Autonomous region division method based on random model predictive control under opportunity constraint

Abstract

The invention discloses an autonomous region division method based on random model prediction control under opportunity constraint. The method comprises the following steps: predicting wind power photovoltaic output by using a differential autoregressive moving average model; considering the uncertainty of wind and light output, and establishing an autonomous region division opportunity constraint system by taking the minimum load loss risk rate of the autonomous region as a target; generating a distributed output scene set in a specified reasonable output interval by adopting a scene analysis method in combination with photovoltaic and fan prediction error probability models, and solving opportunity constraints by applying a stochastic simulation technology; a novel power distribution system autonomous region division optimization model based on random model prediction control is constructed, and continuous forward rolling optimization is carried out through rolling optimization and feedback correction. The method is used for solving the problem that in the prior art, the control dimension is too high due to the fact that distributed energy is connected into a distribution network, it can be guaranteed that the load loss risk probability is kept at a low level under a certain confidence level, and consumption of distributed new energy and the stability of an autonomous region are improved while the power supply reliability is guaranteed.

Description

Technical Field

[0001] The invention relates to the field of distribution network optimization division, and in particular to an autonomous region division method based on random model predictive control under chance constraints. Background Art

[0002] Under the proposal of the "dual carbon" goal, distributed energy generation has gradually transformed from the previous free access power generation mode to a mode of centralized and decentralized photovoltaic access. In some areas, the proportion of distributed energy access has been too high. Due to the uncertainty of distributed output and the shortcomings of the previous centralized voltage control method, the problems caused by large-scale distributed energy access to the distribution network are increasing. The high-dimensional optimization problem of the distribution network with distributed energy access is transformed into an optimization problem of several sub-autonomous regions. Each autonomous region is regulated as a whole, which effectively improves the ability of the distribution network to manage distributed energy, and is conducive to the hierarchical and partitioned control of the distribution network. Therefore, it is necessary to study the dynamic division method of autonomous regions based on stochastic model predictive control under chance constraints under the consideration of system uncertainty, so as to effectively reduce the control difficulty and dimension of the distribution network, and make the system more stable in operation, regulation and management.

[0003] Experts and scholars in the field of power systems have also conducted extensive research on the division of autonomous regions in distribution networks, but it is still in its infancy and no universal division system or method has been established. In existing research on the division of autonomous regions of power distribution, indicators such as improved modularity function and autonomous regional power balance are often used as the basis for distribution network division, and intelligent algorithms are used to solve them. However, the impact of the intermittent and volatile output of distributed resources is not considered enough, and the static division method easily makes the division results difficult to adapt to changes in system operating conditions. Existing divisions are mostly based on intelligent algorithms, and the division results are prone to fall into local optimality, and the impact of the intermittent and volatile output of distributed resources is not considered enough. Summary of the invention

[0004] In response to the defects of the prior art, the present application discloses an autonomous region division method based on random model predictive control under chance constraints. The method is used to solve the problem of excessively high control dimension caused by the access of distributed energy to the distribution network in the prior art. The proposed division method can ensure that the probability of load loss risk is maintained at a low level under a certain confidence level, effectively and reasonably match distributed new energy with autonomous regions, ensure the power balance of autonomous regions and the power supply reliability capabilities of autonomous regions, improve the absorption of distributed new energy, and improve the stability of autonomous regions.

[0005] The technical solution to achieve the purpose of the present invention is:

[0006] The autonomous region partitioning method based on stochastic model predictive control under chance constraints specifically includes the following steps:

[0007] Step 1: After obtaining all possible autonomous region division schemes by the branch and cut enumeration method, the wind and photovoltaic output is predicted using the Autoregressive Integrated Moving Average Model (ARIMA) to generate the output power of photovoltaics, wind turbines and loads in the prediction time domain;

[0008] Step 2: Considering the uncertainty of wind and solar power output, taking the minimum risk rate of load loss in the autonomous region as the objective function, a partition constraint covering the opportunity constraint of power supply rate in the autonomous region, the opportunity constraint of power balance in the autonomous region, and the partition constraint of system safe operation is constructed;

[0009] Step 3: Combined with the predicted output of photovoltaic and wind turbines and the prediction error probability model, the scenario analysis method is used to generate a distributed output scenario set that represents uncertainty within the specified reasonable output range, and the average output of distributed energy is obtained. The random simulation technology is used to solve the opportunity constraint considering the uncertainty of wind and solar output;

[0010] Step 4: Construct a new distribution system autonomous area division optimization model based on stochastic model predictive control (SMPC). In the rolling optimization part, a breadth-first search algorithm is used to collect information on all division schemes, calculate the objective function of all schemes that meet the constraints under a certain confidence level, and determine the autonomous area division scheme within the first rolling optimization time period in the future; during feedback correction, the predicted values ​​of wind turbines, photovoltaics and loads are updated according to the latest real-time measurement values;

[0011] Step 5: Repeat steps 2 to 4 to complete the division of autonomous regions at subsequent times.

[0012] Furthermore, in step 1, the prediction module is used to generate the output power of photovoltaic, wind turbine and load in the prediction time domain, as follows:

[0013] The prediction model uses the ARIMA model. The core idea is to stabilize the series by the difference method, and use the regression moving average model to predict the stabilized time series. The ARIMA model is expressed as ARIMA (p, d, q). The model consists of two models: the autoregressive model (AR) and the moving average model (MA). p is the order of the autoregressive model, q is the order of the moving average, and d is the degree of difference. ARIMA stabilizes the data through the difference operation. Generally speaking, the data can be stabilized after one difference, so the value of d is generally 0, 1, or 2.

[0014] The mathematical formula of ARIMA(p, d, q) can be described as follows:

[0015]

[0016] Where: Y(t) is the value of the time series, t is the time point, c is the constant term, is the autoregressive coefficient, ε t are the interference terms θ1,θ2,…,θ q is the moving average coefficient, p is the number of autoregressive terms in the AR model, and q is the number of moving average terms in the MA model;

[0017] The process of building an ARIMA model is divided into the following three steps:

[0018] Step 1.1: Stationarity test. Check whether the time series meets the stationarity requirements. For non-stationary time series, first perform d-order differences to convert them into stationary time series;

[0019] Step 1.2: Calculate the autocorrelation coefficient (ACF) and partial autocorrelation coefficient (PACF) of the stationary time series, and perform parameter estimation through Akaikes information criterion (AIC) or Bayesian information criterion (BIC) search to obtain the optimal order p and q;

[0020] Step 1.3: Obtain the ARIMA model through the above d, p, q. The obtained model is tested, that is, the residual value ε t Perform analytical checks to determine the suitability of the time series ARIMA model and determine E(ε t ) is zero, which means the obtained model passes the test.

[0021] Furthermore, the objective function and constraints in step 2 are as follows:

[0022] (1) Objective function

[0023] Distributed renewable energy has flexible operation modes and can be connected to the grid or off-grid. When a major accident occurs in the power grid and a large-scale power outage occurs, the autonomous region can rely on the distributed renewable energy in the autonomous region to provide power support for the island operation of the autonomous region, quickly restore the power supply of important loads in the autonomous region, reduce the economic losses caused by large-scale power outages in the autonomous region, and enable the autonomous region to have independent autonomous operation capabilities. This paper takes the minimum risk rate of load loss in the autonomous region as the goal, and establishes the objective function based on opportunity constraints as follows:

[0024]

[0025] In the formula, F psr is the target value of the load loss risk rate in the autonomous region, F psrThe smaller the value, the better the power supply reliability of the autonomous region obtained by division. str.j (t) is the power supply rate of autonomous region j at time t; P DG.i (t) is the output power of the i-th distributed generation in the autonomous region, P str.i (t) is the charging and discharging power of the i-th energy storage device in the autonomous region, P l.i (t) is the load power of node i in the autonomous region, N clu.j.DG 、N clu.j.str 、N clu.j.l 、N clu.j.g are the number of distributed generation, energy storage devices, load nodes, and controllable traditional units in autonomous region j, and T is the optimization period;

[0026] (2) Constraints:

[0027] 1) The probability constraint of power supply rate in autonomous region, that is,

[0028]

[0029] In the formula, F min is the minimum value of the load power supply rate; x is the decision variable, that is, the division scheme of each autonomous area; ξ is the distributed energy output scenario; n is the number of distributed energy output scenarios; is the probability of the event being true; P ξ Contribute to a scenario in the distributed resource contribution scenario set; F(x,P ξ ) is the output P ξ is the load power supply rate of the system; β is the confidence level value satisfied by the objective function;

[0030] 2) Autonomous regional power balance indicators and constraints

[0031] The dynamic power balance of the autonomous region of the partitioning scheme is constrained to fully ensure the stability of power in the autonomous region. As the load and DG power change in real time, it brings certain risks to the safe and stable operation in the autonomous region. Based on this, the power balance index and constraint R of the autonomous region are proposed. p The definition is as follows:

[0032]

[0033] R p ≥β p (6)

[0034] Where: is the total active power demand of loads in the autonomous area; Make positive contributions to distributed energy resources; is the maximum active output of distributed energy storage in the autonomous region; α is the load power supply rate that needs to be met in the autonomous region; β p is the confidence level that the power balance constraint needs to meet;

[0035] Autonomous region power balance constraint R p It reflects the probability that the total output of DG in the autonomous region can meet the load demand to a certain extent when the energy storage is at its maximum output. The higher the index is, the higher the safety of the autonomous region is and the more reasonable the division scheme is.

[0036] 3) System safety operation constraints

[0037] U i,min ≤u i (t)≤U i，max (7)

[0038] |I ij (t)|≤I ij.max (8)

[0039] Where: u i (t) is the voltage amplitude of node i at time t; U i,max and U i,min are the upper and lower limits allowed for node i; I ij (t) is the current amplitude of line ij at time t; I ij.max is the maximum current allowed to flow through branch ij;

[0040] Furthermore, the step 3 is specifically as follows: For the prediction errors of photovoltaic, wind turbine output and load, it is generally believed that they satisfy the normal distribution, and the prediction error probability density function is:

[0041]

[0042]

[0043] Where ΔP is the prediction error of distributed energy output; μ and σ are the expectation and standard deviation of the prediction error, respectively, and μ is 0; e i is the value of each data point; n is the number of data points;

[0044] Latin hypercube sampling can avoid the problem of excessive clustering of sampling samples. The specific sampling steps are as follows:

[0045] Step 3.1: Divide the interval from 0 to 1 into l equal parts, and randomly generate a number P in each part according to uniform distribution. i ;

[0046] Step 3.2: Shuffle the order of these l random numbers in any way, and these random numbers are the probability values ​​of each random sample;

[0047] Step 3.3: From the inverse function of the probability distribution function f -1 (P i )=x i Calculate the sample value corresponding to the probability value l of each sample;

[0048] There are generally two methods for dealing with chance constraints: analytical method and random simulation method. This paper uses random simulation method to verify whether the division of autonomous regions meets the chance constraints. The steps to check whether the division scheme meets the chance constraint confidence level at time t are as follows:

[0049] 1) According to the probability distribution characteristics of the load and uncontrollable DG power output in the autonomous area, a power curve is randomly extracted based on the scenario generation technology in step 2;

[0050] 2) Count the total number of times the autonomous region meets power balance in M ​​samplings. p and the total number of times M that the probability constraint of the objective function of the load loss risk rate of the autonomous region is satisfied v ;

[0051] 3) If It is considered that the autonomous region division scheme at this moment can meet the opportunity constraint condition. Otherwise, the next autonomous region division scheme is solved until the optimal autonomous region division scheme under the confidence level that meets the objective function and the opportunity constraint is found.

[0052] Due to the randomness of wind power and photovoltaic output, there is a widespread deviation in wind power and photovoltaic prediction. At present, most prediction methods have a wind speed prediction deviation of 25% to 40%. Although the ARIMA prediction method proposed in this patent can effectively guarantee the prediction accuracy of distributed output, there is still a prediction deviation. Therefore, if the autonomous area division is performed completely according to the predicted output, there will be a large deviation, and the power quality and stability of the system after the division will also be affected; but if the prediction of distributed output is completely ignored, the distributed output prediction error is randomly generated according to the above Weibull distribution function, and its output is calculated, it has a large blindness and cannot guarantee the rationality of the division result; therefore, if the randomness and the regularity of the prediction are combined, it can avoid blindness and ensure the rationality of the division result. This paper introduces the interval random simulation method to combine the two.

[0053] For the power prediction curve of wind power and photovoltaic power, taking into account the prediction error and volatility of wind power and photovoltaic power, the output deviation is allowed to be ±20%. It is assumed that the output level of wind power and photovoltaic power at each moment fluctuates between the upper and lower deviations, which is the random number generation range of the wind power and photovoltaic power output at each moment.

[0054] Compared with the prior art, the present invention adopts the above technical solution and has the following beneficial effects:

[0055] 1. The autonomous region division method based on random model predictive control under chance constraints proposed in the present invention is used to solve the problem of excessively high control dimension caused by the access of distributed energy to the distribution network in the prior art. It can ensure that the probability of load loss risk is maintained at a low level under a certain confidence level, effectively and reasonably match distributed new energy with autonomous regions, and ensure the power balance of autonomous regions and the power supply reliability capabilities of autonomous regions.

[0056] 2. The present invention proposes an autonomous region partitioning method based on random model predictive control under chance constraints. The chance constraint solving process takes all sampling scenarios into account and performs partitioning and solving at a certain confidence level, thereby effectively balancing the impact of the randomness and volatility of distributed energy output on the partitioning results.

[0057] 3. The autonomous region division method based on random model predictive control under chance constraints proposed in the present invention not only improves the consumption of distributed new energy and the stability of autonomous regions, but also provides data support and theoretical basis for the independent operation, partition coordination and island control of microgrids. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart of an autonomous region partitioning method based on stochastic model predictive control under chance constraints in a specific implementation mode of the present invention;

[0059] Figure 2 It is a modified IEEE33 node structure diagram of a specific implementation mode of the present invention;

[0060] Figure 3 is the photovoltaic prediction value under the ARIMA prediction model of the specific implementation mode of the present invention;

[0061] Figure 4 is the photovoltaic predicted output range of a specific implementation mode of the present invention;

[0062] Figure 5 It is a zoning diagram of the autonomous area of ​​cutting line No. 3 of a specific embodiment of the present invention;

[0063] Figure 6 It is a zoning diagram of the autonomous area of ​​the No. 6 cutting line of the specific embodiment of the present invention;

[0064] Figure 7 It is a zoning diagram of the autonomous areas of cutting lines No. 4 and No. 8 according to a specific implementation mode of the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] In order to verify the effectiveness of the proposed distribution network autonomous area division method, the improved IEEE33-node power grid is used as an example for case analysis. The total base capacity of the system is 1MVA, the voltage level is 12.66kV, the voltage range is 0.95~1.05pu., the total active load is 3715kW, and the total reactive load is 2300kvar. There are 2 wind turbines, 2 photovoltaics, and 3 energy storages in the system. Their installation locations are as follows: Figure 2 As shown, Figure 2 The black nodes are nodes with DG. The DG installation nodes and installation capacity are shown in Table 1. The minimum power supply rate allowed is set to 0.35. Simulation verification is carried out on the Matalb 2016b platform, and the autonomous region partition model is constructed through the Yalmip optimization toolbox, and the optimization solution software Cplex is called for solution.

[0067] Table 1DG installation capacity

[0068]

[0069] Among them, nodes 13, 24 and 30 are installed with energy storage devices, and their active and reactive outputs are controllable; nodes 9 and 14 are installed with photovoltaics, and nodes 25 and 27 are installed with wind turbines. The active power of photovoltaics and wind turbines is not controllable, but the reactive power is controllable.

[0070] Step 1: Use the prediction module to generate the output power of photovoltaic, wind turbine and load in the prediction time domain.

[0071] The prediction model uses the ARIMA model. The core idea is to stabilize the series by the difference method, and use the regression moving average model to predict the stabilized time series. The ARIMA model is expressed as ARIMA (p, d, q). The model consists of two models: the autoregressive model AR (p) and the moving average MA (q). p is the order of the autoregressive model, q is the order of the moving average, and d is the degree of difference. ARIMA stabilizes the data through the difference operation. Generally speaking, the data can be stabilized after one difference, so the value of d is generally 0, 1, or 2.

[0072] The mathematical formula of ARIMA(p, d, q) can be described as follows:

[0073]

[0074] Where: Y(t) is the value of the time series, t is the time point, c is the constant term, is the autoregressive coefficient, ε t are the interference terms θ1,θ2,…,θ q is the moving average coefficient, p is the number of autoregressive terms in the AR model, and q is the number of moving average terms in the MA model.

[0075] The process of building an ARIMA model is divided into the following three steps:

[0076] Step 1.1: Stationarity test. Check whether the time series meets the stationarity requirements. For non-stationary time series, first perform d-order differences to convert them into stationary time series;

[0077] Step 1.2: Calculate the autocorrelation coefficient (ACF) and partial autocorrelation coefficient (PACF) of the stationary time series, and perform parameter estimation through Akaikes information criterion (AIC) or Bayesian information criterion (BIC) search to obtain the optimal order p and q;

[0078] Step 1.3: Obtain the ARIMA model through the above d, p, q. The obtained model is tested, that is, the residual value ε t Perform analytical checks to determine the suitability of the time series ARIMA model and determine E(ε t ) is zero, which means the obtained model passes the test.

[0079] After obtaining all possible autonomous region division schemes by the branch-cut enumeration method, the differential autoregressive moving average model is used to predict the wind power and photovoltaic output, and the output power of photovoltaic, wind turbine and load in the prediction time domain is generated. Taking 720 sets of photovoltaic samples on a certain day as an example, 432 sets of training set data are intercepted, and the ARIMA prediction module is used to generate the photovoltaic output power of 288 sets of samples in the prediction time domain of node 9. Figure 3 shown.

[0080] Step 2: Considering the uncertainty of wind and solar power output, taking the minimum risk rate of load loss in the autonomous region as the objective function, a partition constraint covering the opportunity constraint of power supply rate in the autonomous region, the opportunity constraint of power balance in the autonomous region, and the partition constraint of system safe operation is constructed;

[0081] (1) Objective function

[0082] Distributed renewable energy has flexible operation modes and can be connected to the grid or off-grid. When a major accident occurs in the power grid and a large-scale power outage occurs, the autonomous region can rely on the distributed renewable energy in the autonomous region to provide power support for the island operation of the autonomous region, quickly restore the power supply of important loads in the autonomous region, reduce the economic losses caused by large-scale power outages in the autonomous region, and enable the autonomous region to have independent autonomous operation capabilities. This paper takes the minimum risk rate of load loss in the autonomous region as the goal, and establishes the objective function based on opportunity constraints as follows:

[0083]

[0084] In the formula, F psr is the target value of the load loss risk rate in the autonomous region, F psr The smaller the value, the better the power supply reliability of the autonomous region obtained by division. str.j (t) is the power supply rate of autonomous region j at time t. DG.i (t) is the output power of the i-th distributed generation in the autonomous region, P str.i (t) is the charging and discharging power of the i-th energy storage device in the autonomous region, P l.i (t) is the load power of node i in the autonomous region, N clu.j.DG 、N clu.j.str 、N clu.j.l 、N clu.j.g are the number of distributed generation, energy storage devices, load nodes, and controllable traditional units in autonomous region j, and T is the optimization period.

[0085] (2) Constraints:

[0086] 1) The probability constraint of power supply rate in autonomous region, that is,

[0087]

[0088] In the formula, F min is the minimum value of the load power supply rate; x is the decision variable, that is, the division scheme of each autonomous area; ξ is the distributed energy output scenario; n is the number of distributed energy output scenarios; is the probability of the event being true; P ξ Contribute to a scenario in the distributed resource contribution scenario set; F(x,P ξ ) is the output P ξ is the load power supply rate of the system; β is the confidence level value satisfied by the objective function.

[0089] 2) Autonomous regional power balance indicators and constraints

[0090] The dynamic power balance of the autonomous region of the partitioning scheme is constrained to fully ensure the stability of power in the autonomous region. As the load and DG power change in real time, it brings certain risks to the safe and stable operation in the autonomous region. Based on this, the power balance index and constraint R of the autonomous region are proposed. p The definition is as follows:

[0091]

[0092] R p ≥β p (6)

[0093] Where: is the total active power demand of loads in the autonomous area; Make positive contributions to distributed energy resources; is the maximum active output of distributed energy storage in the autonomous region; α is the load power supply rate that needs to be met in the autonomous region; β p is the confidence level that the power balance constraint needs to meet.

[0094] Autonomous region power balance constraint R p It reflects the probability that the total DG output in the autonomous region can meet the load demand to a certain extent when the energy storage is at maximum output. The higher the index is, the higher the security of the autonomous region is and the more reasonable the division scheme is.

[0095] 3) System safety operation constraints

[0096] U i,min ≤u i (t)≤U i,max (7)

[0097] |I ij (t)|≤I ij.max (8)

[0098] Where: u i (t) is the voltage amplitude of node i at time t; U i,max and U i,min are the upper and lower limits allowed for node i; I ij (t) is the current amplitude of line ij at time t; I ij.max It is the maximum current allowed to flow through branch ij.

[0099] Step 3: Combined with the predicted output of photovoltaic and wind turbines and the prediction error probability model, the scenario analysis method is used to generate a distributed output scenario set that represents uncertainty within the specified reasonable output range, and the average output of distributed energy is obtained. The random simulation technology is used to solve the opportunity constraint considering the uncertainty of wind and solar output;

[0100] For the prediction errors of photovoltaic, wind turbine output and load, it is generally believed that they satisfy the normal distribution, and the probability density function of the prediction error is:

[0101]

[0102] Where ΔP is the prediction error of distributed energy output; μ and σ are the expectation and standard deviation of the prediction error, respectively, and μ is 0; e i is the value of each data point; n is the number of data points.

[0103] Latin hypercube sampling can avoid the problem of excessive clustering of sampling samples. The specific sampling steps are as follows:

[0104] Step 3.1: Divide the interval from 0 to 1 into l equal parts, and randomly generate a number P in each part according to uniform distribution. i .

[0105] Step 3.2: Shuffle the order of these l random numbers in any way, and these random numbers are the probability values ​​of each random sample.

[0106] Step 3.3: From the inverse function of the probability distribution function f -1 (P i )=x i Calculate the sample value corresponding to the probability value l of each sample.

[0107] There are generally two methods for dealing with chance constraints: analytical method and random simulation method. This paper uses random simulation method to verify whether the autonomous region division meets the chance constraint. The steps to check whether the division scheme meets the chance constraint confidence level at time t are as follows:

[0108] 1) According to the probability distribution characteristics of the load and uncontrollable DG power output in the autonomous area, a power curve is randomly extracted based on the scenario generation technology in step 2;

[0109] 2) Count the total number of times the autonomous region meets power balance in M ​​samplings. p and the total number of times M that the probability constraint of the objective function of the load loss risk rate of the autonomous region is satisfied v ;

[0110] 3) If It is considered that the autonomous region division scheme at that moment can meet the opportunity constraint. Otherwise, the next autonomous region division scheme is solved until the optimal autonomous region division scheme is found under the confidence level that meets the objective function and the opportunity constraint.

[0111] Due to the randomness of wind power and photovoltaic output, there is a widespread deviation in wind power and photovoltaic prediction. At present, the deviation of wind speed prediction from multiple prediction methods is 25% to 40%. Although the ARIMA prediction method proposed in this patent can effectively guarantee the prediction accuracy of distributed output, there is still a prediction deviation. Therefore, if the autonomous area division is carried out completely according to the predicted output, there will be a large deviation, and the power quality and stability of the system after the division will also be affected. However, if the prediction of distributed output is completely ignored, the distributed output prediction error is randomly generated according to the above-mentioned Weibull distribution function, and its output is calculated, it has a large blindness and cannot guarantee the rationality of the division result. Therefore, if randomness and the regularity of prediction are combined, it can avoid blindness and ensure the rationality of the division result. This article introduces the method of interval random simulation to combine the two.

[0112] Taking into account the prediction error and volatility of wind power and photovoltaic power, assuming that the output level of wind power and photovoltaic power at each moment fluctuates between the output deviation of ±20%, taking photovoltaic power as an example, its predicted output range is as follows: Figure 4 As shown. Combining the predicted output and prediction error probability model of photovoltaic and wind turbines, random simulation technology is used to solve the opportunity constraint. The six hours from 12:00 to 18:00 are selected as the autonomous area division period. The rolling optimization cycle is 1 hour, the feedback correction cycle is 20 minutes, and the confidence level of the opportunity constraint is set to 80%. The photovoltaic and wind turbines in this period are considered to generate random simulation scenarios for the predicted output range. When the sampling number is 1000, the photovoltaic samples in this period are as follows: Figure 5 shown.

[0113] Step 4: Construct a new autonomous area division optimization model for the distribution system based on SMPC. In the rolling optimization part, the breadth-first search algorithm is used to collect information on all division schemes, calculate the objective function of all schemes that meet the constraints under a certain confidence level, and determine the autonomous area division scheme within the first rolling optimization time period in the future; during feedback correction, the predicted values ​​of wind turbines, photovoltaics and loads are updated according to the latest real-time measurement values.

[0114] The optimization results of the dynamic autonomous region partitioning based on opportunity constraints are shown in Table 2.

[0115] Table 2 Results of dynamic autonomous region division of stochastic model predictive control based on chance constraints

[0116]

[0117] The results of dynamic autonomous area division of stochastic model predictive control based on chance constraints can be obtained from Table 2. The results of division of each time period are as follows: Figure 5-7 The average load loss risk rate of the autonomous region corresponding to the division results of each time period is shown in Table 3.

[0118] Depend on Figure 5 It can be seen that from 12:00 to 15:00, due to the large photovoltaic installed capacity, photovoltaic output cooperates with wind turbines and energy storage as the main active support of the autonomous area of ​​the distribution network during this period. At this time, the autonomous area is divided into two parts from the No. 3 cutting line. The average load loss risk rate in each autonomous area is 36.2%, which is relatively small. The load in each autonomous area can be effectively supported. The results of the autonomous area division during the 15:00-15:20 and 15:20-16:00 periods are as follows: Figure 6-7 As shown, as the photovoltaic output gradually weakens in the evening, wind turbines and energy storage gradually replace photovoltaics as active support to ensure the power supply rate in each autonomous region.

[0119] Table 3 Average load loss risk rate of autonomous regions in the division results

[0120]

[0121]

[0122] Through division, the control dimension and difficulty of the distribution network are effectively reduced, and the node structure of each autonomous area is relatively reasonable. Each autonomous area contains distributed energy nodes, which can meet the power supply rate requirements of each autonomous area, facilitate participation in regulation, and the division results of autonomous areas are relatively reasonable.

Claims

1. An autonomous region partitioning method based on stochastic model predictive control under chance constraints, characterized in that: The method comprises: Step 1: After obtaining all possible autonomous region division schemes by the branch and cut enumeration method, the wind and photovoltaic output is predicted using the Autoregressive Integrated Moving Average Model (ARIMA) to generate the output power of photovoltaics, wind turbines and loads in the prediction time domain; Step 2: Considering the uncertainty of wind and solar power output, taking the minimum risk rate of load loss in the autonomous region as the objective function, a partition constraint covering the opportunity constraint of power supply rate in the autonomous region, the opportunity constraint of power balance in the autonomous region, and the partition constraint of system safe operation is constructed; Step 3: Combined with the predicted output of photovoltaic and wind turbines and the prediction error probability model, the scenario analysis method is used to generate a distributed output scenario set that represents uncertainty within the specified reasonable output range, and the average output of distributed energy is obtained. The random simulation technology is used to solve the opportunity constraint considering the uncertainty of wind and solar output; Step 4: Construct a new distribution system autonomous area division optimization model based on stochastic model predictive control (SMPC). In the rolling optimization part, a breadth-first search algorithm is used to collect information on all division schemes, calculate the objective function of all schemes that meet the constraints under a certain confidence level, and determine the autonomous area division scheme within the first rolling optimization time period in the future; during feedback correction, the predicted values ​​of wind turbines, photovoltaics and loads are updated according to the latest real-time measurement values; Step 5: Repeat steps 2 to 4 to complete the division of autonomous regions at subsequent times.

2. The autonomous region partitioning method based on stochastic model predictive control under chance constraints according to claim 1 is characterized in that: In step 1, the prediction module is used to generate the output power of photovoltaic, wind turbine and load in the prediction time domain, as follows: The prediction model uses the ARIMA model. The core idea is to stabilize the sequence by the difference method, and use the regression moving average model to predict the time series after the stabilization process. The ARIMA model is expressed as ARIMA (p, d, q). The model consists of two models: the autoregressive model (AR) and the moving average model (MA). p is the order of the autoregressive model, q is the order of the moving average, and d is the degree of difference. ARIMA stabilizes the data through the difference operation. Generally speaking, the data can be stabilized after one difference, so the value of d is generally 0, 1, or 2. The mathematical formula of ARIMA(p, d, q) can be described as follows: Where: Y(t) is the value of the time series, t is the time point, c is the constant term, is the autoregressive coefficient, ε t are the interference terms θ1,θ2,…,θ q is the moving average coefficient, p is the number of autoregressive terms in the AR model, and q is the number of moving average terms in the MA model; The process of building an ARIMA model is divided into the following three steps: Step 1.1: Stationarity test: check whether the time series meets the stationarity requirements. For non-stationary time series, d-order differences must be performed first to convert them into stationary time series. Step 1.2: Calculate the autocorrelation coefficient (ACF) and partial autocorrelation coefficient (PACF) of the stationary time series, and perform parameter estimation through Akaikes information criterion (AIC) or Bayesian information criterion (BIC) search to obtain the optimal order p and q; Step 1.3: The ARIMA model is obtained through the above d, p, q, and the obtained model is model-checked, that is, the residual value ε t Perform analytical checks to determine the suitability of the time series ARIMA model and determine E(ε t ) is zero, which means the obtained model passes the test.

3. The autonomous region partitioning method based on stochastic model predictive control under chance constraints according to claim 1 is characterized in that: The objective function and constraints in step 2 are as follows: (1) Objective function Distributed renewable energy has flexible operation modes and can be connected to the grid or off-grid. When a major accident occurs in the power grid and a large-scale power outage occurs, the autonomous region can rely on the distributed renewable energy in the autonomous region to provide power support for the isolated operation of the autonomous region, quickly restore the power supply of important loads in the autonomous region, reduce the economic losses caused by large-scale power outages in the autonomous region, and enable the autonomous region to have independent autonomous operation capabilities. This paper takes the minimum risk rate of load loss in the autonomous region as the goal, and establishes the objective function based on opportunity constraints as follows: In the formula, F psr is the target value of the load loss risk rate in the autonomous region, F psr The smaller the value, the better the power supply reliability of the autonomous region obtained by division. str.j (t) is the power supply rate of autonomous region j at time t; P DG.i (t) is the output power of the i-th distributed generation in the autonomous region, P str.i (t) is the charging and discharging power of the i-th energy storage device in the autonomous region, P l.i (t) is the load power of node i in the autonomous region, N clu.j.DG 、N clu.j.str 、N clu.j.l 、N clu.j.g are the number of distributed generation, energy storage devices, load nodes, and controllable traditional units in autonomous region j, and T is the optimization period; (2) Constraints: 1) The probability constraint of power supply rate in autonomous region, that is, In the formula, F min is the minimum value of the load power supply rate; x is the decision variable, that is, the division scheme of each autonomous area; ξ is the distributed energy output scenario; n is the number of distributed energy output scenarios; is the probability of the event being true; P ξ Contribute to a scenario in the distributed resource contribution scenario set; F(x,P ξ ) is the output P ξ is the load power supply rate of the system; β is the confidence level value satisfied by the objective function; 2) Autonomous regional power balance indicators and constraints The dynamic power balance of the autonomous region of the partitioning scheme is constrained to fully ensure the stability of the power in the autonomous region. As the load and DG power change in real time, it brings certain risks to the safe and stable operation in the autonomous region. Based on this, the power balance index and constraint R of the autonomous region are proposed. p The definition is as follows: R p ≥β p (6) Where: is the total active power demand of loads in the autonomous area; Make positive contributions to distributed energy resources; is the maximum active output of distributed energy storage in the autonomous region; α is the load power supply rate that needs to be met in the autonomous region; β p is the confidence level that the power balance constraint needs to meet; Autonomous region power balance constraint R p It reflects the probability that the total output of DG in the autonomous region can meet the load demand to a certain extent when the energy storage is at its maximum output. The higher the index is, the higher the safety of the autonomous region is and the more reasonable the division scheme is. 3) System safety operation constraints U i,min ≤u i (t)≤U i,max (7) |I ij (t)|≤I ij.max (8) Where: u i (t) is the voltage amplitude of node i at time t; U i,max and U i,min are the upper and lower limits allowed for node i; I ij (t) is the current amplitude of line ij at time t; I ij.max It is the maximum current allowed to flow through branch ij.

4. The autonomous region partitioning method based on stochastic model predictive control under chance constraints according to claim 1 is characterized in that: The specific details of step 3 are as follows: For the prediction errors of photovoltaic, wind turbine output and load, it is generally believed that they satisfy the normal distribution, and the prediction error probability density function is: Where ΔP is the prediction error of distributed energy output; μ and σ are the expectation and standard deviation of the prediction error, respectively, and μ is 0; e i is the value of each data point; n is the number of data points; Latin hypercube sampling can avoid the problem of excessive clustering of sampling samples. The specific sampling steps are as follows: Step 3.1: Divide the interval from 0 to 1 into l equal parts, and randomly generate a number P in each part according to uniform distribution. i ; Step 3.2: Shuffle the order of these l random numbers in any way, and these random numbers are the probability values ​​of each random sample; Step 3.3: From the inverse function of the probability distribution function f -1 (P i )=x i Calculate the sample value corresponding to the probability value l of each sample; There are generally two methods for dealing with chance constraints: analytical method and random simulation method. This paper uses random simulation method to verify whether the division of autonomous regions meets the chance constraints. The steps to check whether the division scheme meets the chance constraint confidence level at time t are as follows: 1) According to the probability distribution characteristics of the load and uncontrollable DG power output in the autonomous area, a power curve is randomly extracted based on the scenario generation technology in step 2; 2) Count the total number of times the autonomous region meets power balance in M ​​samplings. p and the total number of times M that the probability constraint of the objective function of the load loss risk rate of the autonomous region is satisfied v ; 3) If It is considered that the autonomous region division scheme at this moment can meet the opportunity constraint condition. Otherwise, the next autonomous region division scheme is solved until the optimal autonomous region division scheme under the confidence level that meets the objective function and the opportunity constraint is found. Due to the randomness of wind power and photovoltaic output, there is a widespread deviation in wind power and photovoltaic prediction. At present, most prediction methods have a wind speed prediction deviation of 25% to 40%. Although the ARIMA prediction method proposed in this patent can effectively guarantee the prediction accuracy of distributed output, there is still a prediction deviation. Therefore, if the autonomous area division is performed completely according to the predicted output, there will be a large deviation, and the power quality and stability of the system after the division will also be affected; but if the prediction of distributed output is completely ignored, the distributed output prediction error is randomly generated according to the above Weibull distribution function, and its output is calculated, it has a large blindness and cannot guarantee the rationality of the division result; therefore, if the randomness and the regularity of the prediction are combined, it can avoid blindness and ensure the rationality of the division result. This paper introduces the interval random simulation method to combine the two. For the power prediction curve of wind power and photovoltaic power, taking into account the prediction error and volatility of wind power and photovoltaic power, the output deviation is allowed to be ±20%. It is assumed that the output level of wind power and photovoltaic power at each moment fluctuates between the upper and lower deviations, which is the random number generation range of the wind power and photovoltaic power output at each moment.

Images