A power distribution network power quality monitoring and comprehensive evaluation system and method

Abstract

The application relates to the technical field of smart grids, and discloses a power distribution network electric energy quality monitoring and comprehensive evaluation system and method, which comprises the following steps: collecting simulated electric energy quality treatment data; obtaining simulated control parameters of an existing photovoltaic grid-connected inverter according to simulated electric energy quality treatment data processing; generating simulation data; generating two-dimensional structure data according to the simulation data, inputting the two-dimensional structure data into a PVMFGCI strategy model, outputting a strategy, and the strategy containing the capacity of controllable source nodes that need to be added to all leading controlled nodes; and adjusting the power grid by executing the strategy; the application can intelligently divide a power grid treatment area, then dynamically adjust the operation parameters of the photovoltaic grid-connected inverter by applying a deep learning technology, and has high electric energy quality adjustment capacity.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart grid, more particularly, it relates to a power quality monitoring and comprehensive evaluation system and method for distribution network. BACKGROUND

[0002] With the wide application of photovoltaic power generation system in distribution network, photovoltaic inverters bear more responsibilities of power quality management. Traditional photovoltaic inverters mainly focus on realizing maximum power tracking, and have limited processing capacity for power quality problems. SUMMARY

[0003] The present application provides a power quality monitoring and comprehensive evaluation system and method for distribution network, which solves the technical problems in the related art.

[0004] The present application provides a power quality monitoring and comprehensive evaluation method for distribution network, comprising the following steps:

[0005] Step 100, collecting simulated power quality management data, the power quality management data including existing photovoltaic grid-connected inverter operation data, power grid overall operation data, power grid node operation data;

[0006] Step 200, obtaining simulated control parameters of the existing photovoltaic grid-connected inverter according to the simulated power quality management data processing; the control parameters including photovoltaic grid-connected inverter harmonic management, reactive power compensation, three-phase imbalance management coefficients;

[0007] Step 300, generating simulation data, one simulation data including simulated power quality management data and corresponding simulated control parameters collected in a fixed period within a set time period;

[0008] Step 400, generating two-dimensional structure data according to the simulation data, the two-dimensional structure data including a data matrix and a relationship matrix, one cell of the data matrix representing one-dimensional structure data of an independent object, the independent object including a controllable source node, a controlled node, and a power grid overall, and the cell of the data matrix only containing simulated data of the independent object represented thereby;

[0009] The element of the i-th row and the j-th column of the relationship matrix represents the association between the i-th cell and the j-th cell of the data matrix, and if the association exists, the element value of the relationship matrix is 1, otherwise 0;

[0010] One one-dimensional structure data includes n data items sorted by time, and the t-th data item represents the simulated data of the corresponding independent object collected at the t-th time point;

[0011] Step 500, input two-dimensional structure data into the PVMFGCI strategy model, the PVMFGCI strategy model comprising a first intermediate layer, a second intermediate layer and a first output layer, the first intermediate layer inputting one-dimensional structure data, outputting first intermediate representation data to the second intermediate layer, the second intermediate layer further inputting a relationship matrix, outputting second intermediate representation data to the first output layer, the first output layer outputting a strategy, the strategy containing capacities of controllable source nodes required to be added to all dominant controlled nodes;

[0012] Step 600, adjust the power grid according to the strategy output by step 500.

[0013] Further, the strategy is expressed as a vector, and one component of the vector represents a capacity of a controllable source node required to be added to a dominant controlled node.

[0014] Further, the calculation formula of the first intermediate layer is as follows:

[0015] u ( t ) =σ(WuX ( t)+WuH( t -1 ) +bu)

[0016] r ( t ) =σ(WrX(t)+WrH( t -1)+br)

[0017] C(t)=tanh(WcX( t) +Wcr (t) ⊙ H( t - 1) +bc)

[0018] H(t ) =(1-u (t) )⊙ C( t) +u (t) ⊙ H( t - 1 )

[0019] Wherein, W u , Wr, W c represent weight parameters, bu, br, bc represent bias parameters, and represents dot product. u(t), r(t) and C(t) represent first, second and third intermediate states respectively, wherein X(t) represents the tth data item of one-dimensional structure data, H(t) and H(t- 1 ) represent the tth and (t-1)th intermediate representation data respectively, n≥t≥1, n represents the total number of data items of one-dimensional structure data, and t=1. 1) = X(t), tanh is the hyperbolic tangent function, and σ represents the sigmoid function;

[0020] The calculation formula of the second intermediate layer is as follows:

[0021]

[0022] wherein kv represents the second intermediate representation data of the vth unit of the data matrix, M ( v ) is a set of units associated with the vth unit of the data matrix, represents the nth first intermediate representation data output by the first intermediate layer when the one-dimensional structure data input of the u th unit of the data matrix is input, c v is a normalization constant, and the default value is M ( v) is the total number of units, and σ is the sigmoid function, W k represents the third weight parameter.

[0023] Further, the calculation formula of the first output layer is as follows:

[0024]

[0025] wherein y represents the first output vector, the cth component value of which represents the probability value of the cth strategy, the strategy with the maximum probability value is selected as the output, and the strategy group contains all executable strategies, M ( al l ) represents a set of all units of the data matrix, Wy represents the output weight parameter, and by represents the output bias parameter.

[0026] Further, the association between the controllable source node and the controlled node means that the controllable source node is used to adjust the controlled node;

[0027] The association between controllable source nodes means that the controllable source nodes are located in the same power quality comprehensive management area;

[0028] The power grid as a whole is associated with all controllable source nodes and controlled nodes;

[0029] The association between controlled nodes means that the controlled nodes are located in the same power quality comprehensive management area.

[0030] Further, the deep learning model is used to obtain the simulation control parameters of the existing photovoltaic grid-connected inverter, including the following steps:

[0031] The simulated power quality treatment data is processed as sequence data, one sequence unit of the sequence data represents one collection of simulated power quality treatment data, the sequence data is input into a deep learning model, and a result representing a simulated control parameter of the photovoltaic grid-connected inverter is output.

[0032] Further, the calculation formula of the deep learning model is as follows:

[0033] h t =tanh(Whhh t -1+WxhX t +b h )

[0034] Wherein ht represents the tth intermediate feature, ht-1 represents the t-1th first intermediate representation data, xt represents the tth sequence unit of the sequence data, Wh h and Wx h are weight parameters, b h is a bias parameter, and tanh is a hyperbolic tangent function.

[0035] OUT=σ(Wouthmax+bout)

[0036] Wherein OUT represents a first output vector, a component value of which represents a simulated control parameter of one photovoltaic grid-connected inverter, hmax represents the last intermediate feature, Wou t is an out weight parameter, b out is an out bias parameter, and sigma represents a sigmoid function.

[0037] Further, the training of the deep learning model includes the following steps:

[0038] Step 101, initializing the parameters of the deep learning model;

[0039] Step 102, observing the simulated power quality treatment data St at time t, the strategy At executed at time t, the simulated power quality treatment data S t+1 t+1 obtained by executing the strategy At, and the reward Rt obtained by executing the strategy At;

[0040] Step 103, then calculate the strategy error:

[0041]

[0042] Wherein δt represents the strategy error at time t, γ represents a discount coefficient, γ∈[0, 1], represents the maximum probability value in the first output vector output by the deep learning model when St +1 is input, a probability value corresponding to the policy At in the first output vector output when the deep learning model input St is represented;

[0043] R t = w1*△THD + w2*△VUF + w3*△Q - λ*nswitc h

[0044] wherein, △THD is the reduction of total harmonic distortion rate from t to t+1 time;

[0045] △VUF is the reduction of voltage unbalance rate from t to t+1 time;

[0046] △Q is the reduction of reactive power loss from t to t+1 time;

[0047] nswitc h is the proportion of the capacity of the photovoltaic grid-connected inverter changed for use from t to t+1 time;

[0048] w1, w2, w3 are weight coefficients of three electric energy quality indicators, different weight values are given according to importance, λ is a switching penalty coefficient, and the loss of controlling the capacity use change of the photovoltaic grid-connected inverter;

[0049] Step 104, updating the deep learning model, and the updated formula is as follows:

[0050] V(St)←-V(St)+βδt

[0051] β∈[0,1], β represents the step length of deep learning, and ↑ represents transmission update;

[0052] Step 105, iteration steps 102-104, until the deep learning model converges or the iteration number reaches the set value.

[0053] The application provides a power distribution network electric energy quality monitoring and comprehensive evaluation system, comprising:

[0054] The first module collects simulated electric energy quality treatment data, and the electric energy quality treatment data includes existing photovoltaic grid-connected inverter operation data, overall power grid operation data (load current, harmonic component, reactive component and unbalanced current component) and each node operation data of the power grid;

[0055] The second module obtains simulated control parameters of the existing photovoltaic grid-connected inverter according to the simulated electric energy quality treatment data processing; the control parameters include photovoltaic grid-connected inverter harmonic treatment, reactive compensation and three-phase imbalance treatment coefficients;

[0056] The third module generates simulation data, and one simulation data includes simulated electric energy quality treatment data collected in a set time period and corresponding simulated control parameters.

[0057] The fourth module generates two-dimensional structural data based on the simulation data. The two-dimensional structural data includes a data matrix and a relationship matrix. A unit of the data matrix represents the one-dimensional structural data of an independent object. The independent objects include controllable source nodes, controlled nodes, and the entire power grid. A unit of the data matrix only contains the simulation data of the independent object it represents.

[0058] The fifth module inputs the two-dimensional structure data into the PVMFGCI strategy model. The PVMFGCI strategy model includes a first intermediate layer, a second intermediate layer and a first output layer. The first intermediate layer inputs the one-dimensional structure data and outputs the first intermediate representation data to the second intermediate layer. The second intermediate layer also inputs the relationship matrix and outputs the second intermediate representation data to the first output layer. The first output layer outputs the strategy, which includes the capacity of the controllable source nodes that need to be added to all the dominant controlled nodes.

[0059] The present invention provides a computer storage medium for storing computer-readable instructions, which can execute the aforementioned method for monitoring and comprehensively evaluating power quality of distribution network when the computer-readable instructions are read.

[0060] The beneficial effects of the present invention are:

[0061] When a new photovoltaic grid access point is added to the grid, the grid node layout will be changed, disrupting the original grid operation balance. Generally, a period of continuous observation is required to determine whether to add information-controllable source nodes to adjust grid parameters. Before adding a new photovoltaic grid access point to the grid, the present invention uses a method of simulating the addition of a new photovoltaic grid access point to perform simulated operation, and optimizes the existing photovoltaic grid-connected inverter. Then, deep learning technology is used to provide a grid transformation strategy after actual grid connection. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a block diagram of the PVMFGCI structure of the present invention;

[0063] Figure 2 This is the distribution diagram of harmonic distortion rate before and after comprehensive treatment in different areas;

[0064] Figure 3 It is the three-phase imbalance distribution map before and after comprehensive governance in different regions;

[0065] Figure 4 It is a flow chart of the regional division strategy for comprehensive management of power quality in AC and DC distribution networks of the present invention;

[0066] Figure 5 It is a novel distribution network topology of the present invention;

[0067] Figure 6 is the clustering validity function value of the present invention;

[0068] Figure 7 is a power quality comprehensive treatment regional division and dominant node diagram of the present application;

[0069] Figure 8 is a flow chart of a power quality monitoring and comprehensive evaluation method of a distribution network of the present application. DETAILED DESCRIPTION

[0070] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can include changes, modifications, additions or omissions of the functions and arrangements of the elements discussed without departing from the scope of the present description. Various examples can omit, substitute, or add various procedures or components as appropriate. Also, it should be understood that instead of a single arrangement of elements, a plurality of arrangements can be performed, either simultaneously or in succession.

[0071] In at least one embodiment of the present application, a power quality monitoring and comprehensive evaluation method of a distribution network is disclosed, as shown in Figure 8 comprising the following steps:

[0072] Step 100, collecting simulated power quality treatment data, the power quality treatment data including existing photovoltaic grid-connected inverter operation data, overall power grid operation data (load current, harmonic component, reactive component and unbalanced current component), and each node operation data of the power grid;

[0073] Step 200, obtaining simulated control parameters of the existing photovoltaic grid-connected inverter according to the simulated power quality treatment data processing; the control parameters including harmonic treatment, reactive compensation and three-phase imbalance treatment coefficients of the photovoltaic grid-connected inverter;

[0074] Step 300, generating simulation data, one simulation data including simulated power quality treatment data collected in a fixed period and corresponding simulated control parameters in a set time period;

[0075] Step 400, generating two-dimensional structure data according to the simulation data, the two-dimensional structure data including a data matrix and a relationship matrix, one cell of the data matrix representing one-dimensional structure data of an independent object, the independent object including a controllable source node, a controlled node and an overall power grid, and one cell of the data matrix only containing simulation data of the independent object represented thereby;

[0076] The element of the i-th row and the j-th column of the relationship matrix represents the association between the independent objects represented by the i-th cell and the j-th cell of the data matrix, and if there is an association, the element value of the relationship matrix is 1, otherwise it is 0;

[0077] The association between the controllable source node and the controlled node means that the controllable source node is used to adjust the controlled node;

[0078] The association between controllable source nodes means that the controllable source nodes are located in the same power quality comprehensive management area (see below);

[0079] The overall power grid is associated with all controllable source nodes and controlled nodes;

[0080] The association between controlled nodes means that the controlled nodes are located in the same power quality comprehensive management area;

[0081] A one-dimensional structure data includes n data items sorted by time, and the tth data item represents the simulation data of the corresponding independent object collected at the tth moment;

[0082] Step 500, input the two-dimensional structure data into the PVMFGCI strategy model, the PVMFGCI strategy model includes a first intermediate layer, a second intermediate layer and a first output layer, the first intermediate layer inputs one-dimensional structure data, outputs first intermediate representation data to the second intermediate layer, the second intermediate layer also inputs a relationship matrix, outputs second intermediate representation data to the first output layer, and the first output layer outputs a strategy, the strategy contains the capacity of all controllable source nodes needed to be added to the dominant controlled node.

[0083] In an embodiment of the present application, the strategy is represented as a vector, and one component of the vector represents the per unit value of the capacity of the controllable source node needed to be added to the dominant controlled node.

[0084] In an embodiment of the present application, the calculation formula of the first intermediate layer is as follows:

[0085] u ( t)=σ(WuX ( t ) +WuH( t -1 ) +bu)

[0086] r ( t)=σ(WrX(t ) +WrH( t -1)+br)

[0087] C ( t)=tanh(WcX( t )+Wcr (t) ⊙ H( t -1)+bc)

[0088] H(t)=(1-u (t) )⊙ C(t)+u ( t ) ⊙ H(t - 1)

[0089] wherein W u , Wr, W c represent weight parameters, bu, br, bc represent bias parameters, represents dot product, u(t), r(t) and C(t) represent the first, second and third intermediate states respectively, wherein X(t) represents the tth data item of the one-dimensional structure data, H(t) and H(t- 1 ) represent the tth and (t-1)th intermediate representation data respectively, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, H(t-1)=X(t) when t=1, tanh is the hyperbolic tangent function, and σ represents the sigmoid function;

[0090] The calculation formula of the second intermediate layer is as follows:

[0091]

[0092] wherein kv represents the second intermediate representation data of the vth cell of the data matrix, M(v) is a set of cells associated with the vth cell of the data matrix, represents the nth first intermediate representation data output by the one-dimensional structure data input of the uth cell of the data matrix to the first intermediate layer, c v is a normalization constant, and the default value is M ( the total number of cells in M(v), σ is the sigmoid function, W k represents the third weight parameter.

[0093] The calculation formula of the first output layer is as follows:

[0094]

[0095] wherein y represents the first output vector, the cth component value of y represents the probability value of the cth strategy, the strategy with the maximum probability value is selected as the output, the strategy group contains all executable strategies, M ( a 11 represents the set of all cells of the data matrix, W y represents the output weight parameter, and by represents the output bias parameter.

[0096] Step 600: adjusting the power grid by executing the strategy output by step 500.

[0097] It should be noted that the training data of the PVMFGCI strategy model is derived from a simulation system, the simulation system simulates the rail transit information to be dispatched, simulates a seismic source to perform simulation, and obtains the training data in the simulation environment

[0098] The training target of the PVMFGCI strategy model is minimization of total harmonic distortion rate, voltage unbalance rate and reactive power loss.

[0099] In one embodiment of the application, the power distribution network is divided into several power quality comprehensive management areas with high coupling degree between nodes. Due to the strong coupling of nodes in the area, there is a dominant management node. When the power quality pollution of the node is managed, the power quality pollution of the remaining nodes in the same area will also be greatly improved, thereby realizing the effect of greater regional management with smaller management capacity.

[0100] The application proposes a power quality comprehensive management area division strategy for AC / DC power distribution network, and the flow is as shown in Figure 4 The specific steps are as follows:

[0101] (1) Based on harmonic sensitivity, reactive power sensitivity and three-phase imbalance sensitivity, a comprehensive voltage sensitivity model is established, and distributed photovoltaic management resources in the power distribution network are considered to divide photovoltaic control areas considering three main power quality indicators.

[0102] (2) The DDTW-FCM clustering partition algorithm is used to process time series data to obtain the maximum clustering validity function and the nodes contained in each class, thereby forming power quality area division, and the node closest to the clustering center in each area is determined as the dominant node of the area management.

[0103] (3) Based on the spatial relationship between the two types of areas, the following method is adopted to define the power quality comprehensive management area: if the photovoltaic control area is located inside the power quality area, the comprehensive management area is determined as the range of the photovoltaic control area; if the photovoltaic control area overlaps with the power quality area, the photovoltaic control area and the power quality area corresponding to the dominant node covered by the photovoltaic control area are integrated to form the power quality comprehensive management area.

[0104] Such a definition method helps to effectively integrate the management resources of distributed photovoltaics, ensure photovoltaic control and coordination to improve power quality, and is conducive to better implementation of comprehensive management measures.

[0105] To verify the rationality of the partition management method in this paper, a Figure 5The new distribution network for reference model is modified from IEEE 14-bus network. The new distribution network is composed of 16-bus 220V AC subnetwork and 5-bus ±375V DC subnetwork. VSC is used to realize AC-DC interconnection in the distribution network, and the capacity of VSC is set to 200 kV·A. PV is connected to the distribution network through AC three-phase and DC. The system capacitor and ideal transformer are not considered for harmonic mitigation, and the remaining capacity of photovoltaic grid-connected inverter is used for power quality comprehensive management. The remaining capacity of photovoltaic grid-connected inverter at the grid-connected node is set to 18 kVA, and the output of solar cell array is matched with the maximum power. The maximum power is 15 kW. After setting the full capacity of photovoltaic grid-connected inverter for grid connection and harmonic mitigation, there is still a certain margin. This margin is not used for active grid connection or power quality management. In the example solution, the available capacity is 17 kVA. Combined with the typical daily active output curve of photovoltaic, the designed power quality management optimization strategy is verified by example.

[0106] According to the harmonic standard of China, the upper limit of harmonic voltage distortion rate THD is set to 4% in this chapter. N The harmonic distribution of system nodes used in the example is shown in the following table. Harmonic current frequencies of 5, 7, 11 and 13 are considered. Nodes 3, 7, 13 and 15 are connected to 20 kW unbalanced loads. These settings help to consider the impact of harmonic current in the analysis, especially in the case of unbalanced loads connected to nodes 3, 7, 13 and 15, for effective monitoring and control of power quality.

[0107] Table 1 Harmonic pollution data of nodes

[0108]

[0109] Based on the comprehensive voltage sensitivity model, the photovoltaic control area considering three power quality indicators can be divided in the AC subnetwork. Four photovoltaic devices have their control areas, and the controllable source nodes and their controlled nodes are shown in Table 2.

[0110] Through such division, each photovoltaic device has a clear control area, so that the system can be more targeted in considering multiple power quality indicators for management. This fine regional division helps to improve the effect of power quality management, so that each photovoltaic device can more effectively adjust the power quality parameters in its control area.

[0111] Table 2 Photovoltaic control area and its nodes

[0112]

[0113] From the information in the table, it can be seen that the 4 controllable source nodes correspond to different photovoltaic control areas. Since their power quality compensation form is current, this leads to the nodes on the same branch in each control area. Under the power quality governance function of photovoltaic, the power quality in the photovoltaic control area will be effectively improved. Essentially, harmonic compensation, reactive power compensation and unbalanced compensation of the system are all governance of current waveform. Through unified compensation, the system current becomes symmetrical and sinusoidal. This process helps to improve the power quality of the system and make it meet the harmonic standard and other related power quality requirements.

[0114] The power quality time series data of 16 nodes are dimensionally arranged, and the DDTW-FCM method is used for power quality partitioning. The nearest point to the cluster center in each partition is set as the dominant node in the partition, and the clustering effectiveness and partition results are shown in the table. Through this analysis, the distribution of power quality data can be clearly seen, and the dominant nodes in each partition are also marked in the table. This study helps to better understand the spatio-temporal distribution of power quality characteristics and provides important guidance for subsequent comprehensive power quality governance.

[0115] Through the analysis of the clustering effectiveness function, it can be known that when 4 regions are selected, the effect of region division is best. At this time, the coupling relationship between regions is weakest, while the coupling relationship within the region is strongest. According to the data in the table, the dominant control nodes of regions I, II, IV and V are 3, 7, 15 and 11 respectively. In consideration of the distributed photovoltaic governance resources, according to the new type of distribution network power quality comprehensive governance region division strategy, the power quality partitioning can be further determined.

[0116] Table 3 Power quality region division results

[0117]

[0118] Based on the spatial relationship between the two types of regions, the power quality comprehensive governance region is defined, and the AC subnetwork of AC-DC distribution network can be divided into 4 power quality comprehensive governance regions, as shown below. Such a definition method helps to effectively integrate the governance resources of distributed photovoltaic, ensure the photovoltaic control and coordination to improve power quality, and is conducive to better implementation of comprehensive governance measures.

[0119] In the low-voltage distribution network of our country, three-phase four-wire power supply mode is widely used. The four-bridge photovoltaic grid-connected inverter with active power filter function mainly includes grid-connected power generation, active power filter, maximum power point tracking (MPPT), instruction current calculation and tracking control technology. The structure diagram of photovoltaic multi-function grid-connected inverter (PVMFGCI) is as follows.

[0120] The main circuit of the system is composed of photovoltaic array, three-phase four-bridge inverter, filter inductance, parallel load and three-phase four-wire power grid. The control circuit is composed of signal detection unit, maximum power tracking (MPPT) control unit, instruction current operation unit and main control unit. This structure ensures the stable operation of photovoltaic grid-connected inverter in low-voltage distribution network, and has multiple key functions such as active power filter, which provides strong guarantee for the reliability and performance of the system.

[0121] The control circuit of photovoltaic grid-connected inverter monitors the load current of distribution network in real time, extracts harmonic, reactive and unbalanced current components. By generating compensation current reference value and combining with active current value of photovoltaic power generation, the final grid-connected instruction current is formed to ensure the unified regulation and control of grid-connected power generation and active power filter function. This research focuses on the combination of photovoltaic grid connection and parallel APF, aiming to better compensate the unbalanced power problems in the grid while delivering active power to the grid, so the MPPT is not described too much. Through this integrated control strategy, the system can more effectively coordinate photovoltaic power generation and active power filter function, improve power quality and provide reliable support for stable operation of power grid.

[0122] The total capacity of photovoltaic grid-connected inverter is S N , the capacity of DG grid-connected is S DG , and the remaining capacity of grid-connected inverter is S RE .

[0123]

[0124] In the formula, U is the voltage of distribution network, and I1 is the active current output by the inverter.

[0125] The relationship between harmonic control and reactive compensation capacity of photovoltaic grid-connected inverter and its remaining capacity can be expressed as:

[0126]

[0127] In the formula: S h , S q , S2 are the harmonic control, reactive compensation and three-phase imbalance control capacity of DG grid-connected inverter respectively

[0128] By reasonably configuring the location and capacity of the grid-connected inverter of the distribution network distributed generation (DG), the distribution of the remaining capacity of the inverter can be optimized, so as to realize the differentiated management of power quality in different areas. Through the optimization of the compensation strategy, the management effect of the photovoltaic grid-connected inverter in terms of harmonic, reactive, three-phase imbalance and the like can be more economical and reasonable, and the effective allocation of management resources can be realized

[50] -

[51] .

[0129] In order to establish a mathematical model of PVMFGCI capacity allocation, so as to solve the optimal compensation coefficients of the remaining capacity in terms of harmonic, reactive, three-phase imbalance, the following mathematical model is introduced to represent the coefficients in the three aspects:

[0130]

[0131]

[0132] In the formula, I1, I h , I q , I0 are the fundamental, harmonic, reactive and three-phase imbalance currents at the grid side.

[0133] Suppose that the harmonic management, reactive compensation and three-phase imbalance management coefficients of the remaining capacity of PVMFGCI are ε h , ε q , ε2 respectively, and the compensated power quality coefficients are

[0134]

[0135] In the formula, l h0 , l q0 , l 20 and l h , l q , l2 are the harmonic, reactive and three-phase imbalance coefficients before and after compensation respectively.

[0136] Solving PVMFGCI to establish a comprehensive power quality evaluation model under multiple operating scenarios

[0137]

[0138] In the formula, I p1 , I ph , I pq , I p2 are the active current, harmonic management, reactive compensation and three-phase imbalance management current of PVMFGCI output.

[0139] In view of the simplicity and practicability of the Lagrange multiplier method, this paper adopts the method to solve the optimal coefficient problem of the comprehensive power quality management of the remaining capacity of PVMFGCI.

[0140] Let λ be the Lagrange multiplier, and construct the Lagrange function as follows:

[0141]

[0142] The calculation by the Lagrange multiplier method is as follows:

[0143]

[0144] The power quality comprehensive treatment strategy of the residual capacity of the PVMFGCI is as follows:

[0145] (1) Calculate the residual capacity S of the grid-connected inverter RE , and set the residual capacity threshold Q N 0.

[0146] (2) If S RE ≥ QN0, and satisfies

[0147]

[0148] Directly perform full compensation, that is, εh=εq=ε0=1, otherwise, first calculate the compensation coefficients according to formula (10) and then perform compensation.

[0149] (3) If S RE ≤ Q N0 , do not put in compensation, that is, εh=εq=ε0=0.

[0150] In some other embodiments of the present application, a deep learning model is used to obtain the simulated control parameters of the existing photovoltaic grid-connected inverter, including the following steps:

[0151] The simulated power quality treatment data is processed into sequence data, one sequence unit of the sequence data represents once collected simulated power quality treatment data, the sequence data is input into the deep learning model, and a result representing the simulated control parameters of the photovoltaic grid-connected inverter is output.

[0152] The calculation formula of the deep learning model is as follows:

[0153] h t =tanh(W hh h t -1+WxhX t +b h )

[0154] Where ht represents the tth intermediate feature, ht-1 represents the t-1th first intermediate representation data, xt represents the tth sequence unit of the sequence data, W hh and Wx h are weight parameters, and b his a bias parameter, tanh is the hyperbolic tangent function;

[0155] OUT = sigma(Wout St + bout) x

[0156] where OUT represents a first output vector, a component value of which represents a simulation of a control parameter of a photovoltaic grid-connected inverter, hmax represents a last intermediate feature, Wout u t is an out weight parameter, b out is an out bias parameter, and sigma represents a sigmoid function;

[0157] The training of the deep learning model comprises the following steps:

[0158] Step 101, initializing parameters of the deep learning model;

[0159] Step 102, observing simulated power quality treatment data St at time t, a strategy At executed at time t, simulated power quality treatment data St +1 obtained by executing the strategy At, and a reward Rt;

[0160] Step 103, then calculating a strategy error:

[0161]

[0162] where δt represents a strategy error at time t, γ represents a discount coefficient, γ ∈ [0, 1], represents a maximum probability value in a first output vector output by the deep learning model when inputting S t+ 1, and V (S t) represents a probability value corresponding to the strategy At in the first output vector output by the deep learning model when inputting St.

[0163] R t = W1·ΔTHD + W2·ΔVUF + W3·△Q - λ·nswitc h

[0164] where ΔTHD is a reduction amount of total harmonic distortion rate from time t to time t+1;

[0165] ΔVUF is a reduction amount of voltage unbalance rate from time t to time t+1;

[0166] △Q is a reduction amount of reactive power loss from time t to time t+1;

[0167] nswitc h is a proportion of capacity of the photovoltaic grid-connected inverter replaced for use from time t to time t+1;

[0168] ​w1, w2, w3 are weight coefficients of three power quality indicators, different weights are given according to importance, and lambda is a switching penalty coefficient, which controls the loss of capacity use change of the photovoltaic grid-connected inverter;

[0169] In step 104, the deep learning model is updated, and the updated formula is as follows:

[0170] V( S t)←-V( S t)+βδt

[0171] Beta is in [0, 1], beta represents the step length of deep learning, and the arrow up represents transmission update.

[0172] In step 105, steps 102-104 are iterated until the deep learning model converges or the number of iterations reaches the set value.

[0173] Governance optimization result simulation verification:

[0174] In order to solve the optimal compensation coefficient of the power quality optimization strategy of the adjustable capacity of the photovoltaic grid-connected system for harmonics, reactive power and three-phase imbalance, corresponding mathematical modeling is carried out, and the capacity distribution and compensation coefficient are obtained, as shown in the following table.

[0175] Table 4 governance optimization results and compensation coefficients

[0176]

[0177] According to the data in table 1, the maximum output power of the photovoltaic system at 12:00 is selected as 15kW in this project. It is calculated that the power quality compensation capacity does not exceed the limit of the remaining capacity. This shows that the remaining capacity of the distributed photovoltaic system fully meets the demand of the power quality management of the distribution network, so that the active grid connection of the photovoltaic system can be fully realized without reduction. As shown in the following table, the changes of harmonic and three-phase imbalance indexes of each node in the AC subnetwork of the AC-DC distribution network before and after the comprehensive management of power quality. Figures 2-3

[0178] After the comprehensive management of power quality, the harmonic level of each node in the AC subnetwork is lower than 2%, and the three-phase imbalance degree is also significantly reduced. In addition, the ripple coefficient of the DC bus decreases from 1.75% before the management to 0.5% after the management. This shows that by implementing the power quality management measures in the AC subnetwork, the harmonic ripple introduced by the AC subnetwork is successfully reduced, further improving the power quality of the DC subnetwork and ensuring its stable operation.

[0179] In at least one embodiment of the present application, a power quality monitoring and comprehensive evaluation system for distribution network is provided, which is used to execute the power quality monitoring and comprehensive evaluation method for distribution network as described above, comprising:

[0180] ​The first module collects simulated power quality treatment data, the power quality treatment data including operation data of an existing photovoltaic grid-connected inverter, operation data of an overall power grid (load current, harmonic component, reactive component and unbalanced current component), and operation data of each node of the power grid;

[0181] The second module obtains simulated control parameters of the existing photovoltaic grid-connected inverter according to the simulated power quality treatment data processing, the control parameters including harmonic treatment, reactive compensation and three-phase imbalance treatment coefficients of the photovoltaic grid-connected inverter;

[0182] The third module generates simulation data, one simulation data including simulated power quality treatment data collected in a fixed period and corresponding simulated control parameters;

[0183] The fourth module generates two-dimensional structure data according to the simulation data, the two-dimensional structure data including a data matrix and a relationship matrix, one cell of the data matrix representing one-dimensional structure data of an independent object, the independent object including a controllable source node, a controlled node and an overall power grid, and the cell only containing simulated data of the independent object represented by the cell;

[0184] The fifth module inputs the two-dimensional structure data into a PVMFGCI strategy model, the PVMFGCI strategy model including a first intermediate layer, a second intermediate layer and a first output layer, the first intermediate layer inputting one-dimensional structure data, outputting first intermediate representation data to the second intermediate layer, the second intermediate layer further inputting a relationship matrix, outputting second intermediate representation data to the first output layer, and the first output layer outputting a strategy, the strategy containing capacities of controllable source nodes needed to be added to all dominant controlled nodes;

[0185] Step 600: adjusting the power grid according to the strategy output by step 500.

[0186] The above describes the embodiments of the present application, but the embodiments are not limited to the above specific embodiments, the above specific embodiments are only illustrative but not restrictive, and those skilled in the art can make more forms of equivalent embodiments under the inspiration of the embodiments, which are all within the protection scope of the embodiments.

Claims

1. A method for monitoring and comprehensively evaluating power quality in a distribution network, characterized in that: The following steps are involved: Step 100: Collect simulated power quality management data, which includes operating data of existing photovoltaic grid-connected inverters, operating data of the entire power grid, and operating data of each node in the power grid; Step 200: obtaining simulated control parameters of an existing photovoltaic grid-connected inverter based on the simulated power quality control data; the control parameters include harmonic control, reactive power compensation, and three-phase unbalance control coefficient of the photovoltaic grid-connected inverter; Step 300: Generate simulation data, where each simulation data includes simulated power quality management data collected at a fixed period within a set time period and corresponding simulated control parameters; Step 400: Generate two-dimensional structural data based on the simulation data. The two-dimensional structural data includes a data matrix and a relationship matrix. One cell of the data matrix represents one-dimensional structural data of an independent object. The independent object includes a controllable source node, a controlled node, and the entire power grid. A cell of the data matrix only contains simulation data of the independent object it represents. The element in the i-th row and j-th column of the relationship matrix represents the association between the i-th unit of the data matrix and the independent object represented by the j-th unit. If there is an association, the value of this element in the relationship matrix is ​​1, otherwise it is 0; The association between the controllable source node and the controlled node means that: the controllable source node is used to regulate the controlled node; The association between controllable source nodes means that the controllable source nodes are located in the same power quality comprehensive management area; The power grid as a whole is associated with all controllable source nodes and controlled nodes; The association between controlled nodes means that the controlled nodes are located in the same power quality comprehensive management area; A one-dimensional structured data consists of n data items sorted by time, and the t-th data item represents the simulation data of the corresponding independent object collected at the t-th moment; Step 500: Input the two-dimensional structure data into the PVMFGCI strategy model. The PVMFGCI strategy model includes a first intermediate layer, a second intermediate layer, and a first output layer. The first intermediate layer inputs the one-dimensional structure data and outputs first intermediate representation data to the second intermediate layer. The second intermediate layer also inputs the relationship matrix and outputs second intermediate representation data to the first output layer. The first output layer outputs a strategy, which includes the capacity of controllable source nodes that need to be added to all dominant controlled nodes. Step 600: Execute the strategy output in step 500 to adjust the power grid.

2. A method for monitoring and comprehensively evaluating power quality of a distribution network according to claim 1, characterized in that: The strategy is represented as a vector, one component of which represents the per-unit value of the capacity of the controllable source node that needs to be added to a dominant controlled node.

3. A method for monitoring and comprehensively evaluating power quality of a distribution network according to claim 1, characterized in that: The calculation formula for the first intermediate layer is as follows: ； ； ； ； in, 、 、 represents the weight parameter, 、 、 represents the bias parameter, represents the dot product, 、 and Represent the first, second, and third intermediate states respectively, where Represents the tth data item of one-dimensional structured data, and Respectively represent the t-th and t-1-th intermediate representation data, n≥t≥1, n represents the total number of data items of the one-dimensional structure data, when t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function; The calculation formula for the second intermediate layer is as follows: ； in The second intermediate representation data representing the vth cell of the data matrix, is the set of cells associated with the vth cell of the data matrix, The data matrix The nth first intermediate representation data output when the one-dimensional structure data of the units is input into the first intermediate layer, is a normalization constant, the default value is The total number of cells in is the sigmoid function, Represents the third weight parameter.

4. A method for monitoring and comprehensively evaluating power quality of a distribution network according to claim 3, characterized in that: The calculation formula of the first output layer is as follows: ； in Represents the first output vector, whose cth component value represents the probability value of the cth strategy. The strategy with the largest probability value is selected as the output. The strategy group contains all executable strategies. represents the set of all cells of the data matrix, represents the output weight parameter, Represents the output bias parameter.

5. A method for monitoring and comprehensively evaluating power quality of a distribution network according to claim 1, characterized in that: Using a deep learning model to obtain the control parameters of an existing photovoltaic grid-connected inverter simulation includes the following steps: The simulated power quality management data is processed into sequence data, where a sequence unit of the sequence data represents the simulated power quality management data collected once. The sequence data is input into the deep learning model, and the output represents the result of the simulated control parameters of the photovoltaic grid-connected inverter.

6. A method for monitoring and comprehensively evaluating power quality of a distribution network according to claim 5, characterized in that: The calculation formula of the deep learning model is as follows: ； in represents the t-th intermediate feature, represents the t-1th first intermediate representation data, represents the tth sequence unit of the sequence data, and is the weight parameter, is the bias parameter, tanh is the hyperbolic tangent function; ； in represents the first output vector, a component value of which represents a control parameter of a simulation of a photovoltaic grid-connected inverter, represents the last intermediate feature, is the weight parameter, is the bias parameter, Represents the sigmoid function.

7. A method for monitoring and comprehensively evaluating power quality of a distribution network according to claim 6, characterized in that: Training a deep learning model involves the following steps: Step 101, initializing the parameters of the deep learning model; Step 102: Observe the simulated power quality management data at time t , the strategy executed at time t , Simulated power quality management data at time t+1 , execution strategy Rewards received ; Step 103, then calculate the policy error: ； in represents the policy error at time t, represents the discount factor, , Represents the deep learning model input The maximum probability value in the first output vector output when , Represents the deep learning model input The first output vector outputted when the strategy The probability value of ； Where ΔTHD is the reduction in total harmonic distortion from time t to time t+1; is the reduction of the voltage unbalance rate from time t to time t+1; is the reduction in reactive power loss from time t to time t+1; The capacity ratio of the photovoltaic grid-connected inverters that changed their use from time t to time t+1; 、 、 is the weight coefficient of the three power quality indicators, which are assigned different weights according to their importance. λ is the switching penalty coefficient, which controls the loss caused by the change of capacity usage of the photovoltaic grid-connected inverter. Step 104: Update the deep learning model. The updated formula is as follows: ； , represents the step size of deep learning, Indicates delivery update; Step 105, iterate steps 102-104 until the deep learning model converges or the number of iterations reaches the set value.

8. A distribution network power quality monitoring and comprehensive evaluation system, characterized in that: It is used to execute a distribution network power quality monitoring and comprehensive evaluation method according to any one of claims 1 to 7, comprising: The first module collects simulated power quality management data, which includes the operating data of existing photovoltaic grid-connected inverters, the operating data of the entire power grid, and the operating data of each node in the power grid. The overall operating data of the power grid includes load current, harmonic components, reactive components, and unbalanced current components. The second module obtains simulated control parameters of the existing photovoltaic grid-connected inverter based on the simulated power quality control data processing; the control parameters include harmonic control, reactive power compensation, and three-phase unbalance control coefficient of the photovoltaic grid-connected inverter; The third module generates simulation data, which includes simulated power quality management data collected at a fixed period within a set time period and corresponding simulated control parameters; The fourth module generates two-dimensional structural data based on the simulation data. The two-dimensional structural data includes a data matrix and a relationship matrix. A unit of the data matrix represents the one-dimensional structural data of an independent object. The independent objects include controllable source nodes, controlled nodes, and the entire power grid. A unit of the data matrix only contains the simulation data of the independent object it represents. The fifth module inputs the two-dimensional structure data into the PVMFGCI strategy model. The PVMFGCI strategy model includes a first intermediate layer, a second intermediate layer and a first output layer. The first intermediate layer inputs the one-dimensional structure data and outputs the first intermediate representation data to the second intermediate layer. The second intermediate layer also inputs the relationship matrix and outputs the second intermediate representation data to the first output layer. The first output layer outputs the strategy, which includes the capacity of the controllable source nodes that need to be added to all the dominant controlled nodes.

9. A computer storage medium, characterized in that It is used to store computer-readable instructions, and when the computer-readable instructions are read, it can run a distribution network power quality monitoring and comprehensive evaluation method as described in any one of claims 1-7.

Images