CN104967149B - A kind of micro-capacitance sensor wind-light storage model predictive control method - Google Patents

Abstract

The invention discloses a forecasting control method of a micro-grid wind-solar-storage model, comprising the following steps: establishing a forecasting model, using the forecasting model to predict the maximum output of wind turbines and photovoltaic power generation in a micro-grid within a set period of time in the future; using the predicted wind power and The maximum output of photovoltaics is used as a constraint condition, and the output of wind turbines, photovoltaic power generation and energy storage batteries in the microgrid is optimized online, and the reference output of the three is given; according to the real-time adjustable capacity of wind turbines and photovoltaic power generation, the wind turbines , photovoltaic power generation and energy storage battery for feedback adjustment with reference to output. Compared with the traditional micro-grid control method, this control method reduces the requirements for the accuracy of the prediction model of the uncertain process, and makes up for the defects of low accuracy of wind power and photovoltaic prediction models and strong output uncertainty that are difficult to solve by traditional control methods , effectively improving the operating characteristics of the microgrid.

Description

A Model Predictive Control Method for Microgrid Wind, Wind and Storage

technical field

The invention relates to a model predictive control method of wind-solar-storage-storage in a microgrid.

Background technique

Distributed power generation has the advantages of resource and environment friendliness and flexible power supply. The development of distributed power generation on the basis of centralized power generation and large power grid has become an inevitable trend in the development of smart grids at home and abroad. According to the national energy development plan, by 2020 my country's distributed power generation installed capacity will account for 9%, which is an important part of the power supply system. In order to solve the problems brought about by the grid connection of distributed generation, the concept of microgrid was put forward. Based on domestic and foreign research results, microgrid refers to a micro power supply network based on distributed power generation technology, mainly based on renewable energy, using energy storage and control devices to achieve power balance within the network.

The uncertainty of wind power and photovoltaic output is large, it is difficult to predict, and its prediction accuracy is also very low, and the longer the forecast is in advance, the greater the forecast error. The integration of wind power and photovoltaics increases the uncertainty of micro-grid operation. The existing micro-grid control methods require high accuracy of prediction models for uncertain processes. Therefore, it is urgent to find control methods that can better deal with uncertainties. .

Model predictive control (model predictive control, MPC) is an effective way to solve this problem. Model predictive control has been widely used in industrial process control. It has strong adaptability and robustness to the model, and is very suitable for dealing with problems with large uncertainty in the system model. Model predictive control is essentially a model-based finite-time domain closed-loop optimal control algorithm. In each sampling period, the controller takes the current system state as the initial state of control, and based on the prediction results of the future state based on the prediction model, The current control behavior is obtained by solving a finite-time optimal control problem through online scrolling, so that the difference between the future output and the reference trajectory is minimized.

Contents of the invention

In order to solve the above problems, the present invention proposes a model predictive control method for microgrid wind-solar-storage storage. This method uses the idea of model predictive control to establish a control model of prediction-online optimization-feedback. The maximum output of wind turbines and photovoltaic power generation in the grid; the maximum output of wind power and photovoltaic power predicted by the prediction model is used as a constraint condition, and the output of wind turbines, photovoltaic power generation, and energy storage batteries in the microgrid is optimized online, and a certain future is given. The reference output of the three within the time period; according to the real-time adjustable capacity of the wind turbine and photovoltaic power generation, the reference output of the wind, wind and storage is adjusted at the moment of control. The analysis of the embodiment shows that this method can better cope with the uncertainty of wind power and photovoltaic output, and effectively improve the operating characteristics of the microgrid.

In order to achieve the above object, the present invention adopts the following technical solutions:

A microgrid wind-solar-storage model predictive control method, comprising the following steps:

(1) Establish a prediction model, and use the prediction model to predict the maximum output of wind turbines and photovoltaic power generation in the microgrid within a set period of time in the future;

(2) Taking the predicted maximum output of wind power and photovoltaics as a constraint condition, the output of wind turbines, photovoltaic power generation and energy storage batteries in the microgrid is optimized online, and the reference output of the three is given;

(3) According to the real-time adjustable capacity of wind turbines and photovoltaic power generation, feedback and adjust the reference output of wind turbines, photovoltaic power generation and energy storage batteries.

In the step (1), the prediction model realizes the power prediction of the wind turbine and photovoltaic power generation through the neural network prediction technology, and predicts the future output value of the process according to the historical information and future input of the process.

The method for establishing the prediction model in the present invention is the neural network prediction technology, and the process of this method is well known in the industry.

The objective function of the online optimization model in the step ( ₂ ): the objective function f1 is that the deviation of the exchanged power between the microgrid and the distribution network is the smallest, so as to ensure that the microgrid becomes a stable power source or load of the distribution network, and reduce the load of the distribution network. Control difficulty. In the off-grid state of the microgrid, set the exchange power between the microgrid and the distribution network to be ₀ , and the objective function f2 is to minimize the difference between the state of charge of the energy storage battery at the beginning and end of the period, so as to ensure that the energy storage battery has a high electricity;

minf ₂ =SOC(1)-SOC(N)

Among them, N is the total number of periods in the optimization time domain; P _l (i) is the load power in period i; P _tie (i) is the exchange power between the microgrid and the distribution network in period i, and a positive value indicates that the microgrid is Injected power of distribution network, a negative value means that distribution network injects power into microgrid; P _w (i) is the planned output of wind turbines in period i; P _PV (i) is the planned output of photovoltaic power generation in period i; P _b (i) is the planned output of the energy storage battery in the period i, the value is positive when charging, and the value is negative when discharging; State of charge of the energy storage battery.

The constraint conditions of the online optimization model in the step (2): include wind turbine output, photovoltaic power generation output and energy storage battery charge and discharge times constraints:

0<P _w (i)<P _{w_pre} (i)

0<P _PV (i)<P _{PV_pre} (i)

N _{ch_dis} <N _battery

Among them, P _{w_pre} (i) is the predicted value of wind turbine output in period i; P _{pv_pre} (i) is the predicted value of photovoltaic power generation output in period i; N _{ch_dis} is the charge and discharge times of energy storage battery; N _battery is the energy storage battery The maximum allowable charge and discharge times.

The specific method of the solution algorithm of the online optimization model in described step (2) is: this optimization model comprises 2 objective functions, belongs to multi-objective optimization problem, and multi-objective optimization algorithm has 3 performance evaluation indexes: 1. the obtained solution Try to get as close to the Pareto optimal solution as possible; ② try to keep the distribution and diversity of the solution population; ③ prevent the loss of the obtained Pareto optimal solution during the solution process, and use the NSGA-Ⅱ algorithm to solve the online optimization model.

The specific method of the real-time adjustable capacity in the step (3) is: the real-time adjustable capacity ΔP _w of the wind turbine is:

ΔP _w ＝P _{w_est} -P _w

Among them, P _w is the reference output of the wind turbine; P _{w_est} is the estimated value of the real-time output of the wind turbine, which can be quickly calculated from the real-time wind speed and the operating status of the wind turbine;

The real-time adjustable capacity of photovoltaic power generation △P _pv is:

ΔP _pv =P _{pv_est} -P _pv

Among them, P _pv is the reference output of photovoltaic power generation; P _{pv_est} is the estimated value of real-time output of photovoltaic power generation, which is calculated by real-time light intensity and temperature.

Feedback adjustment in the step (3): According to whether the wind turbine and photovoltaic power generation have adjustable capacity, the feedback adjustment link specifically includes the following four situations:

(a) Both wind turbines and photovoltaic power generation have adjustable capacity, that is, the adjustable capacity is positive. At this time, the reference output will be directly sent to wind turbines, photovoltaic power generation, and energy storage batteries;

(b) Both the wind turbine and photovoltaic power generation have no adjustable capacity, that is, the adjustable capacity is negative. At this time, the reference output is directly sent to the wind turbine and photovoltaic power generation, and the energy storage battery compensates for the power shortage;

(c) Wind turbines have adjustable capacity, photovoltaic power generation has no adjustable capacity, and photovoltaic power generation has a surplus plan. At this time, the photovoltaic power generation surplus plan is transferred to the wind turbine generator;

(d) Photovoltaic power generation has adjustable capacity, but wind turbines have no adjustable capacity, and wind turbines have surplus plans. At this time, the surplus plan of wind turbines is transferred to photovoltaic power generation.

In the step (c), the reference output is adjusted as follows:

P _{w_sch} ＝P _w +min(ΔP _w ,-ΔP _pv )

P _{pv_sch} ＝P _pv -min(ΔP _w ,-ΔP _pv )

Among them, P _{w_sch} is the final planned contribution of wind turbines; P _{pv_sch} is the final planned contribution of photovoltaic power generation.

In the step (d), the reference output is adjusted as follows:

P _{pv_sch} ＝P _pv +min(ΔP _pv ,-ΔP _w )

P _{w_sch} = P _{w_est -} min(ΔP _pv , -ΔP _w ).

The beneficial effects of the present invention are:

(1) Drawing on the idea of model predictive control, the present invention provides a model predictive control method for microgrid wind-solar-storage storage, and establishes a control model of prediction-online optimization-feedback;

(2) The control method adopts an online optimization strategy based on the actual output feedback, so that the control process can correct the influence of the prediction error in time;

(3) Compared with the traditional micro-grid control method, this control method reduces the requirements for the accuracy of the prediction model of the uncertain process, and makes up for the low accuracy of the wind power and photovoltaic prediction model and the uncertainty of the output that the traditional control method is difficult to solve. Strong defects can effectively improve the operating characteristics of the microgrid.

Description of drawings

Fig. 1 is a schematic flow chart of a microgrid wind-solar-storage model predictive control method provided by the present invention;

Fig. 2 is a schematic diagram of the predicted maximum output of wind power and the actual maximum output curve in the embodiment of the present invention;

Fig. 3 is a schematic diagram of the predicted maximum output of photovoltaics and the actual maximum output in the embodiment of the present invention;

Fig. 4 is a schematic diagram of a reference output curve for wind-solar-storage optimization in an embodiment of the present invention;

Fig. 5 is a schematic diagram of the reference output curve of wind and rain after feedback adjustment in the embodiment of the present invention;

Fig. 6 is a schematic diagram of the real-time output curve of wind and storage in the embodiment of the present invention;

Fig. 7 is a schematic diagram of the SOC variation curve of the energy storage battery in the embodiment of the present invention.

detailed description:

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

A microgrid wind-solar-storage model predictive control method, comprising the following steps:

Step (1): Use the prediction model to predict the maximum output of wind turbines and photovoltaic power generation in the microgrid within a certain period of time in the future;

Step (2): Taking the maximum output of wind power and photovoltaic power predicted in step (1) as a constraint condition, online optimization is performed on the output of wind turbines, photovoltaic power generation and energy storage batteries in the microgrid, and the output of the three in a certain period of time in the future is given. the reference output;

Step (3): According to the real-time adjustable capacity of wind turbines and photovoltaic power generation, feedback and adjust the reference output of wind, wind and storage in step (2) at the moment of control.

The prediction model in the step (1): the function of the prediction model is to predict the future output value of the process according to the historical information of the process and the future input, and provide prior knowledge for the optimization of the model predictive control. The prediction model only focuses on the function of the model, not the form of the model. As long as the model has the function of predicting the future dynamic function of the system, no matter what form it has, it can be used as a prediction model. In the present invention, the prediction model realizes the power prediction of wind turbines and photovoltaic power generation through neural network prediction technology.

The objective function of the online optimization model in the step ( ₂ ): the objective function f1 is that the deviation of the exchanged power between the microgrid and the distribution network is the smallest, so as to ensure that the microgrid becomes a stable power source or load of the distribution network, and reduce the load of the distribution network. Control difficulty. In the off-grid state of the microgrid, set the exchange power between the microgrid and the distribution network to be 0. _The objective function f2 is to minimize the difference between the state of charge of the energy storage battery at the beginning and end of the period, so as to ensure that the energy storage battery has a relatively high power.

minf ₂ =SOC(1)-SOC(N)

Among them, N is the total number of periods in the optimization time domain; P _l (i) is the load power in period i; P _tie (i) is the exchange power between the microgrid and the distribution network in period i, and a positive value indicates that the microgrid is Injected power of distribution network, a negative value means that distribution network injects power into microgrid; P _w (i) is the planned output of wind turbines in period i; P _PV (i) is the planned output of photovoltaic power generation in period i; P _b (i) is the planned output of the energy storage battery in the period i, the value is positive when charging, and the value is negative when discharging; State of charge of the energy storage battery.

The constraint conditions of the online optimization model in the step (2): include wind turbine output, photovoltaic power generation output, and energy storage battery charge and discharge times constraints.

0<P _w (i)<P _{w_pre} (i)

0<P _PV (i)<P _{PV_pre} (i)

N _{ch_dis} <N _battery

Among them, P _{w_pre} (i) is the predicted value of wind turbine output in period i; P _{pv_pre} (i) is the predicted value of photovoltaic power generation output in period i; N _{ch_dis} is the charge and discharge times of energy storage battery; N _battery is the energy storage battery The maximum allowable charge and discharge times, generally set to 1.

The solution algorithm of the online optimization model in the step (2): the optimization model includes two objective functions, which belongs to the multi-objective optimization problem. The multi-objective optimization algorithm has three main performance evaluation indicators: ① The obtained solution should be as close to the Pareto optimal solution as possible; ② The distribution and diversity of the solution population should be kept as far as possible; The optimal solution is lost. Correspondingly, the fast non-dominated sorting genetic algorithm with elite strategy (NSGA-II) has three key technologies that make it an excellent multi-objective optimization algorithm, namely fast non-dominated sorting, individual crowding distance and elite strategy, so The invention adopts the NSGA-II algorithm to solve the online optimization model.

The real-time adjustable capacity in the step (3): Compared with the control time, the online optimization stage has a large advance amount and a large prediction error, and the optimized wind turbine and photovoltaic power generation reference output curve may be higher than that of the wind turbine. And the actual maximum output of photovoltaic power generation, resulting in the inability to accurately complete the reference output curve, increasing the power compensation pressure of the energy storage battery. In order to improve the completion of the reference output of the wind turbine and photovoltaic power generation, the reference output of the wind turbine and photovoltaic power generation is adjusted according to the real-time adjustable capacity of the wind turbine and photovoltaic power generation at the moment of control, so that the energy storage battery can operate according to the online optimization curve as much as possible.

The real-time adjustable capacity ΔP _w of the wind turbine is:

ΔP _w ＝P _{w_est} -P _w

Among them, P _w is the reference output of the wind turbine; P _{w_est} is the estimated value of the real-time output of the wind turbine, which can be quickly calculated from the real-time wind speed and the operating status of the wind turbine.

The real-time adjustable capacity of photovoltaic power generation △P _pv is:

ΔP _pv =P _{pv_est} -P _pv

Among them, P _pv is the reference output of photovoltaic power generation; P _{pv_est} is the estimated value of real-time output of photovoltaic power generation, which can be quickly calculated through real-time light intensity and temperature.

Feedback adjustment in the step (3): According to whether the wind turbine and photovoltaic power generation have adjustable capacity, the feedback adjustment link specifically includes the following four situations.

(1) Both wind turbines and photovoltaic power generation have adjustable capacity, that is, the adjustable capacity is positive. At this time, the reference output is directly sent to wind turbines, photovoltaic power generation, and energy storage batteries.

(2) Both the wind turbine and photovoltaic power generation have no adjustable capacity, that is, the adjustable capacity is negative. At this time, the reference output is directly sent to the wind turbine and photovoltaic power generation, and the energy storage battery compensates for the power shortage.

(3) Wind turbines have adjustable capacity, photovoltaic power generation has no adjustable capacity, and photovoltaic power generation has a surplus plan. At this time, the surplus plan of photovoltaic power generation is transferred to wind turbine generators. The reference output adjustment is as follows:

P _{w_sch} ＝P _w +min(ΔP _w ,-ΔP _pv )

P _{pv_sch} ＝P _pv -min(ΔP _w ,-ΔP _pv )

Among them, P _{w_sch} is the final planned contribution of wind turbines; P _{pv_sch} is the final planned contribution of photovoltaic power generation.

(4) Photovoltaic power generation has adjustable capacity, but wind turbines have no adjustable capacity, and wind turbines have surplus plans. At this time, the surplus plan of wind turbines is transferred to photovoltaic power generation. The reference output adjustment is as follows:

P _{pv_sch} ＝P _pv +min(ΔP _pv ,-ΔP _w )

P _{w_sch} = P _{w_est -} min(ΔP _pv , -ΔP _w ).

According to the flow of the microgrid wind-solar-storage model predictive control method shown in Fig. 1, the implementation program of the micro-grid wind-solar-storage model predictive control algorithm is compiled. The test parameters in the embodiment are set as follows: fan capacity is 66kW, photovoltaic capacity is 200kW, energy storage is 90kW/270kWh, load capacity is 120kW, the exchange power between distribution network and microgrid is 60kW, and the initial SOC of energy storage battery is 0.5.

The predicted maximum output and actual maximum output curves of wind turbines and photovoltaic power generation are shown in Figure 2 and Figure 3, respectively. The predicted value of wind power is too large, which is greater than the actual maximum output in the whole period. The prediction results of photovoltaic power generation can better match the actual maximum output changes, and the prediction accuracy is relatively high.

Set the number of online optimization periods to 15, each period to 1 minute, and the total optimization time to 15 minutes. The NSGA-Ⅱ algorithm is used for optimization, the population number is 400, and the number of generations is 200. The optimization algorithm takes 1 minute and 10 seconds, which can meet the requirements of online control. Typical Pareto optimization schemes are shown in Table 1.

Table 1 Typical Pareto optimization scheme

f ₁ /kW f ₂ plan 1 0.1613 0.4907 Scenario 2 2.7910 0.6017

Taking the objective function f1 as the primary preference, the objective function _f2 as the secondary preference, and choosing scheme ₁ as the optimal control scheme. The optimized reference output curves of wind turbines, photovoltaic power generation and energy storage batteries in Scheme 1 are shown in Figure 4.

According to real-time wind speed, light and temperature conditions, the reference output of wind turbines and photovoltaic power generation is adjusted. The adjusted reference output curve is shown in Figure 5. Wind turbines transfer excess power to photovoltaic power generation.

The real-time output of wind turbines and photovoltaic power generation can better track the adjusted reference output, as shown in Figure 6. The energy storage battery can basically operate according to the reference output curve given by online optimization, which effectively reduces the power compensation pressure of the energy storage battery. The discharge process of the energy storage battery is shown in Figure 7.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (8)

1. A microgrid wind-solar-storage model predictive control method, characterized in that: comprising the following steps: (1) Establish a prediction model, and use the prediction model to predict the maximum output of wind turbines and photovoltaic power generation in the microgrid within a set period of time in the future; (2) Taking the predicted maximum output of wind power and photovoltaics as a constraint condition, the output of wind turbines, photovoltaic power generation and energy storage batteries in the microgrid is optimized online, and the reference output of the three is given; (3) According to the real-time adjustable capacity of wind turbines and photovoltaic power generation, feedback and adjust the reference output of wind turbines, photovoltaic power generation and energy storage batteries; Feedback adjustment in the step (3): According to whether the wind turbine and photovoltaic power generation have adjustable capacity, the feedback adjustment link specifically includes the following four situations: (a) Both wind turbines and photovoltaic power generation have adjustable capacity, that is, the adjustable capacity is positive. At this time, the reference output will be directly sent to wind turbines, photovoltaic power generation, and energy storage batteries; (b) Both the wind turbine and photovoltaic power generation have no adjustable capacity, that is, the adjustable capacity is negative. At this time, the reference output is directly sent to the wind turbine and photovoltaic power generation, and the energy storage battery compensates for the power shortage; (c) Wind turbines have adjustable capacity, photovoltaic power generation has no adjustable capacity, and photovoltaic power generation has a surplus plan. At this time, the photovoltaic power generation surplus plan is transferred to the wind turbine generator; (d) Photovoltaic power generation has adjustable capacity, but wind turbines have no adjustable capacity, and wind turbines have surplus plans. At this time, the surplus plan of wind turbines is transferred to photovoltaic power generation. 2. A kind of microgrid wind-solar-storage model predictive control method as claimed in claim 1, it is characterized in that: in described step (1), predictive model realizes the power prediction of wind turbine unit and photovoltaic power generation through neural network predictive technology, according to The historical information of the process and the future input predict the future output value of the process. 3. A kind of microgrid wind-solar-storage model predictive control method as claimed in claim 1, it is characterized in that: the objective function of the online optimization model in the described step (2): objective function f ₁ is the microgrid and distribution network The exchange power deviation is the smallest to ensure that the microgrid becomes a stable power source or load of the distribution network and reduce the control difficulty of the distribution network. In the off-grid state of the microgrid, set the exchange power between the microgrid and the distribution network as 0, and the objective function _2. The difference between the state of charge of the energy storage battery at the beginning and end of the period is the smallest, so as to ensure that the energy storage battery has a relatively high power; <mrow> <mi>min</mi> <mi> </mi> <msub> <mi>f</mi> <mn>1</mn> </msub> <mo>=</mo> <mfrac> <msqrt> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msup> <mrow> <mo>&amp;lsqb;</mo> <msub> <mi>P</mi> <mi>l</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>P</mi> <mrow> <mi>t</mi> <mi>i</mi> <mi>e</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>P</mi> <mi>w</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>P</mi> <mrow> <mi>P</mi> <mi>V</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>P</mi> <mi>b</mi> </msub> <mrow> <mo>(</mo> <mi>i</mi> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mn>2</mn> </msup> </mrow> </msqrt> <mi>N</mi> </mfrac> </mrow> min f ₂ =SOC(1)-SOC(N) Among them, N is the total number of periods in the optimization time domain; P _l (i) is the load power in period i; P _tie (i) is the exchange power between the microgrid and the distribution network in period i, and a positive value indicates that the microgrid is Injected power of distribution network, a negative value means that distribution network injects power into microgrid; P _w (i) is the planned output of wind turbines in period i; P _PV (i) is the planned output of photovoltaic power generation in period i; P _b (i) is the planned output of the energy storage battery in the period i, the value is positive when charging, and the value is negative when discharging; State of charge of the energy storage battery. 4. A kind of predictive control method of microgrid wind-solar-storage model as claimed in claim 1, it is characterized in that: the constraints of the online optimization model in the step (2): include wind turbine output, photovoltaic power generation output and energy storage Battery charge and discharge times constraints: 0<P _w (i)<P _{w_pre} (i) 0<P _PV (i)<P _{PV_pre} (i) N _{ch_dis} <N _battery Among them, P _{w_pre} (i) is the predicted value of wind turbine output in period i; P _{pv_pre} (i) is the predicted value of photovoltaic power generation output in period i; N _{ch_dis} is the charge and discharge times of energy storage battery; N _battery is the energy storage battery The maximum allowable charge and discharge times, P _w (i) is the planned output of wind turbines in period i; P _PV (i) is the planned output of photovoltaic power generation in period i. 5. A method for predictive control of a microgrid wind-solar-storage model as claimed in claim 1, characterized in that: the specific method for solving the online optimization model in the step (2) is as follows: the optimization model contains 2 objective functions , belongs to the multi-objective optimization problem. The multi-objective optimization algorithm has three performance evaluation indicators: ① The obtained solution should be as close as possible to the Pareto optimal solution; ② The distribution and diversity of the solution population should be kept as far as possible; To prevent the loss of the obtained Pareto optimal solution, the NSGA-Ⅱ algorithm is used to solve the online optimization model. 6. A microgrid wind-storage-storage model predictive control method as claimed in claim 1, characterized in that: the specific method of the real-time adjustable capacity in the step (3) is: the real-time adjustable capacity of wind turbines ΔP _w for: ΔP _w ＝P _{w_est} -P _w Among them, P _w is the reference output of the wind turbine; P _{w_est} is the estimated value of the real-time output of the wind turbine, which can be quickly calculated from the real-time wind speed and the operating status of the wind turbine; The real-time adjustable capacity of photovoltaic power generation △P _pv is: ΔP _pv =P _{pv_est} -P _pv Among them, P _pv is the reference output of photovoltaic power generation; P _{pv_est} is the estimated value of real-time output of photovoltaic power generation, which is calculated by real-time light intensity and temperature. 7. A microgrid wind-solar-storage model predictive control method as claimed in claim 6, characterized in that: in the step (c), the reference output is adjusted as follows: P _{w_sch} ＝P _w +min(ΔP _w ,-ΔP _pv ) P _{pv_sch} ＝P _pv -min(ΔP _w ,-ΔP _pv ) Among them, P _{w_sch} is the final planned output of wind turbines; P _{pv_sch} is the final planned output of photovoltaic power generation, ΔP _pv is the real-time adjustable capacity of photovoltaic power generation, P _w is the reference output of wind turbines, and P _pv is the reference output of photovoltaic power generation. 8. A microgrid wind-solar-storage model predictive control method as claimed in claim 6, characterized in that: in the step (d), the reference output is adjusted as follows: P _{pv_sch} ＝P _pv +min(ΔP _pv ,-ΔP _w ) P _{w_sch} ＝P _{w_est -min} (ΔP _pv ,-ΔP _w ) Among them, P _{w_sch} is the final planned output of wind turbines; P _{pv_sch} is the final planned output of photovoltaic power generation, ΔP _pv is the real-time adjustable capacity of photovoltaic power generation, P _w is the reference output of wind turbines, and P _pv is the reference output of photovoltaic power generation.