Integrated energy efficiency evaluation method for active distribution network including distributed photovoltaic

Abstract

The present invention discloses an integrated energy efficiency evaluation method for active distribution networks containing high-penetration distributed photovoltaics based on an adversarial interpretation structural model. A corresponding index system is established by comprehensively considering the four aspects of power supply side, grid side, load side, and energy storage side. The index system covers the additional losses caused by power backfeed to the distribution network when the local photovoltaic absorption rate is low. According to the data obtained by the "intelligent online monitoring and control unit", the Hasse matrix based on indicator distance, sample distance, and sample progress is solved, and four directed topological hierarchical graphs are drawn. The energy efficiency level ranking of each area is gradually determined by continuous approximation from multiple dimensions. The energy efficiency level ranking is obtained according to the four directed topological hierarchical graphs, and the accurate energy efficiency level ranking of the area to be evaluated is finally determined.

Description

Technical Field

[0001] The present invention belongs to the field of energy efficiency evaluation of distributed photovoltaic active distribution networks in power systems, and specifically provides an integrated energy efficiency evaluation method for active distribution networks containing distributed photovoltaics. Background Art

[0002] With the growing global demand for sustainable energy and resources, active distribution networks, including distributed photovoltaics, are attracting widespread attention and research in the energy industry as an emerging power system. Characterized by the integration of distributed energy resources, intelligent control, and flexible operation, they hold significant development potential in the renewable energy sector. However, the capacity of solar photovoltaic systems is affected by weather conditions. Energy supply can be unstable, particularly during rainy weather or at night, leading to reduced energy efficiency. Furthermore, low local utilization rates of distributed photovoltaics can cause power to be fed back to the higher-level grid, incurring additional losses in the distribution network.

[0003] Therefore, energy efficiency evaluation of active distribution networks with high penetration of distributed photovoltaic power generation has become an important research area. Energy efficiency evaluation of distribution networks can help assess and improve the efficiency of distribution systems. It can also help identify potential problems and bottlenecks in distributed photovoltaic distribution networks, such as excessive energy volatility, power balance issues, and voltage stability. Evaluation results can identify weaknesses in system operation and propose corresponding improvement plans, thereby optimizing system performance, reducing energy loss and waste, and achieving sustainable energy utilization and effective resource management. Furthermore, the successful operation of distributed photovoltaic distribution networks requires the active participation and support of users. Energy efficiency evaluation can provide users with detailed information on energy consumption and usage, enhancing their understanding of energy consumption. Publicly available and transparent energy efficiency evaluation results can encourage users to adopt more energy-efficient behaviors and promote changes in energy consumption habits.

[0004] Currently, traditional energy efficiency evaluation methods often rely on empirical formulas and statistical methods, relying on the subjective judgment and experience of domain experts in determining weights and setting parameters. This results in highly subjective evaluation results, with different experts potentially producing different results, lacking objectivity and consistency. Traditional energy efficiency evaluation methods often require manual feature selection and parameter setting, leading to subjectivity and limitations in the evaluation results. Energy efficiency evaluation methods based on adversarial explanatory structural models, however, leverage the features and relationships in historical data to avoid subjective factors and better address data uncertainty and complexity. Summary of the Invention

[0005] The present invention provides an integrated energy efficiency evaluation method for an active distribution network including distributed photovoltaics. The technical solution specifically includes the following steps:

[0006] Step 1: Establish an integrated energy efficiency evaluation index system for active distribution networks including distributed photovoltaics. The indicators in this system include:

[0007] First-level indicator: energy efficiency level of active distribution network;

[0008] Secondary indicators: power supply side, grid side, load side, and energy storage side;

[0009] Level 3 indicators:

[0010] The photovoltaic local consumption rate (λ1) and the surplus power grid access rate (λ2) on the power supply side;

[0011] On the grid side, the distribution network bus voltage qualification rate (λ3), voltage harmonic distortion rate (λ4), AC side power factor (λ5), transformer loss (λ6), line economic operation rate (λ7), comprehensive line loss rate (λ8), and DC power supply capacity ratio (λ9);

[0012] Inverter efficiency on the load side (λ10), three-phase load imbalance (λ11), and daily load factor (λ12);

[0013] Energy storage utilization efficiency (λ13), energy storage charge and discharge efficiency (λ14), and energy storage loss rate (λ15) on the energy storage side;

[0014] Clarify the specific calculation formula for each third-level indicator;

[0015] Step 2: Raw data preprocessing

[0016] Based on the normalized energy efficiency evaluation index system for active distribution networks containing distributed photovoltaics, the required index data is obtained using the index calculation formula on the "intelligent online monitoring and control unit" and the phasor measurement unit to generate a raw data matrix; the raw data matrix is ​​normalized; the weights of all indicators are calculated using the coefficient of variation method; and the positive and negative ideal solution sets of the normalized matrix are solved;

[0017] Step 3: Solve the Hasse matrix based on indicator distance and sample distance

[0018] Generate a set of adversarial matrix pairs based on the distance from the normalized matrix to the positive and negative ideal solutions; arrange them in ascending and descending order according to the size of the indicator weights to obtain two sets of adversarial matrix pairs; linearly accumulate these two sets of adversarial matrix pairs to obtain two sets of adversarial matrix pairs; solve the relationship matrix of the two sets of adversarial matrices based on the partial order rule; calculate the reachability matrix based on the relationship matrix, and further obtain the Hasse matrix based on the indicator distance;

[0019] Calculate the sample distances from the evaluation object to the positive and negative ideal points; calculate the progress of the evaluation object based on the sample distances from the evaluation object to the positive and negative ideal solutions; repeatedly solve the Hasse matrix based on the indicator distance; obtain the Hasse matrix based on the sample distance and the Hasse matrix based on the progress of the evaluation object;

[0020] Calculate the distance and closeness between the evaluation object and the positive and negative ideal points; repeat steps 2 and 3 to obtain the Haas matrix based on the distance and closeness between the evaluation object and the positive and negative ideal points;

[0021] Step 4: Draw four directed topological hierarchical graphs;

[0022] Step 5: Ranking the energy efficiency levels of the evaluation objects in the four directed topological hierarchical graphs to finally obtain the accurate ranking of the evaluation objects.

[0023] Beneficial effects:

[0024] Based on some existing indicator systems, the present invention establishes an integrated indicator evaluation system from four aspects: power supply side, grid side, load side, and energy storage side, and comprehensively evaluates the energy efficiency of the distribution network. Considering the high-penetration distributed photovoltaic scenario, when the photovoltaic power generation is greater than the load demand and the local photovoltaic absorption rate is low, the power backflow has an impact on the upper distribution network and the additional losses brought to the distribution network, the adversarial explanatory structural model is used to evaluate the energy efficiency of the active distribution network containing multiple influencing factors. Compared with the traditional explanatory structural model, this method evaluates the energy efficiency of the distribution network from multiple dimensions, and the results obtained are more accurate and reasonable, providing a reference for the scheduling, operation, energy saving and planning of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 , energy efficiency evaluation flow chart;

[0026] Figure 2 ,Flowchart of the adversarial interpretation structural model;

[0027] Figure 3 , distribution network energy efficiency evaluation results based on the indicator distances Pu and Qu;

[0028] Figure 4 , distribution network energy efficiency evaluation results based on indicator distance Pd and Qd;

[0029] Figure 5 , distribution network energy efficiency evaluation results based on the distance to the evaluation object;

[0030] Figure 6 , distribution network energy efficiency evaluation results based on the progress of the evaluation object. DETAILED DESCRIPTION

[0031] The following is combined with Figure 1-6, Examples further illustrate the technical solution of the present invention in detail.

[0032] The present invention discloses an integrated energy efficiency evaluation method for an active distribution network including distributed photovoltaics, comprising the following steps:

[0033] Step 1: Establish an integrated energy efficiency evaluation index system for active distribution networks including distributed photovoltaics. The indicators in this system include:

[0034] First-level indicator: energy efficiency level of active distribution network;

[0035] Secondary indicators: power supply side, grid side, load side, and energy storage side;

[0036] Level 3 indicators:

[0037] The photovoltaic local consumption rate (λ1) and the surplus power grid access rate (λ2) on the power supply side;

[0038] On the grid side, the distribution network bus voltage qualification rate (λ3), voltage harmonic distortion rate (λ4), AC side power factor (λ5), transformer loss (λ6), line economic operation rate (λ7), comprehensive line loss rate (λ8), and DC power supply capacity ratio (λ9);

[0039] Inverter efficiency on the load side (λ10), three-phase load imbalance (λ11), and daily load factor (λ12);

[0040] Energy storage utilization efficiency (λ13), energy storage charge and discharge efficiency (λ14), and energy storage loss rate (λ15) on the energy storage side;

[0041] The specific calculation formula for each third-level indicator is as follows:

[0042] Power supply side

[0043] Photovoltaic local consumption rate

[0044] The photovoltaic local consumption rate is one of the decisive factors for the scale of new photovoltaic power generation, reflecting the dispatching and operation capacity of the distribution system. The photovoltaic local consumption rate is determined by the proportion of the self-generated and self-consumed electricity of the photovoltaic power station to the total power generation of the photovoltaic power station.

[0045]

[0046] Where P self-use is the self-generated and self-used electricity of the photovoltaic power station, P PV is the total power generation of the photovoltaic power station;

[0047] Surplus power grid access rate

[0048] The surplus power grid-connected rate reflects the degree of impact of distributed photovoltaics on the upper power grid, and also reflects the amount of additional losses caused to the distribution network when the local photovoltaic consumption rate is low. The surplus power grid-connected rate is calculated by subtracting the local photovoltaic consumption rate and the ratio of the power loss during local consumption to the total power generation of the photovoltaic power station.

[0049]

[0050] Where P loss The amount of electricity lost when consuming electricity nearby;

[0051] Grid side

[0052] Distribution network bus voltage qualification rate

[0053] The bus voltage of the distribution network directly affects the network loss of the entire distribution system. When the bus voltage decreases, the variable loss of the distribution system increases. When the bus voltage increases, the fixed loss of the distribution system increases. When distributed photovoltaics are connected to the distribution system, the voltage of the node may exceed the limit, resulting in an increase in the fixed loss of the distribution system. The qualified rate of the distribution network bus voltage is characterized by the proportion of the working time of the bus voltage within the specified range in a certain continuous detection time period to the entire detection period. The expression is:

[0054]

[0055] Voltage harmonic distortion rate

[0056] According to the "Technical Guidelines for Rural Power Grid Construction and Reconstruction", the qualified rate of total harmonic distortion of low-voltage distribution network λ4 is based on the low-voltage distribution system meeting the 380V voltage total harmonic distortion rate THD u The ratio of the number of areas with ≤5.0% to the total number of areas is used to determine the proportion.

[0057]

[0058] Where U H is the harmonic voltage content, in kV; U1 is the fundamental voltage (RMS value), in kV; U h is the hth harmonic voltage (RMS value), in kV;

[0059] AC side power factor

[0060] Low power factor will reduce the energy efficiency of the distribution network and affect the power quality.

[0061]

[0062] Where P is active power; S is apparent power,

[0063] Transformer losses

[0064]

[0065] Where P H is the power on the high-voltage side of the distribution transformer, in kW; P L The power on the low-voltage side of the distribution transformer, in kW;

[0066] Line economic operation rate

[0067]

[0068] Where t line-i represents the economic operation time of line i; λi represents the weight coefficient of line i, which is used to distinguish its importance; T represents the statistical time. The line load rate range of the economic operation of the line in this paper is [50%, 70%].

[0069] Comprehensive line loss rate

[0070]

[0071] Where P1 represents the total amount of power supplied to the distribution network, P PV Represents the power generation of distributed photovoltaics, P self-use Indicates the on-site consumption of photovoltaic power; P up P2 represents the amount of electricity generated by photovoltaic power plants; P2 represents the amount of electricity sold by the power supply company;

[0072] DC power supply capacity ratio

[0073] When distributing power at the same power, the power loss of DC distribution is lower than that of AC distribution. The proportion of DC power supply capacity represents the impact of DC power flow on the energy efficiency of the distribution network.

[0074]

[0075] Where P DC Represents DC power supply capacity; P Load Represents the load capacity of the distribution network;

[0076] Load side

[0077] Inverter efficiency

[0078]

[0079] Where N is the number of inverter types; n is the number of inverters of each type; p i is the capacity proportion of the i-th type inverter; η ij is the conversion efficiency of the j-th inverter of the i-th category; N' is the total number of inverters,

[0080] Three-phase load imbalance

[0081]

[0082] Where I max is the typical maximum phase current of the year; I min is the typical minimum phase current of the year;

[0083] Daily load rate

[0084]

[0085] Where P av_load represents the average load of the user; P max_load Represents the maximum load of the user;

[0086] Energy storage side

[0087] Energy storage utilization efficiency

[0088] Energy storage utilization efficiency is calculated by converting the actual amount of electricity transmitted (including charging and discharging) during the statistical period into the number of operating hours when the rated power is used.

[0089]

[0090] Where E cha is the charging amount during the statistical time period, E dis is the discharge amount in the statistical period, P r is the rated power of the energy storage power station;

[0091] Energy storage charging and discharging efficiency

[0092]

[0093] Where E cha E is the charge amount of the energy storage device during the statistical time period; dis The discharge amount of the energy storage device during the statistical time period;

[0094] Energy storage loss rate

[0095] It is determined by the ratio of the difference between the on-grid power and off-grid power during the production and operation of the energy storage power station to the total off-grid power grid during the statistical period.

[0096]

[0097] Where P up is the amount of electricity that is online during the statistical period of the energy storage power station, P down The amount of electricity disconnected from the energy storage power station during the statistical period;

[0098] Step 2: Raw data preprocessing

[0099] Based on the normalized energy efficiency evaluation index system for active distribution networks containing distributed photovoltaics, the required index data is obtained using the index calculation formula on the "intelligent online monitoring and control unit" and the "phasor measurement unit" to generate a raw data matrix; the raw data matrix is ​​normalized; the weights of all indicators are calculated using the coefficient of variation method; and the positive and negative ideal solution sets of the normalized matrix are solved;

[0100] Specifically: According to the attributes of the original data, the indicators are divided into two categories: positive indicators and negative indicators, and the range method is used to normalize the original data matrix:

[0101] Positive indicators

[0102] Negative indicators

[0103] Where z ij ∈Z, Z represents the matrix composed of original data, n ij ∈N, N represents the normalized matrix,

[0104] After the original data matrix is ​​normalized, the positive and negative ideal point sets of the normalized matrix are solved. The positive ideal point set is recorded as S+ and the negative ideal point set is recorded as S-, which can be expressed as:

[0105]

[0106] Step 3: Solve the Hasse matrix based on indicator distance and sample distance

[0107] Generate a set of adversarial matrix pairs based on the distance from the normalized matrix to the positive and negative ideal solutions; arrange them in ascending and descending order according to the size of the indicator weights to obtain two sets of adversarial matrix pairs; linearly accumulate these two sets of adversarial matrix pairs to obtain two sets of adversarial matrix pairs; solve the relationship matrix of the two sets of adversarial matrices based on the partial order rule; calculate the reachability matrix based on the relationship matrix, and further obtain the Hasse matrix based on the indicator distance;

[0108] Calculate the sample distances from the evaluation object to the positive and negative ideal points; calculate the progress of the evaluation object based on the sample distances from the evaluation object to the positive and negative ideal solutions; repeatedly solve the Hasse matrix based on the indicator distance; obtain the Hasse matrix based on the sample distance and the Hasse matrix based on the progress of the evaluation object;

[0109] Specifically: Based on the distance from the normalized matrix to the positive and negative points, a pair of adversarial matrices K+ and K- with opposite properties are generated. The weight of each indicator is calculated using the coefficient of variation method. K+ and K- are rearranged in ascending and descending order according to the proportion of the indicator weight, respectively, to obtain two sets of adversarial matrices, which are recorded as and Linearly add the components of each column of the four matrices to obtain matrices Pa, Pd and Qa, Qd, respectively, where Pa and Qa are a pair of adversarial matrices, and Pd and Qd are a pair of adversarial matrices;

[0110] First, take the partial order rule for matrices P and Q respectively to obtain the relationship matrix A. The calculation formula is as follows:

[0111]

[0112] Where x ij ,y ij are the elements of any two rows in the P and Q matrices;

[0113] For the relationship matrix A, the calculation formula of its reachability matrix is ​​as follows:

[0114] B=A+I

[0115] Where B is the multiplication matrix; A is the relationship matrix; I is the identity matrix;

[0116] Use Boolean algebra to multiply B until

[0117] B k-1 ≠B k =B k+1 =R

[0118] Where R is the reachability matrix;

[0119] The calculation formula from the reachability matrix R to the skeleton matrix is ​​as follows:

[0120] S=HS=R-(RI) 2 -I

[0121] Where S is the skeleton matrix and HS is the Hasse matrix;

[0122] Evaluate the distance and closeness of the object to the positive and negative ideal points,

[0123] The calculation formula for the evaluation object to the positive and negative ideal points d+ and d- is as follows:

[0124]

[0125] Where ω j is the weight of the j-th indicator, n ij ∈N,

[0126] The closeness of the evaluation object is determined by d + , d - Calculated, γ + Indicates the closeness between the evaluation object and the negative ideal point, γ -Indicates the closeness between the evaluation object and the positive ideal point. The calculation formula is as follows:

[0127]

[0128] Calculate the distance and closeness between the evaluation object and the positive and negative ideal points; repeat steps 2 and 3 to obtain the Haas matrix based on the distance and closeness between the evaluation object and the positive and negative ideal points;

[0129] Step 4: Draw four directed topological hierarchical graphs

[0130] First, for Boolean square matrices, there are reachable set R, precedent set Q, and common set T, where T = R∩Q. Taking the relation matrix A as an example, its element e1,

[0131] The reachable set of e1 is denoted as R(e1), which is all elements whose corresponding row value is 1;

[0132] The predecessor set of e1 is denoted as Q(e1), which is all elements whose corresponding column value is 1;

[0133] The common set of e1 is T(e1), that is, R(e1)∩Q(e1);

[0134] Secondly, four extraction rules for drawing directed topological hierarchical graphs are established;

[0135] specific:

[0136] UP type hierarchy diagram: T(ei)=R(ei), placed from top to bottom, results take priority;

[0137] DOWN type hierarchy diagram: T(ei)=Q(ei), placed from bottom to top, cause priority;

[0138] UD type hierarchy diagram: first T(ei)=R(ei), then T(ei)=Q(ei), placed alternately at the top and bottom of the hierarchy diagram;

[0139] DU type hierarchy diagram: first T(ei)=Q(ei), then T(ei)=R(ei), placed alternately at the bottom and top of the hierarchy diagram;

[0140] Then, four directed topological hierarchical graphs are obtained according to the Hasse matrix based on indicator distance, sample distance, and evaluation object progress.

[0141] Step 5: Ranking the energy efficiency levels of the evaluation objects in the four directed topological hierarchical graphs to finally obtain the accurate ranking of the evaluation objects.

[0142] Examples

[0143] To verify the energy efficiency evaluation method proposed in this invention for a distributed photovoltaic active distribution network with high penetration, the following experiments were conducted to verify the method by obtaining the weights of the indicators based on historical data in the face of complex distribution systems and low local consumption rates of distributed photovoltaics, so as to avoid the influence of subjective factors on the objectivity of the energy efficiency evaluation results.

[0144] Eight typical medium and low voltage substations with high penetration rates of distributed photovoltaics in Rizhao City, Shandong Province were selected for analysis. Table 2 shows the penetration rates of the eight areas.

[0145] Table 2. PV penetration rates in eight regions

[0146] Area Number Area 1 Area 2 Area 3 Area 4 Area 5 Area 6 Area 7 Area 8 Penetration rate / % 127.6％ 112.7％ 89.6％ 85.1％ 92.4％ 79.7％ 83.4％ 88.35％

[0147] The indicator data in the index system are all derived from the real-time data of the "intelligent online monitoring and control unit" in Rizhao City, Shandong Province. The original data of the eight regions are evaluated according to the energy efficiency evaluation method for the active distribution network with high penetration rate distributed photovoltaics described in this invention. The distribution networks of the 10 selected regions are numbered according to 8 respectively. The original data matrix of the eight regions is shown in Table 3.

[0148] Table 3. Raw energy efficiency data of eight regional distribution networks

[0149] <![CDATA[λ1]]> <![CDATA[λ2]]> <![CDATA[λ3]]> <![CDATA[λ4]]> <![CDATA[λ5]]> <![CDATA[λ6]]> <![CDATA[λ7]]> <![CDATA[λ8]]> <![CDATA[λ9]]> <![CDATA[λ 10 ]]> <![CDATA[λ 11 ]]> <![CDATA[λ 12 ]]> <![CDATA[λ 13 ]]> <![CDATA[λ 14 ]]> <![CDATA[λ 15 ]]> Area 1 0.460 0.412 0.70 0.080 0.800 0.050 0.650 0.077 0.287 0.890 0.910 0.424 0.068 0.830 0.180 Area 2 0.330 0.520 0.760 0.090 0.700 0.067 0.470 0.080 0.309 0.870 0.930 0.417 0.084 0.810 0.210 Area 3 0.430 0.450 0.790 0.063 0.800 0.052 0.550 0.062 0.198 0.850 0.965 0.430 0.062 0.790 0.130 Area 4 0.450 0.430 0.830 0.065 0.790 0.045 0.600 0.073 0.453 0.960 0.970 0.349 0.106 0.860 0.190 Area 5 0.530 0.370 0.860 0.070 0.820 0.035 0.650 0.053 0.264 0.930 1.000 0.370 0.073 0.850 0.150 Area 6 0.590 0.320 0.890 0.060 0.850 0.032 0.760 0.045 0.154 0.880 0.990 0.389 0.060 0.840 0.140 Area 7 0.500 0.390 0.800 0.030 0.820 0.039 0.680 0.059 0.312 0.840 0.980 0.514 0.077 0.850 0.200 Area 8 0.650 0.280 0.890 0.035 0.890 0.028 0.850 0.034 0.365 0.830 1.000 0.434 0.090 0.790 0.200

[0150] According to the steps in the specific implementation plan 3, the directed topological hierarchy diagram of the regional distribution network energy efficiency ranking based on the indicator distance, the directed topological hierarchy diagram of the regional distribution network energy efficiency ranking based on the evaluation object distance, and the directed topological hierarchy diagram of the regional distribution network energy efficiency ranking based on the evaluation object scheme progress are obtained respectively. Figure 3-Figure 6 ,Table 4 shows the data of the eight regions based on the evaluation objects and ,scheme posting progress.

[0151] Table 4. Evaluation object distance and solution posting progress

[0152]

[0153] It can be seen that the energy efficiency levels among the eight regions are continuously approximated by the directed topological hierarchical graph based on the indicator distance, the evaluation object distance and the program progress, and finally the accurate energy efficiency ranking among the eight regions is obtained. The accurate ranking of energy efficiency among the eight regions cannot be obtained only from the directed topological hierarchical graph of the indicator distance or the evaluation object distance. The energy efficiency ranking of the eight regional distribution networks is Figure 6 shown.

[0154] To quantify the energy efficiency ranking of the distribution network, the energy efficiency level is divided into four levels based on the degree of closeness γ- (the smaller the γ-, the better, indicating a closer fit to the positive ideal solution and higher energy efficiency) between the evaluation area and the positive ideal point set. Table 5 shows the quantified energy efficiency level evaluation table.

[0155] Table 5. Energy efficiency level evaluation table

[0156]

[0157] The γ-values ​​of the eight regions are calculated according to the following formula, and the energy efficiency level evaluation results of the eight regions are shown in Table 6.

[0158]

[0159] Table 6. Regional distribution network energy efficiency evaluation results

[0160] Area Number Energy efficiency evaluation 4、8、2 good 1、3、6、5、7 generally

[0161] The present invention discloses an integrated energy efficiency evaluation method based on an adversarial explanatory structural model, suitable for active distribution networks with high penetration rates of distributed photovoltaics. This innovative technology combines adversarial learning with an explanatory structural model to accurately and efficiently evaluate the energy efficiency of distribution networks. The adversarial explanatory structural model integrates adversarial thinking with traditional explanatory structural model methods and is suitable for energy efficiency evaluation of nonlinear systems with multiple factors, multiple indicators, and multi-level coupling. The energy efficiency ranking of the evaluation object is ultimately determined through layer-by-layer approximation of indicator distance, sample distance, and the proximity of the evaluation object. Furthermore, the energy efficiency index system established in traditional energy efficiency evaluation methods is not suitable for active distribution networks with a large amount of distributed energy. This invention establishes an energy efficiency evaluation system suitable for high-penetration distributed photovoltaic access distribution networks by focusing on four aspects of the distribution network: the power supply side, the grid side, the load side, and the energy storage side. Based on the shared data of the "intelligent online monitoring and control unit," data from the energy efficiency evaluation index system is selected for energy efficiency measurement. Finally, the energy efficiency evaluation method proposed in this invention is used for evaluation, providing a reference for energy conservation and optimized dispatching operations of the distribution network.

Claims

1. Integrated energy efficiency evaluation method for active distribution network including distributed photovoltaics, It is characterized in that The following steps are involved: Step 1: Establish an integrated energy efficiency evaluation index system for active distribution networks including distributed photovoltaics. The indicators in this system include: First-level indicator: energy efficiency level of active distribution network; Secondary indicators: power supply side, grid side, load side, and energy storage side; Level 3 indicators: On the power supply side, the photovoltaic local consumption rate λ1 and the surplus power grid rate λ2; On the grid side, the distribution network bus voltage qualification rate λ3, voltage harmonic distortion rate λ4, AC side power factor λ5, transformer loss λ6, line economic operation rate λ7, comprehensive line loss rate λ8, and DC power supply capacity ratio λ9; Inverter efficiency λ10, three-phase load imbalance λ11, and daily load rate λ12 on the load side; Energy storage utilization efficiency λ13, energy storage charge and discharge efficiency λ14, and energy storage loss rate λ15 on the energy storage side; The specific calculation formula for each third-level indicator is as follows: Power supply side Photovoltaic local consumption rate The photovoltaic local consumption rate is one of the decisive factors for the scale of new photovoltaic power generation, reflecting the dispatching and operation capacity of the distribution system. The photovoltaic local consumption rate is determined by the proportion of the self-generated and self-consumed electricity of the photovoltaic power station to the total power generation of the photovoltaic power station. Where P self-use is the self-generated and self-used electricity of the photovoltaic power station, P PV is the total power generation of the photovoltaic power station; Surplus power grid access rate The surplus power grid-connected rate reflects the degree of impact of distributed photovoltaics on the upper power grid, and also reflects the amount of additional losses caused to the distribution network when the local photovoltaic consumption rate is low. The surplus power grid-connected rate is calculated by subtracting the local photovoltaic consumption rate and the ratio of the power loss during local consumption to the total power generation of the photovoltaic power station. Where P loss The amount of electricity lost when consuming electricity nearby; Grid side Distribution network bus voltage qualification rate The bus voltage of the distribution network directly affects the network loss of the entire distribution system. When the bus voltage decreases, the variable loss of the distribution system increases. When the bus voltage increases, the fixed loss of the distribution system increases. When distributed photovoltaics are connected to the distribution system, the voltage of the node may exceed the limit, resulting in an increase in the fixed loss of the distribution system. The qualified rate of the distribution network bus voltage is characterized by the proportion of the working time of the bus voltage within the specified range in a certain continuous detection time period to the entire detection period. The expression is: Voltage harmonic distortion rate According to the "Technical Guidelines for Rural Power Grid Construction and Reconstruction", the qualified rate of total harmonic distortion of low-voltage distribution network λ4 is based on the low-voltage distribution system meeting the 380V voltage total harmonic distortion rate THD u The ratio of the number of areas with ≤5.0% to the total number of areas is used to determine the proportion. Where U H is the harmonic voltage content, in kV; U1 is the root mean square value of the fundamental voltage, in kV; U h is the root mean square value of the hth harmonic voltage, in kV; AC side power factor Low power factor will reduce the energy efficiency of the distribution network and affect the power quality. Where P is active power; S is apparent power, Transformer losses Where P H is the power on the high-voltage side of the distribution transformer, in kW; P L The power on the low-voltage side of the distribution transformer, in kW; Line economic operation rate Where t line-i represents the economic operation time of line i; λi represents the weight coefficient of line i, which is used to distinguish its importance; T represents the statistical time. The line load rate range of the economic operation of the line in this paper is [50%, 70%]. Comprehensive line loss rate Where P1 represents the total amount of electricity supplied to the distribution network, P PV Represents the power generation of distributed photovoltaics, P self-use Indicates the on-site consumption of photovoltaic power; P up P2 represents the amount of electricity generated by photovoltaic grid; P2 represents the amount of electricity sold by the power supply company; DC power supply capacity ratio When distributing power at the same power, the power loss of DC distribution is lower than that of AC distribution. The proportion of DC power supply capacity represents the impact of DC power flow on the energy efficiency of the distribution network. Where P DC Represents DC power supply capacity; P Load Represents the load capacity of the distribution network; Load side Inverter efficiency Where N is the number of inverter types; n is the number of inverters of each type; p i is the capacity proportion of the i-th type inverter; η ij is the conversion efficiency of the j-th inverter of the i-th category; N' is the total number of inverters, Three-phase load imbalance Where I max is the typical maximum phase current of the year; I min is the typical minimum phase current of the year; Daily load rate Where P av_load represents the average load of the user; P max_load Represents the maximum load of the user; Energy storage side Energy storage utilization efficiency Energy storage utilization efficiency is calculated by converting the actual amount of electricity transmitted during the statistical period, including charging and discharging, into the number of operating hours at rated power. Where E cha is the charging amount during the statistical time period, E dis is the discharge amount in the statistical period, P r is the rated power of the energy storage power station; Energy storage charging and discharging efficiency Where E cha E is the charge amount of the energy storage device during the statistical time period; dis The discharge amount of the energy storage device during the statistical time period; Energy storage loss rate It is determined by the ratio of the difference between the on-grid power and off-grid power during the production and operation of the energy storage power station to the total off-grid power grid during the statistical period. Where P up is the amount of electricity that is online during the statistical period of the energy storage power station, P down The amount of electricity disconnected from the energy storage power station during the statistical period; Step 2: Raw data preprocessing Based on the normalized energy efficiency evaluation index system for active distribution networks including distributed photovoltaics, the "intelligent online monitoring and control unit" and phasor measurement unit are used to obtain the required index data according to the index calculation formula to generate a raw data matrix; the raw data matrix is ​​normalized; the coefficient of variation method is used to calculate the weights of all indicators; and the positive and negative ideal solution sets of the normalized matrix are solved; Specifically: According to the attributes of the original data, the indicators are divided into two categories: positive indicators and negative indicators, and the range method is used to normalize the original data matrix: Positive indicators In the negative indicator formula, z ij ∈Z, Z represents the matrix composed of original data, n ij ∈N, N represents the normalized matrix, After the original data matrix is ​​normalized, the positive and negative ideal point sets of the normalized matrix are solved. The positive ideal point set is recorded as S+ and the negative ideal point set is recorded as S-, which can be expressed as: Step 3: Solve the Hasse matrix based on indicator distance and sample distance Generate a set of adversarial matrix pairs based on the distance from the normalized matrix to the positive and negative ideal solutions; arrange them in ascending and descending order according to the size of the indicator weights to obtain two sets of adversarial matrix pairs; linearly accumulate these two sets of adversarial matrix pairs to obtain two sets of adversarial matrix pairs; solve the relationship matrix of the two sets of adversarial matrices based on the partial order rule; calculate the reachability matrix based on the relationship matrix, and further obtain the Hasse matrix based on the indicator distance; Calculate the sample distances from the evaluation object to the positive and negative ideal points; calculate the progress of the evaluation object based on the sample distances from the evaluation object to the positive and negative ideal solutions; repeatedly solve the Hasse matrix based on the indicator distance; obtain the Hasse matrix based on the sample distance and the Hasse matrix based on the progress of the evaluation object; Specifically: Based on the distance from the normalized matrix to the positive and negative points, a pair of adversarial matrices K+ and K- with opposite properties are generated. The weight of each indicator is calculated using the coefficient of variation method. K+ and K- are rearranged in ascending and descending order according to the proportion of the indicator weight, respectively, to obtain two sets of adversarial matrices, which are recorded as and Linearly add the components of each column of the four matrices to obtain the matrix P a , P d and Q a , Q d , where P a and Q a is a pair of adversarial matrices, P d and Q d , is a pair of adversarial matrices; First, take the partial order rule for matrices P and Q respectively to obtain the relationship matrix A. The calculation formula is as follows: Where x ij ,y ij are the elements of any two rows in the P and Q matrices; For the relationship matrix A, the calculation formula of its reachability matrix is ​​as follows: B=A+I Where B is the multiplication matrix; A is the relationship matrix; I is the identity matrix; Use Boolean algebra to multiply B until B k-1 ≠B k =B k+1 =R Where R is the reachability matrix; The calculation formula from the reachability matrix R to the skeleton matrix is ​​as follows: S=HS=R-(RI) 2 -I Where S is the skeleton matrix and HS is the Hasse matrix; Evaluate the distance and proximity of the object to positive and negative ideal points The calculation formula for the evaluation object to the positive and negative ideal points d+ and d- is as follows: Where ω j is the weight of the j-th indicator, n ij ∈N, The closeness of the evaluation object is determined by d + , d - Calculated, γ + Indicates the closeness between the evaluation object and the negative ideal point, γ - Indicates the closeness between the evaluation object and the positive ideal point. The calculation formula is as follows: Calculate the distance and closeness between the evaluation object and the positive and negative ideal points; repeat steps 2 and 3 to obtain the Haas matrix based on the distance and closeness between the evaluation object and the positive and negative ideal points; Step 4: Draw four directed topological hierarchical graphs First, for Boolean square matrices, there are reachable set R, precedent set Q, and common set T, where T = R∩Q. Taking the relation matrix A as an example, its element e1, The reachable set of e1 is denoted as R(e1), which is all elements whose corresponding row value is 1; The predecessor set of e1 is denoted as Q(e1), which is all elements whose corresponding column value is 1; The common set of e1 is T(e1), that is, R(e1)∩Q(e1); Secondly, four extraction rules for drawing directed topological hierarchical graphs are established; specific: UP type hierarchy diagram: T(ei)=R(ei), placed from top to bottom, results take priority; DOWN type hierarchy diagram: T(ei)=Q(ei), placed from bottom to top, cause priority; UD type hierarchy diagram: first T(ei)=R(ei), then T(ei)=Q(ei), placed alternately at the top and bottom of the hierarchy diagram; DU type hierarchy diagram: first T(ei)=Q(ei), then T(ei)=R(ei), placed alternately at the bottom and top of the hierarchy diagram; Then, four directed topological hierarchical graphs are obtained according to the Hasse matrix based on indicator distance, sample distance, and evaluation object progress. Step 5: Ranking the energy efficiency levels of the evaluation objects in the four directed topological hierarchical graphs to finally obtain the accurate ranking of the evaluation objects.

Images