Safety risk assessment method and system for distribution area considering photovoltaic and electric vehicles

Abstract

The present invention belongs to the technical field of risk assessment, and in particular to a method and system for assessing the safety risks of distribution substations considering photovoltaic and electric vehicles. The method comprises: obtaining environmental information data of the area to be tested; obtaining a load probability model, inputting the environmental information data into the load probability model, and obtaining a target data probability distribution result; inputting the target data probability distribution result into a power grid flow model to obtain a flow calculation result; and analyzing the flow calculation result using a DCNN deep neural network model to obtain a risk assessment result. The present invention first calculates the load and output status in the distribution substation area through a power grid impact analysis probability model, and then calculates the distribution network flow in the substation area under fine scene division, and finally performs a safety risk assessment of the distribution substation area under multiple scenarios through a DCNN deep neural network.

Description

Technical Field

[0001] The present invention belongs to the technical field of risk assessment, and in particular relates to a method and system for assessing the safety risks of distribution stations taking into account photovoltaic and electric vehicles. Background Art

[0002] Electric vehicles are a powerful tool to address global warming. With the attention of governments around the world, their penetration rate is increasing year by year. The emergence of electric vehicles (EVs) and residential photovoltaic (PV) solar power generation has transformed the traditional distribution model from a one-way transaction between the grid and consumers to a dynamic two-way interaction between the grid and prosumers. The shift from overall power generation to a distributed network of flexible loads and generators puts traditional equipment (such as distribution transformers) at risk of failure when faced with new demands.

[0003] Traditionally, planners can design residential load distribution systems based on existing information, but this design approach requires regular updates and is predictable and difficult to change. Modern systems that mix electric vehicles, photovoltaic power generation, and other distributed energy resources (DER) will bring unprecedented levels of daily load variation and system load variation. In addition to the thermal, power quality, and protection limit constraints of distribution system planning, the uncertainty of distributed energy adoption has the potential to affect transmission and large-capacity power systems. At present, there have been some studies on the impact of electric vehicles and DER on distribution substations. Li Juan et al. proposed a feeder load balancing method for flexible distribution substations under large-scale electric vehicle access. Under the premise of ensuring that the line load rate is within a reasonable range, the charging behavior of electric vehicles is guided by time-of-use electricity prices, thereby improving the economic efficiency of electric vehicle electricity purchase. Zhang Bo et al. proposed a power supply restoration strategy that fully considers the characteristics and regulation capabilities of flexible DC interconnection for urban AC / DC hybrid distribution substations with integrated photovoltaic storage and charging power stations. However, deterministic predictions are not reliable enough in formulating long-term distribution system planning and operation strategies.

[0004] Therefore, with the massive access of electric vehicles and DERs, there is an urgent need for a distribution station area safety risk assessment method that can predict risks based on real-time data. Summary of the invention

[0005] In view of the above problems, the present invention provides a method for assessing the safety risks of power distribution areas taking into account photovoltaic and electric vehicles, the method comprising:

[0006] Obtain environmental information data of the area to be tested;

[0007] Obtain a load probability model, input environmental information data into the load probability model, and obtain a target data probability distribution result;

[0008] Input the target data probability distribution result into the power grid power flow model to obtain the power flow calculation result;

[0009] The DCNN deep neural network model is used to analyze the power flow calculation results to obtain risk assessment results.

[0010] Preferably, the environmental information data includes transformer capacity, the number of electric vehicles in the distribution network under study, the average charging power of electric vehicles, the number, average size and environmental condition data of rooftop photovoltaics.

[0011] Preferably, the load probability model includes an electric vehicle load probability model, a rooftop photovoltaic probability model and a residential load probability model.

[0012] Preferably, the target data probability distribution result includes the probability distribution of electric vehicle arrival, the probability distribution of residential load and the probability distribution of solar radiation.

[0013] Preferably, establishing the electric vehicle load probability model includes:

[0014] Preset the average value of the total number of electric vehicles;

[0015] Determine the number of electric vehicles charged in the distribution network based on the average number of total electric vehicles;

[0016] The total load function of electric vehicles is determined according to the number of electric vehicles.

[0017] Preferably, establishing the rooftop photovoltaic probability model includes:

[0018] Get solar PV array size and solar efficiency;

[0019] A rooftop solar power generation function is determined according to the solar photovoltaic array size and the solar energy efficiency.

[0020] Preferably, establishing the residential load probability model comprises:

[0021] Calculate the probability density function of the hourly residential load using the Rayleigh distribution.

[0022] Preferably, the DCNN deep neural network model consists of a convolutional component and a fully connected component.

[0023] The present invention also proposes a distribution station area safety risk assessment system considering photovoltaic and electric vehicles, the system comprising:

[0024] The first acquisition module is used to acquire environmental information data of the area to be tested;

[0025] The second acquisition module is used to acquire a load probability model, input the environmental information data into the load probability model, and obtain a target data probability distribution result;

[0026] An input module is used to input the target data probability distribution result into the power grid power flow model to obtain the power flow calculation result;

[0027] The analysis module is used to analyze the power flow calculation results using a DCNN deep neural network model to obtain risk assessment results.

[0028] Preferably, the second acquisition module is used to acquire the load probability model, including:

[0029] The second acquisition module is used to obtain the electric vehicle load probability model, the rooftop photovoltaic probability model and the residential load probability model.

[0030] Preferably, the target data probability distribution result includes the probability distribution of electric vehicle arrival, the probability distribution of residential load and the probability distribution of solar radiation.

[0031] Preferably, the second acquisition module is used to acquire the electric vehicle load probability model, including:

[0032] The second acquisition module is used to preset the average value of the total number of electric vehicles;

[0033] Determine the number of electric vehicles charged in the distribution network based on the average number of total electric vehicles;

[0034] The total load function of electric vehicles is determined according to the number of electric vehicles.

[0035] Preferably, the second acquisition module is used to acquire the rooftop photovoltaic probability model, including:

[0036] The second acquisition module is used to obtain the size of the solar photovoltaic array and the solar energy efficiency;

[0037] A rooftop solar power generation function is determined according to the solar photovoltaic array size and the solar energy efficiency.

[0038] Preferably, the second acquisition module is used to acquire a residential load probability model, including:

[0039] The second acquisition module is used to calculate the probability density function of the hourly residential load using Rayleigh distribution.

[0040] The present invention also proposes an electronic device for assessing the safety risk of distribution stations in photovoltaic and electric vehicles, including:

[0041] Processor and memory;

[0042] The processor calls the computer program stored in the memory to execute any of the above-mentioned methods for assessing safety risks in distribution station areas considering photovoltaic and electric vehicles.

[0043] The present invention also provides a computer-readable storage medium.

[0044] The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor is enabled to execute any of the above-mentioned methods for assessing safety risks in distribution station areas.

[0045] The present invention has the following beneficial effects:

[0046] The present invention firstly calculates the load and output status in the distribution station area through the grid impact analysis probability model, then calculates the distribution network flow in the station area under the fine scenario division, and finally evaluates the safety risk of the distribution station area under multiple scenarios through the DCNN deep neural network; the present invention is simulated and verified through a 9-bus, 12-household distribution system, which can verify that the proposed method can make accurate predictions on the safety risks of the distribution station area under different EV penetration rates and different PV access numbers. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A diagram showing a method for assessing the safety risks of distribution stations in consideration of photovoltaic and electric vehicles in an embodiment of the present invention is shown;

[0049] Figure 2 A flow chart of a method for assessing the safety risks of power distribution areas taking into account photovoltaic and electric vehicles in an embodiment of the present invention is shown;

[0050] Figure 3 The DCNN structure diagram in the embodiment of the present invention is shown;

[0051] Figure 4 The network structure diagram of DCNN in an embodiment of the present invention is shown;

[0052] Figure 5 A simulation example scenario setting diagram is shown in an embodiment of the present invention;

[0053] Figure 6 The transformer daily load curve diagram under 25% penetration rate EV+PV (scenario 2) in an embodiment of the present invention is shown;

[0054] Figure 7 A transformer load curve diagram under various scenarios in an embodiment of the present invention is shown;

[0055] Figure 8 The RMSE curve diagram of DCNN training in an embodiment of the present invention is shown;

[0056] Fig. 9 A probability curve diagram of the occurrence of transformer load predicted by DCNN in an embodiment of the present invention is shown;

[0057] Fig.10 A diagram of a distribution station area safety risk assessment system considering photovoltaic and electric vehicles in an embodiment of the present invention is shown;

[0058] Fig.11 A diagram of an electronic device for assessing the safety risks of a distribution station area taking into account photovoltaic and electric vehicles in an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0059] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0060] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware units or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0061] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the steps. For example, some steps may be decomposed, while some steps may be combined or partially combined, so the actual execution order may change according to the actual situation.

[0062] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example.

[0063] In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or sub-modules is not necessarily limited to those steps or sub-modules explicitly listed, but may include other steps or sub-modules not explicitly listed or inherent to these processes, methods, products, or apparatuses.

[0064] like Figure 1 As shown, the present invention also proposes a method for assessing the safety risks of distribution stations in consideration of photovoltaic and electric vehicles, the method comprising the following steps:

[0065] S1 obtains environmental information data of the area to be tested;

[0066] S2 obtains a load probability model, inputs environmental information data into the load probability model, and obtains a target data probability distribution result;

[0067] S3 inputs the target data probability distribution result into the power grid power flow model to obtain the power flow calculation result;

[0068] S4 uses the DCNN deep neural network model to analyze the power flow calculation results to obtain risk assessment results.

[0069] Specifically, in S1, the environmental information data includes transformer capacity, the number of electric vehicles in the distribution network under study, the average charging power of electric vehicles, the number and average size of rooftop photovoltaics, and environmental condition data. In this embodiment, the environmental information data comes from user input, that is, it is input according to the actual situation of the distribution network under study.

[0070] Specifically, in S2, the load probability model includes an electric vehicle load probability model, a rooftop photovoltaic probability model, and a residential load probability model; the target data probability distribution results include an electric vehicle arrival probability distribution, a residential load probability distribution, and a solar radiation probability distribution.

[0071] Specifically, establishing the electric vehicle load probability model includes: presetting an average value of the total number of electric vehicles; determining the number of electric vehicles charged in the distribution network based on the average value of the total number of electric vehicles; and determining the total load function of electric vehicles according to the number of electric vehicles;

[0072] Establishing the rooftop photovoltaic probability model includes: obtaining the size of the solar photovoltaic array and the solar energy efficiency; determining the rooftop solar energy power generation function according to the size of the solar photovoltaic array and the solar energy efficiency;

[0073] Establishing the residential load probability model includes: calculating the probability density function of the residential load per hour using Rayleigh distribution.

[0074] Specifically, the technical solution adopted by the present invention is: a method for safety risk assessment of distribution area considering electric vehicles and distributed photovoltaics, including the following stages:

[0075] In general, the distribution substation safety analysis model estimates the probabilistic results of transformer overload and aging risks by performing power flow analysis on given environmental conditions and input data of the distribution network.

[0076] like Figure 2 As shown in the figure, the distribution substation area safety risk assessment method consists of five parts: input data, distribution substation area load probability model, scenario generation and power flow calculation, DCNN multi-dimensional safety risk assessment and output.

[0077] First, the distribution secondary system model is defined as a network graph G = (H, L), where H and L represent the set of system main lines and secondary lines, respectively;

[0078] The user then inputs the transformer capacity, the number of EVs in the distribution network under study, the average EV charging power (kW), the number of rooftop PV panels, the average size (m 2 ) and environmental conditions data;

[0079] Next, probability distribution functions are derived from the input data to calculate the confidence space of possible values ​​of the data during the day, sampling the probability distributions of electric vehicle arrivals, residential loads, and solar radiation, and using them as inputs for power flow analysis;

[0080] The calculated probability distribution samples of transformer load and ambient temperature are then used to calculate transformer life loss.

[0081] The specific stages include:

[0082] 1) Establish a load probability model within the distribution station area. The specific steps are as follows:

[0083] First, define the time scale, where t is the time variable of the studied area, in hours.

[0084] (1) Electric vehicle load probability modeling:

[0085] For EV load modeling, the arrival of EVs at charging stations can be considered as independent events. Assume that the total number of EVs in the distribution network is the average value of the statistical value Assuming that the probability of an electric vehicle arriving at a charging station follows a normally distributed random variable, the number of electric vehicles charging in the distribution network is for:

[0086] (1);

[0087] In the formula, and are the expected value and variance of EV charging, respectively, which are calculated based on the historical statistical data of the distribution station area analyzed.

[0088] Then, multiply the number of electric vehicles in equation (1) by the average electric vehicle charging rated power to obtain the total load at time t shown in equation (2):

[0089] (2);

[0090] In the formula, Charging power for an average electric vehicle.

[0091] (2) Probabilistic modeling of rooftop photovoltaics considering temperature:

[0092] Similarly, for modeling rooftop solar and ambient temperature, kernel density estimation can be used to approximate the distribution of random variables in ambient conditions, such as temperature and solar radiation. This estimation uses a Gaussian kernel and the bandwidth specified by the Scott normal reference rule. At the same time, historical weather data for a specific geographic location can provide a probability distribution for the simulation of ambient conditions. The power generation of the i-th rooftop solar in the system is as follows:

[0093] (3);

[0094] In the formula, the power generated by the solar panel is calculated by the light intensity, solar panel area and efficiency. The power generation of the i-th rooftop solar energy in the system is By solar photovoltaic array size (Unit: m 2 ) and efficiency And the light intensity distribution function The Beta distribution function of solar radiation is used to calculate :

[0095] (4);

[0096] In the formula, , is the shape parameter, and is the real-time light intensity and the maximum light intensity during the calculation time.

[0097] (3) Residential load probability modeling:

[0098] For residential load modeling, we adopt Rayleigh distribution to calculate the probability density function of hourly residential load.

[0099] (5);

[0100] In the formula, is the expected value of a single residential load, is the historical average of the load of a single residence, is the number of residential buildings in the area.

[0101] 2) Distribution area safety risk assessment method. This method is divided into two parts, namely scenario analysis and power flow calculation, and DCNN evaluation. The detailed algorithm is given below.

[0102] After the establishment of the load probability model for grid impact analysis in the previous article, the loads of electric vehicles, rooftop photovoltaics considering temperature, and residential loads are modeled. The power flow submodel of the load probability model for grid impact analysis is responsible for calculating the secondary winding load, bus load, bus generation, and total transformer load. The power flow model is responsible for calculating the multi-cycle DC power flow model, and the sampled input data for each time period in the study day are given. Equation (6) represents the node power balance constraint.

[0103] (6);

[0104] In the formula, , and represent the solar generation, electric vehicle load, and residential load on bus j, respectively. The total transformer load value generated by this power flow submodel is used as input to the transformer aging model submodel when calculating the life loss impact. The transformer secondary winding power is given by:

[0105] (7);

[0106] In the formula, represents the secondary line flow from bus j to bus k, is the system admittance and and are both bus voltage angles.

[0107] 3) DCNN-based distribution area risk assessment method

[0108] According to the load probability model of power grid impact analysis and the power grid flow calculated by historical data, a DCNN-based security risk assessment method for distribution substations is proposed. The security risk of distribution substations is evaluated through the active and reactive matrix of nodes and lines.

[0109] The DCNN deep learning network is used to perform safety risk assessment based on the line characteristics in the scenario, including the active power matrix and reactive power matrix of the nodes and lines. The unique features of DCNN include sparse interactions (also known as sparse connections or sparse weights), parameter sharing, and equivariant representation, which enable the prediction model to automatically extract high-level features from feature data and make full use of these features to assess risk values. The structure of the DCNN used in this article refers to Figure 3 It should be noted that the DCNN network is a deep learning architecture whose structure and parameters are not fixed. The DCNN used in this article is just one of them. DCNN mainly consists of two types of layers: convolutional layers and fully connected layers. The convolutional layer is an operation on the input and the kernel, used to perform linear operations.

[0110] Specifically, each unit in a convolutional layer is connected to the feature map of the previous layer through a set of weights called filters. The result of this local weighted sum is then passed through various layers, including rectified linear units (ReLU) and batch normalization (BN), to form the feature map of the next layer. The fully connected layer performs matrix multiplication using a separate weight matrix that describes the pairwise interactions between all input and output units. That is, the value of the i-th unit in the second fully connected layer (i.e. ) is obtained by adding all units in the previous layer (i.e. , ,..., ) multiplied by a weight matrix.

[0111] like Figure 4 As shown in Figure 1, the DCNN model is a deep convolutional neural network whose overall architecture consists of eight components.

[0112] The first six components build the convolutional layers, which include one first convolutional layer and five second convolutional layers. The second convolutional layer contains a convolutional layer, a batch normalization, and a linear rectification unit component. The first convolutional layer also contains an additional maximum pooling layer on the basis of the second convolutional layer for the output of the linear rectification unit at this stage. Batch normalization is a widely used technique in deep learning that helps reduce the distribution of input changes in each layer during training to improve training results. The linear rectification unit is used to introduce nonlinear characteristics into the neural network to improve the expressiveness of the model.

[0113] The last four components are fully connected components, which include three first fully connected layers and one second fully connected layer. The first fully connected layer is connected after the second convolutional layer, the second fully connected layer is connected after the first fully connected layer, and the output of the second fully connected layer is received by the regression layer, which outputs the target capacity estimation.

[0114] The DCNN model in the present invention has more than 28,000 parameters and about 6,000 neurons. It consists of five convolutional layers and three fully connected layers, and can be used to complete tasks such as classification based on image data and target capacity estimation. Its deep convolutional structure can extract features from images and realize hierarchical representation of features through multiple levels of convolution and pooling operations. This hierarchical feature representation can overcome the defects in traditional feature extraction methods, thereby improving the accuracy of classification and estimation.

[0115] During the training phase, the DCNN model needs to identify unknown parameters (e.g., weights and biases of convolutional and fully connected layers) during the training process. In order to correctly identify these parameters, the present invention defines a cost function J to measure the difference (or generalization error) between the model prediction and the relevant ground truth. The expected generalization error given by the cost function is minimized by a widely used optimization method, namely stochastic gradient descent (SGD) with momentum. SGD with momentum updates the parameters θ (weights ω and biases b) to minimize the generalization error by taking small steps in the direction of the negative gradient of the cost function. The above process is repeated multiple times, and each iteration is performed on a small set of sample training samples until the generalization error is close to zero. The cost function J(θ) with a regularization term is defined as

[0116] (8);

[0117] in, represents the cost function, represents the regularization term, represents the L2 regularization factor that measures the relative contribution of the norm penalty term, and represents the hypothesis function, m represents the number of samples used in each iteration, represents the i-th input sample matrix, Indicates the corresponding target value.

[0118] Assumption Function It can be expressed as:

[0119] (9);

[0120] in, and Respectively represent the nth unknown parameter and its corresponding input variable, is the deviation, .parameter It can be updated iteratively according to the following formula:

[0121] (10);

[0122] (11);

[0123] where g represents the estimate of the exact gradient for sampling the minimum set of q samples, represents the i-th input matrix of the minimum sample set in the j-th selection generation, Represents the corresponding target value, which determines the gradient contribution of the previous iteration to the current iteration. represents the parameter estimate in the jth iteration.

[0124] This embodiment is verified by simulation through a 9-bus, 12-household power distribution system.

[0125] Step 1: Model parameter setting

[0126] The Grid Impact Analysis Load Probabilistic Model is flexible enough to define custom distribution systems with arbitrary configurations of electric vehicle chargers, rooftop solar generators, transformer capacities, and environmental conditions for analysis. As a result, the Grid Impact Analysis Load Probabilistic Model can be applied to a wide range of load scenarios, study dates, and locations.

[0127] The examples in this paper are implemented on a server with a 2.60GHz Intel Pentium G620 CPU and 2GB RAM, and the system is Ubuntu Linux 16.04 LTS.

[0128] like Figure 5 As shown in Figure 1, the simulation example in this paper is a 9-bus, 12-household distribution system with a 50kVA distribution transformer and 8 secondary lines. The secondary lines are 4 / 0A WG aluminum cables. The rated power of all electric vehicle chargers in the system is 12.9kW, which is an average between the commonly used 6.6kW and the highest power of commercially available residential charging systems of 19.2kW. The peak power of the photovoltaic system in the study system is 6kW DC.

[0129] The allocation of four EV chargers and four PV arrays in the study case is as follows: Figure 5 As shown. Define system parameters for each study case according to the required analysis scope Study date δ, EV charger power and PV array size , where the parameters and is the average EV charger power (in kW) and PV array size (in m2) for the households in the studied system. 2 ). Study date parameters is the day of the year for the analysis.

[0130] After the model is initialized, the study date Input datasets used to determine the required study days for the year, such as residential loads, solar radiation, and ambient temperature.

[0131] Step 2: Scenario Analysis and Power Flow Calculation

[0132] A data-driven approach is used to perform multi-cycle power flow analysis on the distribution secondary system to quantify the load and casualty risks probabilistically. The simulation scenario settings are shown in Table 1.

[0133] Table 1

[0134]

[0135] The median transformer load curve under the 25% penetration rate EV and PV load scenario is as follows: Figure 6 As shown. Figure 6 The results of the analysis of the possibility and consequences of transformer load overload events. It can be seen that the peak of the load occurs around 20h, when the sun has set, the photovoltaic power generation power is almost zero, and EV users have started charging, so the median transformer load has a peak. In addition, although the median load value exceeds the 130% load threshold of 65kVA for only a few hours, the maximum possible load value will cause a transformer with a capacity of 50kVA to be severely overloaded, and even a 75kVA transformer with a load threshold of 97.5kVA to be overloaded.

[0136] Figure 7 The transformer load curves for different scenarios are shown. The base scenario does not include any EVs and PVs. The load on a specific bus and the location of the generator do not affect the total transformer load, so Figure 7 The above factors are not considered in the scenario. Transformer load illustrates the overall change in the load profile caused by EV load and PV generation based on the settings of the base scenario. During the midday period, solar generation contributes to the reduction of the total transformer load in the EV+PV load profile. In all scenarios, EVs increase the total load throughout the day compared to the base case, and the increase in total load is particularly large during the 17:00-22:00 period (the period when EV users are expected to return home and start charging). The contribution of rooftop PV reduces the load during the midday, and this phenomenon demonstrates the impact of the time desynchronization between the peak solar generation at noon and the peak residential EV charging demand in the evening. Compared with the base case, the higher load caused by the start of EV charging and the extended late night load lead to a shorter transformer cooling time, which increases transformer losses and aging.

[0137] Based on the load results, the total overload risk of the 50kVA study transformer and the 75kVA study transformer is predicted , as shown below:

[0138] (11);

[0139] Where, the total overload risk In equation (11), any single load event k(t) causes an overload condition Assume that the overload threshold τ is 130% of the transformer rating, i.e. 65kVA and 97.5kVA respectively (the threshold is Figure 7 The above formula gives the calculation method of the total overload risk of the transformer.

[0140] Step 3: DCNN network training and prediction

[0141] The DCNN is trained using a dataset (provided by the user). The root mean square error (RMSE) during training is as follows: Figure 8 As shown, it is a measure of the difference between the value predicted by the model or estimator and the observed value. The calculation formula of RMSE is as follows

[0142] (12);

[0143] in, is the predicted value of the ith observation in the data set, is the actual value of the ith observation in the data set, and n is the sample size. The smaller the RMSE, the better the model fits the data.

[0144] Use the trained DCNN to predict the transformer load. Fig. 9 The load prediction results of DCNN in five scenarios are shown. The average expected value of transformer load in the base scenario is the lowest among all scenarios, which is about 27kW.

[0145] Table 2

[0146]

[0147] The risk of transformer overload is shown in Table 2, which summarizes the risk values ​​of overload for a 50kVA transformer or a 75kVA transformer. In this case, upgrading from a 50kVA transformer to a 75kVA transformer reduces the overload risk by a factor of 10; adding rooftop solar power reduces the overload risk of both transformer types by 10% compared to the pure electric vehicle study scenario.

[0148] like Fig.10 As shown, the present invention also proposes a power distribution area safety risk assessment system considering photovoltaic and electric vehicles, the system comprising:

[0149] The first acquisition module 10 is used to acquire environmental information data of the area to be tested;

[0150] The second acquisition module 20 is used to acquire a load probability model, input the environmental information data into the load probability model, and obtain a target data probability distribution result;

[0151] An input module 30 is used to input the target data probability distribution result into the power grid power flow model to obtain the power flow calculation result;

[0152] The analysis module 40 is used to analyze the power flow calculation results using a DCNN deep neural network model to obtain risk assessment results.

[0153] like Fig.11 As shown, corresponding to the above-mentioned distribution station area safety risk assessment method considering photovoltaic and electric vehicles, the present invention also provides an electronic device for distribution station area safety risk assessment considering photovoltaic and electric vehicles. Since the embodiment of the device is similar to the above-mentioned method embodiment, the description is relatively simple. Please refer to the description of the above-mentioned method embodiment for relevant points. The device described below is only schematic. The device may include: a processor (processor) 1, a memory (memory) 2 and a communication bus (i.e., the above-mentioned device bus) and a search engine, wherein the processor 1 and the memory 2 communicate with each other through the communication bus and communicate with the outside through the communication interface. The processor 1 can call the logic instructions in the memory 2 to execute the distribution station area safety risk assessment method considering photovoltaic and electric vehicles.

[0154] In addition, the logic instructions in the above-mentioned memory 2 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: a storage chip, a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, and other media that can store program codes.

[0155] On the other hand, an embodiment of the present invention further provides a processor-readable storage medium, on which a computer program 3 is stored. When the computer program 3 is executed by the processor 1, the method for assessing the safety risk of distribution station areas considering photovoltaic and electric vehicles provided in the above-mentioned embodiments is implemented.

[0156] The processor-readable storage medium can be any available medium or data storage device that can be accessed by the processor 1, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NANDFLASH), solid-state drive (SSD)), etc.

[0157] Those skilled in the art should understand that although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible to modify the technical solutions described in the aforementioned embodiments, or to make equivalent replacements for some of the technical features therein; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assessing the safety risk of distribution stations considering photovoltaic and electric vehicles, characterized in that: The method comprises: Obtaining environmental information data of the area to be tested; the environmental information data includes transformer capacity, the number of electric vehicles in the distribution network under study, the average charging power of electric vehicles, the number, average size and environmental condition data of rooftop photovoltaics; Obtain a load probability model, input environmental information data into the load probability model, and obtain a target data probability distribution result; The load probability model includes an electric vehicle load probability model, a rooftop photovoltaic probability model and a residential load probability model; the target data probability distribution result includes an electric vehicle arrival probability distribution, a residential load probability distribution and a solar radiation probability distribution; The electric vehicle load probability model is: ； In the formula, is the total load at time t; Charging power for average electric vehicles; is the number of electric vehicles charged in the distribution network; the arrival of electric vehicles at the charging pile is considered an independent event, and the probability of electric vehicles arriving at the charging pile obeys a normally distributed random variable, then ,in and are the expected value and variance of EV charging, respectively, calculated based on the historical statistical data of the distribution station area analyzed; Indicates that the total number of electric vehicles of users in the distribution network is the average value of the statistical value; The rooftop photovoltaic probability model is: ； In the formula, is the power generation of the i-th rooftop solar energy at time t; is the solar photovoltaic array size; for solar photovoltaic array efficiency; is the light intensity distribution function, calculated using the Beta distribution function of solar radiation: ,in and is the shape parameter, is the real-time light intensity received by the i-th rooftop solar energy, is the maximum light intensity received by the i-th rooftop solar energy during the calculation time; The residential load probability model is: ； In the formula, is the expected value of a single residential load, is the historical average of the load of a single residence, is the number of residential buildings in the area, For the moment, is the probability density function of residential load; Input the target data probability distribution result into the power grid power flow model to obtain the power flow calculation result; Among them, the power flow model is responsible for calculating the multi-cycle DC power flow model, and the sampling input data for each time period in the study day is given. The node power balance constraint formula is as follows; ； In the formula, , ,and Respectively represent Solar power generation, electric vehicle load and residential load on bus j in time period; When calculating the impact of life loss, the total transformer load value generated by the power flow model is used as the input of the transformer aging model sub-model, and the transformer secondary winding power is calculated as follows: ； In the formula, represents the secondary line flow from bus j to bus k, is the system admittance, and All are bus voltage angles; The DCNN deep neural network model is used to analyze the power flow calculation results according to the line characteristics in the scenario to obtain risk assessment results; the line characteristics include active power matrix and reactive power matrix of nodes and lines.

2. The method for assessing the safety risk of distribution stations considering photovoltaic and electric vehicles according to claim 1 is characterized in that: Establishing the electric vehicle load probability model includes: Preset the average value of the total number of electric vehicles; Determine the number of electric vehicles charged in the distribution network based on the average number of total electric vehicles; The total load function of electric vehicles is determined according to the number of electric vehicles.

3. The method for safety risk assessment of distribution area considering photovoltaic and electric vehicles according to claim 1 is characterized in that: The rooftop photovoltaic probability model is established, including: Get solar PV array size and solar efficiency; A rooftop solar power generation function is determined according to the solar photovoltaic array size and the solar energy efficiency.

4. The method for safety risk assessment of distribution station areas considering photovoltaic and electric vehicles according to claim 1 is characterized in that: Establishing the residential load probability model includes: Calculate the probability density function of the hourly residential load using the Rayleigh distribution.

5. The method for safety risk assessment of distribution area considering photovoltaic and electric vehicles according to claim 1 is characterized in that: The DCNN deep neural network model consists of convolutional components and fully connected components.

6. A distribution area safety risk assessment system considering photovoltaic and electric vehicles, used to execute the method according to any one of claims 1 to 5, characterized in that: The system comprises: The first acquisition module is used to acquire environmental information data of the area to be tested; the environmental information data includes transformer capacity, the number of electric vehicles in the distribution network under study, the average charging power of electric vehicles, the number and average size of rooftop photovoltaics, and environmental condition data; The second acquisition module is used to acquire a load probability model, input the environmental information data into the load probability model, and obtain a target data probability distribution result; The second acquisition module is used to acquire a load probability model, including: The second acquisition module is used to obtain an electric vehicle load probability model, a rooftop photovoltaic probability model and a residential load probability model; the target data probability distribution results include an electric vehicle arrival probability distribution, a residential load probability distribution and a solar radiation probability distribution; The electric vehicle load probability model is: ； In the formula, is the total load at time t; Charging power for average electric vehicles; is the number of electric vehicles charged in the distribution network; the arrival of electric vehicles at the charging pile is considered an independent event, and the probability of electric vehicles arriving at the charging pile obeys a normally distributed random variable, then ,in and are the expected value and variance of EV charging, respectively, calculated based on the historical statistical data of the distribution station area analyzed; Indicates that the total number of electric vehicles of users in the distribution network is the average value of the statistical value; The rooftop photovoltaic probability model is: ； In the formula, is the power generation of the i-th rooftop solar energy at time t; is the solar photovoltaic array size; for solar photovoltaic array efficiency; is the light intensity distribution function, calculated using the Beta distribution function of solar radiation: ,in and is the shape parameter, is the real-time light intensity received by the i-th rooftop solar energy, is the maximum light intensity received by the i-th rooftop solar energy during the calculation time; The residential load probability model is: ； In the formula, is the expected value of a single residential load, is the historical average of the load of a single residence, is the number of residential buildings in the area, For the moment, is the probability density function of residential load; An input module is used to input the target data probability distribution result into the power grid power flow model to obtain the power flow calculation result; Among them, the power flow model is responsible for calculating the multi-cycle DC power flow model, and the sampling input data for each time period in the study day is given. The node power balance constraint formula is as follows; ； In the formula, , and Respectively represent Solar power generation, electric vehicle load and residential load on bus j in time period; When calculating the impact of life loss, the total transformer load value generated by the power flow model is used as the input of the transformer aging model sub-model, and the transformer secondary winding power is calculated as follows: ； In the formula, represents the secondary line flow from bus j to bus k, is the system admittance, and All are bus voltage angles; The analysis module is used to analyze the power flow calculation results using a DCNN deep neural network model to obtain risk assessment results.

7. The power distribution area safety risk assessment system considering photovoltaic and electric vehicles according to claim 6 is characterized in that: The second acquisition module is used to acquire the electric vehicle load probability model, including: The second acquisition module is used to preset the average value of the total number of electric vehicles; Determine the number of electric vehicles charged in the distribution network based on the average number of total electric vehicles; The total load function of electric vehicles is determined according to the number of electric vehicles.

8. The power distribution area safety risk assessment system considering photovoltaic and electric vehicles according to claim 6 is characterized in that: The second acquisition module is used to acquire the rooftop photovoltaic probability model, including: The second acquisition module is used to obtain the size of the solar photovoltaic array and the solar energy efficiency; A rooftop solar power generation function is determined according to the solar photovoltaic array size and the solar energy efficiency.

9. The power distribution area safety risk assessment system considering photovoltaic and electric vehicles according to claim 6 is characterized in that: The second acquisition module is used to acquire a residential load probability model, including: The second acquisition module is used to calculate the probability density function of the hourly residential load using Rayleigh distribution.

10. Consider the safety risk assessment of electronic equipment for photovoltaic and electric vehicle distribution areas, characterized in that, include: Processor and memory; The processor calls the computer program stored in the memory to execute the method for assessing the safety risks of power distribution areas taking into account photovoltaic and electric vehicles as described in any one of claims 1 to 5.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor is enabled to execute the method for assessing the safety risk of a distribution station area as described in any one of claims 1 to 5.

Images