 | in real time in each reactive power control cycle:

If |ΔQ _sv | is greater than k ₂ ·|ΔQ _s1 |, it means that the reactive power of the wind turbine is still approaching the reactive power point corresponding to the lowest loss in one direction with a large step size, then continue to change the absolute value of Δp _v | Δp _v | is compared with ε _w1 : if |Δp _v | is greater than ε _w1 , it means that the wind speed change has caused a sudden change in active power, and the maximum wind energy should be re-tracked. At this time, the optimal reactive power search control is stopped. The set reactive power rating adjusts the reactive power of the wind turbine and switches to the maximum wind energy tracking control; if |Δp _v |≤ε _w1 , return to step 3];

If |ΔQ _sv | is less than or equal to k ₂ ·|ΔQ _s1 |, it means that the reactive power of the wind turbine is not far from the reactive power point corresponding to the lowest loss, and it is necessary to conduct a more accurate calculation of the optimal reactive power point positioning, enter step 5];

Among them, k ₂ is an adjustment factor obtained from empirical data, and k ₂ takes a value between 0.3 and 0.4;

5] The current Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle are used as the two input variables of the third fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the third fuzzy inference rule table, according to the theoretical The reactive power change value calculates the optimized reactive power value of the wind-driven generator, and the doubly-fed control system continuously and dynamically adjusts the reactive power of the wind-driven generator according to the optimized reactive power value; enter step 6];

6] In the control process of step 5], the real-time comparison between |Δp _v | and ε _w1 is performed in each reactive power control cycle:

If the condition of |Δp _v |>ε _w1 is met, it means that the change of wind speed has caused a sudden change in active power, and the maximum wind energy should be tracked again, and the optimal reactive power search control should be stopped. Adjust the reactive power of the wind turbine and switch to the maximum wind energy tracking control; if the condition of |Δp _v |≤ε _w1 is met, it means that the wind speed is relatively stable, and then return to step 5];

Among them, _εw1 is the critical switching value corresponding to optimal reactive power search control, and its value can be determined according to experimental or empirical data.

The logical block diagram of the present invention is shown in FIG. 14 .

Claims (1)

1. A double-fed wind generator double-efficiency fuzzy-optimized control method, including a wind generator controlled by a double-fed control system; when the wind speed is greater than or equal to the cut-in wind speed, the wind generator is connected to the grid to generate electricity Large, when the wind turbine reaches the allowable maximum speed, it enters the state of constant speed power generation, which is characterized in that: after the start of grid-connected power generation and before the wind turbine reaches the allowable maximum speed, the following control method is used to control the wind turbine: When the wind turbine is first connected to the grid for power generation, the maximum wind energy tracking control should be carried out as follows: 1) When the wind speed is greater than or equal to the cut-in wind speed, the wind generator is connected to the grid to generate electricity, and the speed and active power of the wind generator are continuously sampled. In each sampling period, the change value of the speed and the change value of the active power are calculated. ; Let the rotational speed change value recorded in a single sampling period be Δω _ri , the active power change value recorded in a single sampling period be Δp _ei , i be the number of samples, i=1, 2, 3, 4...n; set Δω _{Both r1} and Δp _e1 are positive; 2) In the second sampling period, Lp _e1 and Δω _r1 are used as the two input variables of the first fuzzy inference rule table, and the theoretical speed change value is obtained according to the first fuzzy inference rule table, and the wind turbine generator is calculated according to the theoretical speed change value The optimal speed value of the double-fed control system adjusts the speed of the wind turbine according to the optimal speed value; enter step 3); the second sampling cycle also forms the first control cycle; Among them, Lp _e1 is the active power input variable corresponding to k _p and Δp _e1 , Lp _e1 = k _p Δp _e1 , k _p is the quantization factor of the input variable corresponding to Δp _ei ; 3) In the subsequent control cycle corresponding to the first control cycle, the current Lp _ei and the theoretical rotational speed change value obtained in the previous control cycle are used as the two input variables of the first fuzzy inference rule table, according to the first fuzzy inference The rule table obtains the theoretical speed change value, and calculates the optimal speed value of the wind turbine according to the theoretical speed change value; the double-fed control system continuously and dynamically adjusts the wind turbine speed according to the corresponding optimal speed value; enter step 4); Among them, Lp _ei =k _p ·Δp _ei ; Lp _ei is the active power input variable corresponding to k _p and Δp _ei ; 4) In the control process of step 3), compare the absolute value of Δω _ri |Δω _ri | corresponding to the current sampling period with Δω _r1 /k ₁ in real time: if |Δω _ri | is greater than Δω _r1 /k ₁ , return Step 3); if |Δω _ri | is less than or equal to Δω _r1 /k ₁ , go to step 5); Among them, k ₁ is an adjustment factor obtained from empirical data, and k ₁ takes a value between 2.5 and 3.5; 5) Take the current Lp _ei ^* and the theoretical rotational speed change value obtained in the previous control cycle as two input variables of the first fuzzy inference rule table, obtain the theoretical rotational speed change value according to the first fuzzy inference rule table, and obtain the theoretical rotational speed change value according to the theoretical rotational speed change value to calculate the optimal speed value of the wind-driven generator; the doubly-fed control system carries out continuous and dynamic adjustment to the wind-driven generator speed according to the corresponding optimized speed value; enter step 6); Among them, Lp _ei ^* = Δp _ei · k _y , Lp _ei ^* is the active power input variable corresponding to ky and Δp _ei , k _y = k _p · k _{1 ;} k _y is the quantization of the input variable corresponding to Δp _ei after adjustment factor; 6) Compare |Δω _ri | corresponding to the current _sampling period with ε _wmin in real time: if |Δω _ri | is greater than or equal to ε _wmin , return to step 5); if |Δω _ri | The control is switched to optimal reactive power search control; Among them, _εwmin is the critical switching value corresponding to the maximum wind energy tracking control; After entering the optimal reactive power search control, control according to the following steps: 1] Continuously sample the reactive power and active power of the wind turbine, and calculate the change value of reactive power and the change value of active power in each reactive power sampling cycle; set a single reactive power sampling cycle The recorded reactive power change value is ΔQ _sv , and the recorded active power change value in a single reactive power sampling period is Δp _v , where v is the number of samples, v=1, 2, 3, 4...n; 2] In the second reactive power sampling cycle, ΔQ _s1 and Δf ₁ are used as the two input variables of the second fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the second fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the theoretical reactive power change value Calculate the optimal reactive power value of the wind-driven generator, and the doubly-fed control system adjusts the reactive power of the wind-driven generator according to the optimized reactive power value; enter step 3]; the second reactive power sampling cycle also forms the first Reactive power control cycle; Among them, Δf ₁ is the active power input variable corresponding to Δp ₁ , Δf ₁ =-Δp ₁ k _f , and k _f is the quantization factor of the input variable corresponding to Δp _v ; 3] In the subsequent reactive power control cycle, the current Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle are used as the two input variables of the second fuzzy inference rule table, and according to the second fuzzy inference rule table, the theoretical Reactive power change value, calculate the optimized reactive power value of the wind turbine according to the theoretical reactive power change value, and the double-fed control system continuously and dynamically adjusts the reactive power of the wind turbine according to the optimized reactive power value; enter step 4 ]; Among them, Δf _v is the active power input variable corresponding to Δp _v , Δf _v =-Δp _v k _f ; 4] In the control process of step 3], the absolute value of current ΔQ _sv |ΔQ _sv | is compared with k ₂ |ΔQ _s1 | in real time in each reactive power control cycle: If |ΔQ _sv | is greater than k ₂ |ΔQ _s1 |, continue to compare the absolute value of Δp _v |Δp _v | with ε _w1 : if |Δp _v | is greater than ε _w1 , stop the optimal reactive power search control, At the same time, the doubly-fed control system adjusts the reactive power of the wind turbine according to the set reactive power rating, and switches to the maximum wind energy tracking control; if |Δp _v |≤ε _w1 , return to step 3]; If |ΔQ _sv | is less than or equal to k ₂ ·|ΔQ _s1 |, enter step 5]; Among them, k ₂ is an adjustment factor obtained from empirical data, and k ₂ takes a value between 0.3 and 0.4; 5] The current Δf _v and the theoretical reactive power change value obtained in the previous reactive power control cycle are used as the two input variables of the third fuzzy inference rule table, and the theoretical reactive power change value is obtained according to the third fuzzy inference rule table, according to the theoretical The reactive power change value calculates the optimized reactive power value of the wind-driven generator, and the doubly-fed control system continuously and dynamically adjusts the reactive power of the wind-driven generator according to the optimized reactive power value; enter step 6]; 6] In the control process of step 5], the real-time comparison between |Δp _v | and ε _w1 is performed in each reactive power control cycle: If the condition of |Δp _v |>ε _w1 is satisfied, the optimal reactive power search control is stopped, and at the same time, the doubly-fed control system adjusts the reactive power of the wind turbine according to the set reactive power rating, and switches to Maximum wind energy tracking control; if the condition of |Δp _v |≤ε _w1 is met, return to step 5]; Among them, ε _w1 is the critical switching value corresponding to the optimal reactive power search control.