CN107947166B - Multi-energy micro-grid time-varying scheduling method and device based on dynamic matrix control - Google Patents

Abstract

The invention provides a dynamic matrix control-based multi-energy microgrid time-varying scheduling method and a device, wherein the method comprises the following steps of S1: predicting the total load demand of the multi-energy microgrid and the total output power of the uncontrollable micro sources according to historical data and weather information, and solving the total output power of the controllable micro sources; step S2: respectively solving the optimal power flow by using the total output power of the controllable micro-sources as constraint conditions through an economic optimal objective function to obtain the expected value of the output power of each controllable micro-source; step S3: taking the output power expected value as an input parameter, and respectively solving the output power predicted values of various controllable micro sources according to a prediction model controlled by a dynamic matrix; step S4: and correcting the output power predicted value and the output power actual measured value, and determining a scheduling scheme according to the corrected output power predicted value. The method has the characteristics of simple PID structure, convenient parameter adjustment and strong robustness of DMC rolling optimization, and introduces a feedback correction link to correct the influence of random factors on the scheduling result.

Description

A multi-energy microgrid time-varying scheduling method and device based on dynamic matrix control

technical field

The invention relates to the field of electric power technology, in particular to a multi-energy microgrid time-varying scheduling method and device based on dynamic matrix control.

Background technique

In recent years, due to the large demand for energy and the contradiction that traditional fossil energy is on the verge of exhaustion has become increasingly prominent, the development of renewable clean energy such as wind energy and solar energy has also received people's attention. Multi-energy complementary independent power system refers to a small-scale power system formed by fully utilizing renewable energy such as hydropower, wind energy, and solar energy according to local conditions. It can also be called a hybrid power generation system or a new energy microgrid. The proposal of multi-energy complementary independent power system also solves the power supply problem in quite a few remote areas. However, as these new energy micro-grids are connected to the power grid, the safe operation of the power grid is facing severe challenges, especially the access of a large number of renewable intermittent distributed power sources, energy storage devices, and flexible loads to ensure the economic operation and safety of the power grid. Optimal scheduling is a key problem to be solved urgently. The current optimal scheduling method is mainly based on the open-loop optimal scheduling control of an optimal period in the future, which cannot timely and effectively correct the deviation of optimal scheduling results caused by random factors such as wind power and photovoltaic forecast errors.

Contents of the invention

In order to solve the problems existing in the prior art, the present invention provides a multi-energy micro-grid time-varying scheduling method and device based on dynamic matrix control, which has the advantages of simple PID structure, convenient parameter adjustment, DMC rolling optimization, and strong robustness It introduces a feedback correction link to correct the influence of random factors such as wind power and photovoltaic forecast errors on the scheduling results.

The present invention provides a multi-energy microgrid time-varying scheduling method based on dynamic matrix control, including:

Step S1: According to the historical data and weather information, the neural network time series is used to predict the total load demand of the multi-energy microgrid and the total output power of the uncontrollable micro-sources in the multi-energy micro-grid, and calculate the controllable micro-sources in the multi-energy microgrid. Total output power;

Wherein, the total load demand of the multi-energy micro-grid includes the total output power of the uncontrollable micro-source and the total output power of the controllable micro-source;

Step S2: Taking the total output power of the controllable micro-source obtained in step S1 as the constraint condition, the optimal power flow is solved respectively through the economic optimal objective function, and the flexible load, energy storage battery and controllable distribution in the controllable micro-source are obtained The expected value of the output power of the power supply;

Step S3: taking the expected output power values of the three controllable micro-sources obtained in step S2 as input parameters, and solving the output powers of the three controllable micro-sources respectively according to the prediction model of dynamic matrix control and the optimized performance index function Predictive value;

Step S4: Correct the predicted output power values of the three controllable micro-sources obtained in step S3 and the actual measured output power values of the three controllable micro-sources, and determine the schedule according to the corrected output power predicted values Program;

Wherein, the economic optimal objective function is used to solve the three kinds of controllable micro-sources when the operation scheduling cost of the multi-energy microgrid is the lowest and the impact of the pollutants on the environment during the operation of the multi-energy microgrid is guaranteed to be the lowest. output power expectations.

The method provided by the present invention first uses historical data and weather information to perform time series prediction through the existing neural network to obtain the total load demand of the multi-energy micro-grid and the total output power of the uncontrollable micro-sources in the multi-energy micro-grid. Subtract the total output power of the uncontrollable micro-source from the load demand to obtain the total output power of the controllable micro-source. Because in power dispatching, only controllable micro-sources can be regulated, so the method of the present invention needs to obtain the total output power of controllable micro-sources through the total load demand and the total output power of uncontrollable micro-sources before calculating the controllable micro-sources. The specific output power of different types of micro-sources in micro-sources. After the total output power of the controllable micro-source is obtained, the total output power is used as a constraint condition, and the output power of the flexible load, energy storage battery and controllable distributed power is solved under the economic optimal condition through the economic optimal objective function expectations. Finally, with the expected output power values of the above three controllable micro-sources as input parameters, the rolling optimization processing steps and optimized performance index functions of the dynamic matrix control model are used to solve the output power prediction values of the above-mentioned three controllable micro-sources. At the same time, in order to ensure the accuracy of the predicted value of the output power, the actual measured value of the output power is used to correct the predicted value, so as to obtain a more accurate predicted value of the output power, and determine the scheduling scheme according to the predicted value.

Further, the expected output power values of the three controllable micro-sources are input into the dynamic matrix control prediction model as input parameters, so as to solve the output power prediction values of the three controllable micro-sources, and the dynamic matrix control prediction model is :

Among them, P ₀ (k) is the initial value of the controllable micro-source output power obtained through actual measurement, Δu ^T (k+t|k) is the controllable micro-source output power increment of the k+t period predicted at k time Matrix, the dimension is M, P(k+i|k) is the predictive value of the controllable micro-source output power at k+i time, i=1,2,...,N and M≤N, N is the number of initial predicted output values for which the control action remains unchanged at time k.

Further, the economic optimal objective function is:

minF(t)={F ₁ (t), F ₂ (t)}, t=1,2,...,24

F ₁ (t)＝C _grid (t)+C _flex (t)+C _st o _r (t)+C _DP (t)+C _OM (t)+C _DG (t)

F ₂ (t) = C _e (t)

Among them, F ₁ (t) is the operation scheduling cost of the multi-energy micro-grid at time t, F ₂ (t) is the environmental pollution cost of the multi-energy micro-grid at time t, and C _grid (t) is the cost of the multi-energy micro-grid at time t. The power purchase cost of the multi-energy microgrid to the external grid, C _flex (t) is the scheduling cost of the flexible load in the controllable micro-source at time t, and C _stor (t) is the energy storage in the controllable micro-source at time t The scheduling cost of the battery, C _DP (t) is the scheduling cost of the controllable distributed power source in the controllable micro-source at time t, C _OM (t) is the investment depreciation cost of the controllable micro-source at time t, C _DG (t) is the operation and maintenance cost of the controllable micro-source at time t, and C _e (t) is the environmental pollution cost of the controllable micro-source at time t.

The inventor screened out the specific calculation formulas of the above various costs through research:

C _grid (t) = cp (t) × P _grid (t)

In the formula, cp(t) is the electricity price of the external grid at time t, and P _grid (t) is the electricity purchased from the multi-energy microgrid to the external grid at time t;

m is the number of flexible loads, α _i and β _i are the cost coefficients of flexible load scheduling, P _flex0 is the initial value of output power before flexible load scheduling, and ΔP _flexi is the power variation of flexible loads;

n is the number of energy storage batteries, λ _stori is the scheduling cost coefficient of the i-th energy storage battery at time t, P _stori (t) is the charging and discharging power of the i-th energy storage battery at time t;

j is the number of distributed power sources, C _az,i is the unit capacity installation cost of the i-th distributed power source, k _i is the capacity factor of the i-th distributed power source, and Q is the Annual power generation (namely rated power), r is the annual interest rate, n _i is the investment repayment period of the i-th distributed power source, P _i (t) is the output power of the i-th distributed power source at time t;

is the unit power operation and maintenance cost coefficient of the i-th controllable micro-source;

k is the number of controllable micro-sources, a _i , b _i , and c _i are the scheduling cost coefficients of the i-th controllable micro-source, respectively, _PDGi (t) is the output power of the i-th controllable micro-source at time t;

h is the type of pollutant, V _ek and V _k are the environmental value and penalty of the k-th pollutant respectively, and Qi _ik is the unit electricity of the i-th micro-source.

Further, the optimized performance index function is:

minJ＝K _i A(k) ^T QA(k)+K _p ΔA(k) ^T QΔA(k)+K _d Δ ² A(k) ^T QΔ ² A(k)+ΔU(k) ^T RΔU(k )

Among them, A(k)=W(k)-P(k), P(k)=P ₀ (k)+ΔU(k);

In the formula, K _p is the proportional matrix, K _i is the integral matrix, K _d is the differential matrix, A(k) is the prediction error between the expected value and the predicted output power of the controllable micro-source, ΔA(k) is the controllable micro-source The prediction error increment between the expected value and the predicted value of the output power of the source, Δ ² A(k) is the increment of the prediction error increment between the expected value and the predicted value of the controllable micro-source output power, W(k) is the Controlled micro-source output power expectation matrix, P(k) is the controllable micro-source output power predicted value matrix, Q is the error weight matrix, R is the control weight matrix, ΔU(k) is the control increment matrix; P ₀ (k) is the actual measured value of the output power of the controllable micro-source at the initial moment.

Wherein, the present invention adopts the optimized performance index function improved by the inventor, and adopts a novel model predictive control algorithm combining multivariable dynamic matrix control (DMC) and fractional order PID (FPID) control in the state space equation form (FPID-DMC). The control algorithm is based on FPID and DMC control, reconstructs the DMC control objective function, and gives full play to the advantages of FPID controller, which is simple in structure, easy to implement, simple in DMC modeling, less in calculation, and strong in robustness. Effects are optimized.

Further, the correction formula used in step S4 is:

P _c o _r (k+1)＝P ₁ (k+1|k)+hE(k+1)

E(k+1)＝P _sj (k+1)-P ₁ (k+1|k)

In the formula, P ₁ (k+1|k) is the predicted value of the output power of the controllable micro-source at time k+1, P _cor (k+1) is the predicted value of the output power of the controllable micro-source after correction, and h is the error correction coefficient, E(k+1) is the output error, and P _sj (k+1) is the actual measured value of the controllable micro-source at time k+1.

Small hydropower stations, wind power, photovoltaics, etc. are greatly affected by environmental factors, and the predicted value based on DMC cannot guarantee that there is no deviation from the output of each micro-source. Therefore, in the real-time scheduling process based on DMC, a feedback correction link should be added to form a closed loop The control system performs rolling prediction again based on the current active output of the system to overcome the uncertainty of micro-source output such as small hydropower, photovoltaic, and wind power.

The present invention also provides a multi-energy micro-grid time-varying scheduling device based on dynamic matrix control, including:

The total output power prediction module of controllable micro-sources is used to predict the total load demand of multi-energy micro-grids and the total output power of uncontrollable micro-sources in multi-energy micro-grids by using neural network time series based on historical data and weather information, and calculate the multi-energy The total output power of the controllable micro-sources in the energy micro-grid;

Wherein, the total load demand of the multi-energy micro-grid includes the total output power of the uncontrollable micro-source and the total output power of the controllable micro-source;

The optimal power flow module is used to use the total output power of the controllable micro-source obtained by the controllable micro-source total output power prediction module as a constraint condition, and solve the optimal power flow through the economic optimal objective function to obtain the controllable micro-source Expected output power values of flexible loads, energy storage batteries and controllable distributed power sources in micro-sources;

The output power prediction module is used to use the output power expectations of the three controllable micro-sources obtained by the optimal power flow module as input parameters, and solve the three controllable micro-sources according to the prediction model of dynamic matrix control and the optimized performance index function respectively The output power prediction value of a controllable micro-source;

A correction module, configured to correct the predicted output power values of the three controllable micro-sources obtained by the output power prediction module and the actual measured output power values of the three controllable micro-sources, and based on the corrected The output power prediction value determines the scheduling scheme;

Wherein, the economic optimal objective function is used to solve the three kinds of controllable micro-sources when the operation scheduling cost of the multi-energy microgrid is the lowest and the impact of the pollutants on the environment during the operation of the multi-energy microgrid is guaranteed to be the lowest. output power expectations.

Further, the output power prediction module uses the expected output power values of the three controllable micro-sources as input parameters into the dynamic matrix control prediction model, so as to solve the output power prediction values of the three controllable micro-sources, and the dynamic matrix The control prediction model is:

Among them, P ₀ (k) is the initial value of the controllable micro-source output power obtained through actual measurement, Δu ^T (k+t| k) is the controllable micro-source output power increment of the k+t period predicted at k time Matrix, the dimension is M, P(k+i|k) is the predictive value of the controllable micro-source output power at k+i time, i=1,2,...,N and M≤N, N is the number of initial predicted output values for which the control action remains unchanged at time k.

Further, the economic optimal objective function adopted by the optimal power flow module is:

minF(t)={F ₁ (t), F ₂ (t)}, t=1,2,...,24

F ₁ (t)＝C _grid (t)+C _flex (t)+C _st o _r (t)+C _DP (t)+C _OM (t)+C _DG (t)

F ₂ (t) = C _e (t)

Among them, F ₁ (t) is the operation scheduling cost of the multi-energy micro-grid at time t, F ₂ (t) is the environmental pollution cost of the multi-energy micro-grid at time t, and C _grid (t) is the cost of the multi-energy micro-grid at time t. The power purchase cost of the multi-energy microgrid to the external grid, C _flex (t) is the scheduling cost of the flexible load in the controllable micro-source at time t, and C _st o _r (t) is the cost of flexible load in the controllable micro-source at time t. The scheduling cost of the energy storage battery, C _DP (t) is the scheduling cost of the controllable distributed power source in the controllable micro-source at time t, C _OM (t) is the investment depreciation cost of the controllable micro-source at time t, C _DG (t) is the operation and maintenance cost of the controllable micro-source at time t, and C _e (t) is the environmental pollution cost of the controllable micro-source at time t.

Further, the optimized performance index function adopted by the output power prediction module is:

minJ＝K _i A(k) ^T QA(k)+K _p ΔA(k) ^T QΔA(k)+K _d Δ ² A(k) ^T QΔ ² A(k)+ΔU(k) ^T RΔU(k )

Among them, A(k)=W(k)-P(k), P(k)=P ₀ (k)+ΔU(k);

In the formula, K _p is the proportional matrix, K _i is the integral matrix, K _d is the differential matrix, A(k) is the prediction error between the expected value and the predicted output power of the controllable micro-source, ΔA(k) is the controllable micro-source The prediction error increment between the expected value and the predicted value of the output power of the source, Δ ² A(k) is the increment of the prediction error increment between the expected value and the predicted value of the controllable micro-source output power, W(k) is the Controlled micro-source output power expectation matrix, P(k) is the controllable micro-source output power predicted value matrix, Q is the error weight matrix, R is the control weight matrix, ΔU(k) is the control increment matrix; P ₀ (k) is the actual measured value of the output power of the controllable micro-source at the initial moment.

Further, the correction formula adopted by the correction module is:

P _c o _r (k+1)＝P ₁ (k+1|k)+hE(k+1)

E(k+1)＝P _sj (k+1)-P ₁ (k+1|k)

In the formula, P ₁ (k+1|k) is the predicted value of the output power of the controllable micro-source at time k+1, P _cor (k+1) is the predicted value of the output power of the controllable micro-source after correction, and h is the error correction coefficient, E(k+1) is the output error, and P _sj (k+1) is the actual measured value of the controllable micro-source at time k+1.

Beneficial effect

A multi-energy micro-grid time-varying scheduling method and device based on dynamic matrix control provided by the present invention establishes a prediction model based on dynamic matrix control, through multivariable dynamic matrix control (DMC) and fractional-order PID in the form of state space equations The new model predictive control algorithm (FPID-DMC) combined with (FPID) control improves the control quality and rolls the forecast control increment, and then predicts the active output of each controllable distributed power supply, energy storage battery and flexible load in the limited time domain in the future. The method and device provided by the present invention have the characteristics of simple PID structure, convenient parameter adjustment, DMC rolling optimization, and strong robustness. The feedback correction link is introduced to correct the influence of random factors such as wind power and photovoltaic prediction errors on the scheduling results.

Description of drawings

FIG. 1 is a schematic flowchart of a multi-energy microgrid time-varying scheduling method based on dynamic matrix control provided by Embodiment 1 of the present invention;

Fig. 2 is a real-time electricity price of the external network in a specific example of a multi-energy micro-grid time-varying scheduling method based on dynamic matrix control provided by Embodiment 1 of the present invention;

Fig. 3 is the total load demand of the multi-energy micro-grid in the specific example of Embodiment 1 of the present invention;

Fig. 4 is the output power of the small hydropower station in the concrete example of embodiment one of the present invention;

Fig. 5 is the output power of photovoltaic power generation in the concrete example of embodiment one of the present invention;

Fig. 6 is the expected value of the output power of the flexible load, the storage battery and the micro gas turbine in the specific example of the first embodiment of the present invention;

FIG. 7 is the finalized scheduling scheme in the specific example of Embodiment 1 of the present invention;

Fig. 8 is a comparison diagram of the scheduling results obtained by the scheme provided by the present invention and the prior art method;

Fig. 9 is a schematic structural diagram of a multi-energy microgrid time-varying scheduling device based on dynamic matrix control provided by Embodiment 2 of the present invention.

Detailed ways

In order to facilitate a better understanding of the content of the solution of the present invention, the solution of the present invention will be further described below in conjunction with specific examples.

Embodiment one

Figure 1 is a schematic flow chart of a multi-energy microgrid time-varying scheduling method based on dynamic matrix control provided by Embodiment 1 of the present invention. The method provided by the present invention includes: Step S1: Using neural network time series according to historical data and weather information Predict the total load demand of the multi-energy micro-grid and the total output power of the uncontrollable micro-sources in the multi-energy micro-grid, and calculate the total output power of the controllable micro-sources in the multi-energy micro-grid; wherein, the total load of the multi-energy micro-grid The demand includes the total output power of the uncontrollable micro-source and the total output power of the controllable micro-source; step S2: taking the total output power of the controllable micro-source obtained in step S1 as a constraint condition, through the economic optimal objective function Solve the optimal power flow separately to obtain the expected output power values of flexible loads, energy storage batteries and controllable distributed power sources in the controllable micro-sources; step S3: use the expected output power values of the three controllable micro-sources obtained in step S2 as Input parameters, according to the prediction model of dynamic matrix control and the optimized performance index function, solve the output power prediction value of described three kinds of controllable micro-sources respectively; Step S4: the output of described three kinds of controllable micro-sources obtained in step S3 The predicted power value and the actual measured value of the output power of the three controllable micro-sources are corrected, and the scheduling scheme is determined according to the corrected output power predicted value; wherein, the economic optimal objective function is used to solve when the multi-energy micro- The expected value of the output power of the three controllable micro-sources when the operation scheduling cost of the multi-energy micro-grid is the lowest and the impact on the environment caused by the pollutants during the operation of the multi-energy micro-grid is the lowest.

In the present invention, the expected output power values of the three controllable micro-sources are input into the dynamic matrix control prediction model as input parameters, thereby solving the output power prediction values of the three controllable micro-sources, and the dynamic matrix control prediction The model is:

Among them, P ₀ (k) is the initial value of the controllable micro-source output power obtained through actual measurement, Δu ^T (k+t| k) is the controllable micro-source output power increment of the k+t period predicted at k time Matrix, the dimension is M, P(k+i|k) is the predictive value of the controllable micro-source output power at k+i time, i=1,2,...,N and M≤N, N is the number of initial predicted output values for which the control action remains unchanged at time k.

In the above steps, the economic optimal objective function adopted is specifically:

minF(t)={F ₁ (t), F ₂ (t)}, t=1,2,...,24

F ₁ (t)＝C _grid (t)+C _flex (t)+C _st o _r (t)+C _DP (t)+C _OM (t)+C _DG (t)

F ₂ (t) = C _e (t)

Among them, F ₁ (t) is the operation scheduling cost of the multi-energy micro-grid at time t, F ₂ (t) is the environmental pollution cost of the multi-energy micro-grid at time t, and C _grid (t) is the cost of the multi-energy micro-grid at time t. The power purchase cost of the multi-energy microgrid to the external grid, C _flex (t) is the scheduling cost of the flexible load in the controllable micro-source at time t, and C _st o _r (t) is the cost of flexible load in the controllable micro-source at time t. The scheduling cost of the energy storage battery, C _DP (t) is the scheduling cost of the controllable distributed power source in the controllable micro-source at time t, C _OM (t) is the investment depreciation cost of the controllable micro-source at time t, C _DG (t) is the operation and maintenance cost of the controllable micro-source at time t, and C _e (t) is the environmental pollution cost of the controllable micro-source at time t.

Through research, the inventor screened out the specific calculation formulas of the above various costs:

C _grid (t) = cp (t) × P _grid (t)

In the formula, cp(t) is the electricity price of the external grid at time t, and P _grid (t) is the electricity purchased from the multi-energy microgrid to the external grid at time t;

m is the number of flexible loads, α _i and β _i are the cost coefficients of flexible load scheduling, P _flex0 is the initial value of output power before flexible load scheduling, and ΔP _flexi is the power variation of flexible loads;

n is the number of energy storage batteries, λ _stori is the scheduling cost coefficient of the i-th energy storage battery at time t, P _stori (t) is the charging and discharging power of the i-th energy storage battery at time t;

j is the number of distributed power sources, C _az,i is the unit capacity installation cost of the i-th distributed power source, k _i is the capacity factor of the i-th distributed power source, and Q is the Annual power generation (namely rated power), r is the annual interest rate, n _i is the investment repayment period of the i-th distributed power source, P _i (t) is the output power of the i-th distributed power source at time t;

K _OM,i is the unit electricity operation and maintenance cost coefficient of the i-th controllable micro-source;

k is the number of controllable micro-sources, a _i , b _i , and c _i are the scheduling cost coefficients of the i-th controllable micro-source, respectively, _PDGi (t) is the output power of the i-th controllable micro-source at time t;

h is the type of pollutant, V _ek and V _k are the environmental value and penalty of the k-th pollutant respectively, and Qi _ik is the unit electricity of the i-th micro-source.

What the present invention adopted is the optimized performance index function improved by the inventor, and adopts the novel model predictive control algorithm (FPID) that combines multivariable dynamic matrix control (DMC) and fractional order PID (FPID) control under state space equation form -DMC). The control algorithm is based on FPID and DMC control, reconstructs the DMC control objective function, and gives full play to the advantages of FPID controller, which is simple in structure, easy to implement, simple in DMC modeling, less in calculation, and strong in robustness. Effects are optimized. The specific optimization performance index function is:

minJ＝K _i A(k) ^T QA(k)+K _p ΔA(k) ^T QΔA(k)+K _d Δ ² A(k) ^T QΔ ² A(k)+ΔU(k) ^T RΔU(k )

Among them, A(k)=W(k)-P(k), P(k)=P ₀ (k)+ΔU(k);

In the formula, K _p is the proportional matrix, K _i is the integral matrix, K _d is the differential matrix, A(k) is the prediction error between the expected value and the predicted output power of the controllable micro-source, ΔA(k) is the controllable micro-source The prediction error increment between the expected value and the predicted value of the output power of the source, Δ ² A(k) is the increment of the prediction error increment between the expected value and the predicted value of the controllable micro-source output power, W(k) is the Controlled micro-source output power expectation matrix, P(k) is the controllable micro-source output power predicted value matrix, Q is the error weight matrix, R is the control weight matrix, ΔU(k) is the control increment matrix; P ₀ (k) is the actual measured value of the output power of the controllable micro-source at the initial moment.

The correction formula adopted in step S4 of the present invention is:

P _c o _r (k+1)＝P ₁ (k+1|k)+hE(k+1)

E(k+1)＝P _sj (k+1)-P ₁ (k+1|k)

In the formula, P ₁ (k+1|k) is the predicted value of the output power of the controllable micro-source at time k+1, P _c o _r (k+1) is the predicted value of the output power of the controllable micro-source after correction, and h is Error correction coefficient, E(k+1) is the output error, P _sj (k+1) is the actual measurement value of the controllable micro-source at the moment k+1.

Small hydropower stations, wind power, photovoltaics, etc. are greatly affected by environmental factors, and the predicted value based on DMC cannot guarantee that there is no deviation from the output of each micro-source. Therefore, in the real-time scheduling process based on DMC, a feedback correction link should be added to form a closed loop The control system performs rolling prediction again based on the current active output of the system to overcome the uncertainty of micro-source output such as small hydropower, photovoltaic, and wind power.

Specifically, this embodiment aims at the rich practical characteristics of small hydropower in Hunan Province, and establishes a system consisting mainly of small hydropower, photovoltaics, a controllable distributed power source (in this embodiment, a micro gas turbine), and an energy storage battery (in this embodiment That is, a simple multi-energy micro-grid system composed of batteries) and flexible loads, and considering that the power demand in the micro-grid is relatively small and the power lines in the system are short, this embodiment assumes that there is no loss in line transmission, the system voltage is stable, and there is no harmonic interference etc. Figure 2 shows the real-time electricity price of the external network for one day; the fine standards, environmental value standards and emission data of various micro-sources and different pollutants. Li Yang, Xia Anbang. A distributed optimization strategy for power generation dispatching that takes into account both environmental protection and economic benefits [J]. Chinese Journal of Electrical Engineering, 2009, 29(16): 63-68.) and "Environmental Benefit Analysis of Distributed Power Generation" ( Qian Kejun, Yuan Yue, Shi Xiaodan, et al. Environmental Benefit Analysis of Distributed Power Generation[J]. Chinese Journal of Electrical Engineering, 2008,28(29):11-15.); Table 1-Table 5 shows Specific parameters of each device in the multi-energy microgrid in this embodiment.

Table 1 Parameters of photovoltaic power generation and hydropower units

Table 2 Battery parameters

Table 3 Micro gas turbine parameters

Table 4 Flexible load parameters

Note: P _load0 is the initial value of flexible load prediction at the current moment.

Table 5 The remaining parameters of each micro-source

Step S1: According to the historical data and weather information, the neural network time series is used to predict the total load demand of the multi-energy microgrid (as shown in Figure 3) and the total output power of the uncontrollable micro-sources in the multi-energy microgrid (including the The output power of the small hydropower station and the output power of photovoltaic power generation shown in Figure 5), and the total output power of the controllable micro-source in the multi-energy micro-grid is obtained;

Step S2: Taking the total output power of the controllable micro-source obtained in step S1 as a constraint condition, solve the optimal power flow through the economic optimal objective function respectively, and obtain the output power of the flexible load, battery and micro gas turbine in the controllable micro-source Expected value (as shown in Figure 6);

Step S3: Using the expected output power values of the flexible load, storage battery and micro gas turbine obtained in step S2 as input parameters, solve the output power predictions of the three controllable micro sources respectively according to the prediction model of dynamic matrix control and the optimized performance index function value;

Step S4: Correct the predicted output power values of flexible loads, storage batteries, and micro gas turbines obtained in step S3 and their actual measured output power values, and determine a dispatching plan based on the corrected predicted output power values (as shown in Figure 7).

By comparing the results of the traditional open-loop optimal scheduling method shown in Figure 8 (i.e. the conventional DMC scheduling method) with the scheduling method provided by the present invention (i.e. the FPID-DMC optimal scheduling method), it can be known that the active power trends of the two types of scheduling are the same, However, the active power deviation between traditional open-loop optimal scheduling and day-ahead scheduling is much larger than that of FPID-DMC optimal scheduling. At the same time, comparing the two scheduling result curves, it can be found that the optimal scheduling curve based on FPID-DMC is smoother, which can ensure a higher tolerance to the intermittent and fluctuating output of photovoltaic and small hydropower stations, and can also reduce the energy storage device. , flexible loads and other adjustable equipment mechanical loss, prolong service life.

By defining the stability index of schedulable equipment such as energy storage devices and flexible loads, quantitatively compare the dispatching method proposed in this paper with the traditional open-loop dispatching method during the dispatching process. flexible load as an example).

In the formula, P _flexi is the active output of the flexible load at time i, is the average value of the active output of the flexible load, and n is the number of moments corresponding to the scheduling result.

Compare the stability indexes of active power output of dispatchable equipment such as energy storage devices and flexible loads, as shown in Table 6.

Table 6 Comparison of the smoothness index of each schedulable equipment output

By quantitatively comparing the stability indicators of the active output of each schedulable equipment, it is proved that the FPID-DMC proposed in this paper is effective in optimizing the dispatching of stable distributed energy active output fluctuations, and has a higher ability to accommodate distributed energy with strong gaps. .

The operating cost and environmental cost of the real-time scheduling results of the two optimization methods are calculated by the calculation method proposed in this paper, and the calculation results are shown in Table 7.

Table 7 One-day operating costs under different scheduling strategies

Table 7 shows that the operation cost based on FPID-DMC optimization scheduling method is much lower than the traditional one-day optimization, and the environmental cost is similar, which proves that the method proposed in this paper has better economic benefits.

In summary, the present invention provides a multi-energy microgrid time-varying scheduling method based on dynamic matrix control, which establishes a predictive model based on dynamic matrix control, through multivariable dynamic matrix control (DMC) and New Model Predictive Control Algorithm Combining Fractional PID (FPID) Control (FPID-DMC) New Model Predictive Control Combining Multivariable Dynamic Matrix Control (DMC) and Fractional PID (FPID) Control in State Space Equation Form The algorithm (FPID-DMC) improves the control quality and rolls the forecast control increment, and then predicts the active output of each controllable distributed power supply, energy storage battery and flexible load in the future limited time domain. The method and device provided by the present invention have the characteristics of simple PID structure, convenient parameter adjustment, DMC rolling optimization, and strong robustness. The feedback correction link is introduced to correct the influence of random factors such as wind power and photovoltaic prediction errors on the scheduling results.

Embodiment two

Figure 2 shows a multi-energy microgrid time-varying scheduling device based on dynamic matrix control provided by Embodiment 2 of the present invention, including:

The controllable micro-source total output power prediction module 100 is used to predict the total load demand of the multi-energy micro-grid and the total output power of the uncontrollable micro-source in the multi-energy micro-grid by using the neural network time series based on historical data and weather information, and calculate The total output power of the controllable micro-sources in the multi-energy micro-grid;

Wherein, the total load demand of the multi-energy micro-grid includes the total output power of the uncontrollable micro-source and the total output power of the controllable micro-source;

The optimal power flow module 200 is used to use the total output power of the controllable micro-source obtained by the controllable micro-source total output power prediction module as a constraint condition, and solve the optimal power flow through the economic optimal objective function to obtain Output power expectations of flexible loads, energy storage batteries and controllable distributed power sources in micro-controlled sources;

The output power prediction module 300 is used to use the expected output power values of the three controllable micro-sources obtained by the optimal power flow module as input parameters, and respectively solve the described Predicted output power values of three controllable micro-sources;

The correction module 400 is used to correct the predicted output power values of the three controllable micro-sources obtained by the output power prediction module and the actual measured output power values of the three controllable micro-sources, and according to the corrected The predicted value of the output power determines the scheduling scheme;

Wherein, the economic optimal objective function is used to solve the three kinds of controllable micro-sources when the operation scheduling cost of the multi-energy microgrid is the lowest and the impact of the pollutants on the environment during the operation of the multi-energy microgrid is guaranteed to be the lowest. output power expectations.

Wherein, the output power prediction module inputs the expected output power values of the three controllable micro-sources into the dynamic matrix control prediction model as input parameters, thereby solving the output power prediction values of the three controllable micro-sources, and the dynamic matrix control The predictive model is:

Among them, P ₀ (k) is the initial value of the controllable micro-source output power obtained through actual measurement, Δu ^T (k+t|k) is the controllable micro-source output power increment of the k+t period predicted at k time Matrix, the dimension is M, P(k+i|k) is the predictive value of the output power of the controllable micro-source at k+i time predicted at time k, i=1,2,...,N and M≤N.

Among them, the economic optimal objective function adopted by the optimal power flow module is:

minF(t)={F ₁ (t), F ₂ (t)}, t=1,2,...,24

F ₁ (t)＝C _grid (t)+C _flex (t)+C _st o _r (t)+C _DP (t)+C _OM (t)+C _DG (t)

F ₂ (t) = C _e (t)

Among them, F ₁ (t) is the operation scheduling cost of the multi-energy micro-grid at time t, F ₂ (t) is the environmental pollution cost of the multi-energy micro-grid at time t, and C _grid (t) is the cost of the multi-energy micro-grid at time t. The power purchase cost of the multi-energy microgrid to the external grid, C _flex (t) is the scheduling cost of the flexible load in the controllable micro-source at time t, and C _stor (t) is the energy storage in the controllable micro-source at time t The scheduling cost of the battery, C _DP (t) is the scheduling cost of the controllable distributed power source in the controllable micro-source at time t, C _OM (t) is the investment depreciation cost of the controllable micro-source at time t, C _DG (t) is the operation and maintenance cost of the controllable micro-source at time t, and C _e (t) is the environmental pollution cost of the controllable micro-source at time t.

Among them, the optimized performance index function adopted by the output power prediction module is:

minJ＝K _i A(k) ^T QA(k)+K _p ΔA(k) ^T QΔA(k)+K _d Δ ² A(k) ^T QΔ ² A(k)+ΔU(k) ^T RΔU(k )

Among them, A(k)=W(k)-P(k), P(k)=P ₀ (k)+ΔU(k);

In the formula, K _p is the proportional matrix, K _i is the integral matrix, K _d is the differential matrix, A(k) is the prediction error between the expected value and the predicted output power of the controllable micro-source, ΔA(k) is the controllable micro-source The prediction error increment between the expected value and the predicted value of the output power of the source, Δ ² A(k) is the increment of the prediction error increment between the expected value and the predicted value of the controllable micro-source output power, W(k) is the Controlled micro-source output power expectation matrix, P(k) is the controllable micro-source output power predicted value matrix, Q is the error weight matrix, R is the control weight matrix, ΔU(k) is the control increment matrix; P ₀ (k) is the actual measured value of the output power of the controllable micro-source at the initial moment.

Among them, the correction formula adopted by the correction module is:

P _c o _r (k+1)＝P ₁ (k+1|k)+hE(k+1)

E(k+1)＝P _sj (k+1)-P ₁ (k+1|k)

In the formula, P ₁ (k+1|k) is the predicted value of the output power of the controllable micro-source at time k+1, P _c o _r (k+1) is the predicted value of the output power of the controllable micro-source after correction, and h is Error correction coefficient, E(k+1) is the output error, P _sj (k+1) is the actual measurement value of the controllable micro-source at the moment k+1.

For the specific working principle and description of each module in the above method embodiment, reference may be made to the description of the corresponding part of the specific processing flow of each step in the above method embodiment, which will not be repeated here.

In summary, the present invention provides a multi-energy microgrid time-varying scheduling device based on dynamic matrix control, which establishes a predictive model based on dynamic matrix control, through multivariable dynamic matrix control (DMC) and New Model Predictive Control Algorithm Combining Fractional PID (FPID) Control (FPID-DMC) New Model Predictive Control Combining Multivariable Dynamic Matrix Control (DMC) and Fractional PID (FPID) Control in State Space Equation Form The algorithm (FPID-DMC) improves the control quality and rolls the forecast control increment, and then predicts the active output of each controllable distributed power supply, energy storage battery and flexible load in the future limited time domain. The method and device provided by the present invention have the characteristics of simple PID structure, convenient parameter adjustment, DMC rolling optimization, and strong robustness. The feedback correction link is introduced to correct the influence of random factors such as wind power and photovoltaic prediction errors on the scheduling results.

The above description is only an embodiment of the present invention, and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention .

Claims (8)

1. A multi-energy micro-grid controllable micro-source time-varying scheduling method based on dynamic matrix control, characterized in that it comprises: Step S1: According to the historical data and weather information, the neural network time series is used to predict the total load demand of the multi-energy microgrid and the total output power of the uncontrollable micro-sources in the multi-energy micro-grid, and calculate the controllable micro-sources in the multi-energy microgrid. Total output power; Wherein, the total load demand of the multi-energy micro-grid includes the total output power of the uncontrollable micro-source and the total output power of the controllable micro-source; Step S2: Taking the total output power of the controllable micro-source obtained in step S1 as the constraint condition, the optimal power flow is solved respectively through the economic optimal objective function, and the flexible load, energy storage battery and controllable distribution in the controllable micro-source are obtained The expected value of the output power of the power supply; Step S3: taking the expected output power values of the three controllable micro-sources obtained in step S2 as input parameters, and solving the output powers of the three controllable micro-sources respectively according to the prediction model of dynamic matrix control and the optimized performance index function Predictive value; Step S4: Correct the predicted output power values of the three controllable micro-sources obtained in step S3 and the actual measured output power values of the three controllable micro-sources, and determine the schedule according to the corrected output power predicted values Program; Wherein, the economic optimal objective function is used to solve the three kinds of controllable micro-sources when the operation scheduling cost of the multi-energy microgrid is the lowest and the impact of the pollutants on the environment during the operation of the multi-energy microgrid is guaranteed to be the lowest. output power expectation; The optimized performance index function is: minJ＝K _i A(k) ^T QA(k)+K _p ΔA(k) ^T QΔA(k)+K _d Δ ² A(k) ^T QΔ ² A(k)+ΔU(k) ^T RΔU(k ) Among them, A(k)=W(k)-P(k), P(k)=P ₀ (k)+ΔU(k); In the formula, K _p is the proportional matrix, K _i is the integral matrix, K _d is the differential matrix, A(k) is the prediction error between the expected value and the predicted output power of the controllable micro-source, ΔA(k) is the controllable micro-source The prediction error increment between the expected value and the predicted value of the output power of the source, Δ ² A(k) is the increment of the prediction error increment between the expected value and the predicted value of the controllable micro-source output power, W(k) is the Controlled micro-source output power expectation matrix, P(k) is the controllable micro-source output power predicted value matrix, Q is the error weight matrix, R is the control weight matrix, ΔU(k) is the control increment matrix; P ₀ (k) is the actual measured value of the output power of the controllable micro-source at the initial moment; T represents the matrix transposition. 2. method according to claim 1, it is characterized in that, with the output power expectation value of described three kinds of controllable micro-sources as input parameter input dynamic matrix control prediction model, thereby solve the output of described three kinds of controllable micro-sources Power prediction value, the dynamic matrix control prediction model is: Among them, P ₀ (k) is the initial value of the controllable micro-source output power obtained through actual measurement, Δu ^T (k+t|k) is the controllable micro-source output power increment of the k+t period predicted at k time Matrix, the dimension is M, P(k+i|k) is the predictive value of the controllable micro-source output power at k+i time, i=1,2,...,N and M≤N, N is the number of initial predicted output values for which the control action remains unchanged at time k. 3. method according to claim 2, is characterized in that, described economic optimal objective function is: minF(t)={F ₁ (t), F ₂ (t)}, t=1,2,...,24 F ₁ (t)＝C _grid (t)+C _flex (t)+C _stor (t)+C _DP (t)+C _OM (t)+C _DG (t) F ₂ (t) = C _e (t) Among them, F ₁ (t) is the operation scheduling cost of the multi-energy micro-grid at time t, F ₂ (t) is the environmental pollution cost of the multi-energy micro-grid at time t, and C _grid (t) is the cost of the multi-energy micro-grid at time t. The power purchase cost of the multi-energy microgrid to the external grid, C _flex (t) is the scheduling cost of the flexible load in the controllable micro-source at time t, and C _stor (t) is the energy storage in the controllable micro-source at time t The scheduling cost of the battery, C _DP (t) is the scheduling cost of the controllable distributed power in the controllable micro-source at time t, C _OM (t) is the investment depreciation cost of the controllable micro-source at time t, C _DG (t) is the operation and maintenance cost of the controllable micro-source at time t, and C _e (t) is the environmental pollution cost of the controllable micro-source at time t. 4. The method according to claim 3, characterized in that the correction formula adopted in the step S4 is: P _cor (k+1)＝P ₁ (k+1|k)+hE(k+1) E(k+1)＝P _sj (k+1)-P ₁ (k+1|k) In the formula, P ₁ (k+1|k) is the predicted value of the output power of the controllable micro-source at time k+1, P _cor (k+1) is the predicted value of the output power of the controllable micro-source after correction, and h is the error correction coefficient, E(k+1) is the output error, and P _sj (k+1) is the actual measured value of the output power of the controllable micro-source at time k+1. 5. A multi-energy micro-grid controllable micro-source time-varying scheduling device based on dynamic matrix control, characterized in that it includes: The total output power prediction module of controllable micro-sources is used to predict the total load demand of multi-energy micro-grids and the total output power of uncontrollable micro-sources in multi-energy micro-grids by using neural network time series based on historical data and weather information, and calculate the multi-energy The total output power of the controllable micro-sources in the energy micro-grid; The optimal power flow module is used to use the total output power of the controllable micro-source obtained by the controllable micro-source total output power prediction module as a constraint condition, and solve the optimal power flow through the economic optimal objective function to obtain the controllable micro-source Expected output power values of flexible loads, energy storage batteries and controllable distributed power sources in micro-sources; The output power prediction module is used to use the output power expectations of the three controllable micro-sources obtained by the optimal power flow module as input parameters, and solve the three controllable micro-sources according to the prediction model of dynamic matrix control and the optimized performance index function respectively The output power prediction value of a controllable micro-source; A correction module, configured to correct the predicted output power values of the three controllable micro-sources obtained by the output power prediction module and the actual measured output power values of the three controllable micro-sources, and based on the corrected The output power prediction value determines the scheduling scheme; Wherein, the total load demand of the multi-energy micro-grid includes the total output power of the uncontrollable micro-source and the total output power of the controllable micro-source; The economic optimal objective function is used to solve the output power of the three controllable micro-sources when the operation scheduling cost of the multi-energy micro-grid is the lowest and the impact of pollutants on the environment during the operation of the multi-energy micro-grid is guaranteed to be the lowest. expected value; The optimized performance index function adopted by the output power prediction module is: minJ＝K _i A(k) ^T QA(k)+K _p ΔA(k) ^T QΔA(k)+K _d Δ ² A(k) ^T QΔ ² A(k)+ΔU(k) ^T RΔU(k ) Among them, A(k)=W(k)-P(k), P(k)=P ₀ (k)+ΔU(k); In the formula, K _p is the proportional matrix, K _i is the integral matrix, K _d is the differential matrix, A(k) is the prediction error between the expected value and the predicted output power of the controllable micro-source, ΔA(k) is the controllable micro-source The prediction error increment between the expected value and the predicted value of the output power of the source, Δ ² A(k) is the increment of the prediction error increment between the expected value and the predicted value of the output power of the controllable micro-source, W(k) is the controllable Micro-source output power expected value matrix, P(k) is the controllable micro-source output power predicted value matrix, Q is the error weight matrix, R is the control weight matrix, ΔU(k) is the control increment matrix; P ₀ (k) is The actual measured value of the output power of the controllable micro-source at the initial moment; T represents the matrix transposition. 6. The device according to claim 5, wherein the output power prediction module uses the expected output power values of the three controllable micro-sources as input parameters to input the dynamic matrix control prediction model, thereby solving the three controllable micro-sources The output power prediction value of the micro-source, the dynamic matrix control prediction model is: Among them, P ₀ (k) is the initial value of the controllable micro-source output power obtained through actual measurement, Δu ^T (k+t|k) is the controllable micro-source output power increment of the k+t period predicted at k time Matrix, the dimension is M, P(k+i|k) is the predictive value of the controllable micro-source output power at k+i time, i=1,2,...,N and M≤N, N is the number of initial predicted output values for which the control action remains unchanged at time k. 7. The device according to claim 6, characterized in that the economic optimal objective function adopted by the optimal power flow module is: minF(t)={F ₁ (t), F ₂ (t)}, t=1,2,...,24 F ₁ (t)＝C _grid (t)+C _flex (t)+C _stor (t)+C _DP (t)+C _OM (t)+C _DG (t) F ₂ (t) = C _e (t) Among them, F ₁ (t) is the operation scheduling cost of the multi-energy micro-grid at time t, F ₂ (t) is the environmental pollution cost of the multi-energy micro-grid at time t, and C _grid (t) is the The power purchase cost of the multi-energy microgrid to the external network, C _flex (t) is the scheduling cost of the flexible load in the controllable micro-source at time t, C _st o _r (t) is the cost of the controllable micro-source at time t The scheduling cost of the energy storage battery, C _DP (t) is the scheduling cost of the controllable distributed power source in the controllable micro-source at time t, C _OM (t) is the investment depreciation cost of the controllable micro-source at time t, C _DG (t) is the operation and maintenance cost of the controllable micro-source at time t, and C _e (t) is the environmental pollution cost of the controllable micro-source at time t. 8. The device according to claim 7, wherein the correction formula adopted by the correction module is: P _cor (k+1)＝P ₁ (k+1|k)+hE(k+1) E(k+1)＝P _sj (k+1)-P ₁ (k+1|k) In the formula, P ₁ (k+1|k) is the predicted value of the controllable micro-source output power at time k+1, P _cor (k+1) is the corrected controllable micro-source output power predicted value, h is the error correction coefficient, E(k+1) is the output error, and P _sj (k+1) is the actual measured value of the output power of the controllable micro-source at time k+1.