Optimal configuration method of distributed photovoltaic microclimate monitoring device considering space-time characteristics

Abstract

The application discloses a distributed photovoltaic microclimate monitoring device optimization configuration method considering space-time characteristics, comprising the following steps: obtaining a comprehensive node criticality index of a distributed photovoltaic system by comprehensively considering electrical and microclimate sensitivities; dividing an initial period, a development period and a mature period for a growth stage of the distributed photovoltaic system based on a Bass diffusion model; constructing different optimization configuration models for different periods, solving the optimization configuration models, and obtaining microclimate monitoring device optimization configuration results; and performing observability index and output prediction calculation on the optimization configuration results, and testing and judging the economy, rationality and accuracy of the optimization configuration results. The method can consider the sensitivity change of electrical and microclimate factors of a distribution network system, and can also dynamically arrange a measurement device according to the change of distributed photovoltaic access, so that the flexibility, accuracy and economy of the monitoring and prediction device planning of the distribution network with high proportion of urban environment distributed photovoltaic access are effectively ensured.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of distributed photovoltaic system monitoring and optimal configuration, and particularly relates to a distributed photovoltaic micro-meteorological monitoring device optimal configuration method considering time-space characteristics. BACKGROUND

[0002] Large-scale roof distributed photovoltaic construction in cities has become a new trend of urban green development, and the development and construction of roof distributed photovoltaic has great potential, which is conducive to promoting the optimization of power distribution network resources and the improvement of green level development. However, roof distributed photovoltaic is sensitive to meteorological changes, and the intermittent output characteristics are present, and the complex geographical and meteorological environment in the city increases the uncertainty of the output. The short-term explosive growth of the number of distributed photovoltaic construction leads to safety problems in system operation. How to realize comprehensive monitoring of the running state of a large number of distributed photovoltaics is the basis of power distribution network situation awareness and safe dispatching control, and is also the key to guarantee the low-carbon safe operation of the urban power distribution network. The existing photovoltaic monitoring equipment cannot meet the comprehensive "observable and measurable" demand of the running state of a large number of distributed photovoltaics, which is specifically manifested in that: photovoltaic power generation lacks independent meteorological monitoring devices and power prediction systems; the traditional configuration method of photovoltaic monitoring is regional grid division, which does not consider the influence of city geography and micro-meteorology; the installation of monitoring devices is limited by time and cost. Therefore, the monitoring device planning problem in line with the large-scale growth trend of distributed photovoltaics needs to be solved.

[0003] At present, there are many studies on the optimization configuration of measurement devices in power distribution networks. The main methods for determining the weak links of power distribution systems are electrical sensitivity analysis, key meteorological environmental factors, observability index, cost-weighted factor, and state change sensitive node determination method; the main methods for solving the configuration problem of measurement devices are clustering analysis method, particle swarm optimization artificial intelligence algorithm, and principal component analysis method to identify and complete measurement configuration. However, the above methods have the problems that it is difficult to consider the influence of electrical and meteorological factors in determining the weak links of the system, the traditional planning configuration method does not consider the overall change of the power distribution network, it is difficult to develop measurement configuration for growing power grids, and it is difficult to coordinate and optimize the observability and investment cost economy. Therefore, in the present application, photovoltaic output monitoring is carried out for the scene of high-proportion roof distributed photovoltaic access, and the sensitivity monitoring of electrical operation state and micro-meteorological change is required. How to comprehensively consider the influence of the two is still lack of corresponding research methods. SUMMARY

[0004] The purpose of the present application is to provide a distributed photovoltaic micro-meteorological monitoring device optimal configuration method considering time-space characteristics, which aims to maximize observability, optimize economy, accurately monitor and predict effects, and has high flexibility.

[0005] Technical scheme: The distributed photovoltaic microclimate monitoring device optimization configuration method considering the space-time characteristics comprises the following steps:

[0006] For the roof distributed photovoltaic system with a high proportion of access in the city, the distribution network parameters and microclimate information data are calculated, the electrical and microclimate sensitivities are comprehensively obtained, and the node comprehensive criticality index of the distributed photovoltaic system is obtained.

[0007] According to the growth influencing factors of the distributed photovoltaic system, the Bass diffusion model is used to divide the growth stages of the distributed photovoltaic system into an initial stage, a development stage and a mature stage; for the initial stage, according to the node comprehensive criticality index, the total weight of the photovoltaic node is maximized, and the economic constraint weight is minimized as the objective function I, and the optimization configuration model of the geographic space is constructed; for the development stage and the mature stage, according to the comprehensive node criticality index, the layout of the microclimate monitoring device is optimized, and the growth expectation of the future photovoltaic access in the time dimension is considered as the objective function II, and the optimization configuration model of the microclimate characteristics is constructed; the optimization configuration model is solved, and the optimization configuration result of the microclimate monitoring device is obtained.

[0008] For the optimization configuration result, the observability index and the output prediction calculation are performed, and the economy, rationality and accuracy of the optimization configuration result are tested and judged.

[0009] Further, the node comprehensive criticality index of the distributed photovoltaic system is obtained by comprehensively considering the electrical and microclimate sensitivities, and specifically:

[0010] The node comprehensive criticality index expression of the distributed photovoltaic system in the system is as follows:

[0011]

[0012] Wherein, K i is the node criticality coefficient corresponding to each distributed photovoltaic node; δ i represents the electrical comprehensive sensitivity index, ρ i represents the microclimate comprehensive sensitivity index, represents the normalized electrical comprehensive sensitivity index, represents the normalized microclimate comprehensive sensitivity index; ω P represents the active weight coefficient; ω Q represents the reactive weight coefficient; s PU represents the active sensitivity factor; s QU represents the reactive sensitivity factor; ω E represents the irradiation coefficient; ω T represents the temperature coefficient; ω PM represents the PM concentration coefficient; E i represents the total radiation sensitivity factor of node i; T iIndicates the ambient temperature sensitivity factor; C PM,i This represents the air quality sensitivity factor.

[0013] Furthermore, the comprehensive sensitivity index of micrometeorology is determined based on the micrometeorological impact index analysis method, specifically as follows:

[0014] Based on the analysis of the impact of micro-regions and various micro-meteorological factors on photovoltaic output power, firstly, at the urban micro-region level, the urban heat island effect, the density and coverage of building clusters and trees, rainfall and snowfall, air quality, wind and cloud cover, and other factors interact in different micro-region environments, resulting in different micro-meteorological index values. Secondly, at the specific level of micro-meteorological factors, a strong correlation is found between PM concentration and atmospheric aerosol optical thickness (AOD). After correction for ambient temperature, aerosol elevation, relative humidity, and PM concentration factors, an AOD parameter estimation model is obtained. Simultaneously, a linear fitting model for precipitable water is proposed based on the influence of temperature, relative humidity, and surface water vapor pressure. AOD and precipitable water are the two most important parameters affecting atmospheric radiation transmission. The REST2 model is used to calculate diffuse radiation and direct radiation to obtain the total radiation intensity received by the photovoltaic system. The degree of haze accumulation on photovoltaic panels in cities is affected by three micro-meteorological factors: PM cumulative concentration, wind speed and direction, and precipitation, which is reflected in the photovoltaic power generation attenuation rate and photoelectric conversion efficiency.

[0015] In the analysis of factors affecting photovoltaic output power, various micrometeorological factors were ultimately summarized into three strongly correlated influencing factors: photovoltaic cell operating temperature, radiation intensity, and haze concentration.

[0016] Furthermore, an engineering model for photovoltaic output power is used to analyze the factors affecting photovoltaic output power. The expression for the photovoltaic output power engineering model is as follows:

[0017] P S =(1-η)η PV SE[1-k T (T C +α R E+T S )]

[0018] Among them, P S η represents the output power of photovoltaic power generation; η is the photovoltaic power attenuation rate, and the impact of PM concentration on output power is directly represented by the photovoltaic power attenuation rate; η PV S is the photovoltaic cell conversion efficiency; S is the photovoltaic panel area; E is the total irradiance received by the photovoltaic system; k T T is the temperature correction factor; C T represents the current ambient temperature. C +α R E represents the operating temperature of the photovoltaic cell, where α RT is the temperature correction coefficient of the photovoltaic array; T S T is the surface temperature of the photovoltaic array under the standard test condition.

[0019] Further, the growth stage of the distributed photovoltaic system is divided into initial stage, development stage and mature stage based on the Bass diffusion model, specifically:

[0020] According to the cumulative access amount and market potential of the distributed photovoltaic, the initial stage, development stage and mature stage of the distributed photovoltaic installation degree are divided; the initial stage and mature stage defined by the Bass diffusion model are static spatial scenarios, and the development stage is a dynamic spatial scenario, which can better describe and manage different development stages of high-proportion distributed photovoltaic access distribution network system; the factors in the static spatial scenario are relatively fixed, which is helpful for the planning and system design of short-term stable scenarios; the factors in the dynamic spatial scenario change frequently and need real-time monitoring and management to ensure the efficient and stable operation of the system, which is suitable for long-term and flexible planning.

[0021] Further, the cumulative access amount expression of the distributed photovoltaic is:

[0022]

[0023] Wherein, N adopt is the number of users installing distributed PV; M is the market limit, indicating the number of potential installation users in the city; p is the innovation coefficient, indicating the possibility of users not installing PV being affected by external factors and choosing to install distributed PV; q is the imitation coefficient, indicating the possibility of users not installing distributed PV being affected by other users and choosing to install distributed PV; t is the time variable;

[0024] The market saturation is represented by the installation rate of distributed PV, and its expression is:

[0025]

[0026] Wherein, f(t) is the installation rate of distributed PV, which is used to represent the market saturation, N total is the total number of users in the whole city; let f min be the lower limit of the installation rate, f max be the upper limit of the installation rate; when f(t)≤f min , it is in the initial stage; when f min <f(t)<f max , it is in the development stage; when f(t)≥f max , it is in the mature stage.

[0027] Further, the optimized configuration model of geographic space is constructed as:

[0028] (1) Determine the model decision variables, including installation decision variables and observable decision variables;

[0029] Installation decision variable z i The expression is:

[0030]

[0031] Wherein, the installation decision variable z i is a binary variable, indicating whether to install a monitoring device at photovoltaic node i, z i = 1 indicates that a monitoring device is installed at node i; z i = 0 indicates that it is not installed;

[0032] Observable decision variable c ij is to take the observation radius of the monitoring device as the geographic space threshold, and the expression is:

[0033]

[0034] Wherein, the observable decision variable c ij is a binary variable, indicating whether the jth monitoring device covers the ith distributed photovoltaic node; ρ is the observation range radius of the monitoring device; D s,ij is the distance between nodes i and j; c ij = 1 indicates that node i can be observed by the monitoring device configured by node j; c ij = 0 indicates that it cannot be observed;

[0035] (2) The objective function I is to maximize the total weight of the covered photovoltaic nodes and minimize the economic constraint weight, and the specific expression is:

[0036]

[0037] Wherein, F1 is the objective function I, ω e is the economic coefficient; N s is the total number of distributed photovoltaic nodes in the whole region, then represents the total number of distributed photovoltaic nodes in the whole region; K i is the node criticality coefficient corresponding to each distributed photovoltaic node, then represents the coverage of the criticality of the monitored distributed photovoltaic nodes in the whole region; y Penal as a penalty term, limits the number of monitoring devices in the output optimal solution;

[0038] (3) The constraint conditions include observable constraints and monitoring device number constraints;

[0039] For the observable constraint condition of the node, if a node is covered by the observation range of at least one monitoring device, it indicates that the node is in an observable state, and the specific expression is:

[0040]

[0041] Further, the expression of the observable state of node i is obtained as:

[0042]

[0043] wherein, m represents the total number of monitoring devices that can observe node i in the whole domain; m i represents the observable state of node i, when m i = 1, indicating that node i can be observed by at least one monitoring device, so node i is observable; otherwise, m i = 0, indicating that node i is unobservable;

[0044] The specific expression of the monitoring device number constraint for the whole system is:

[0045]

[0046] wherein, N o is the upper limit value of the number of monitoring devices;

[0047] The penalty function is introduced to handle the constraint number exceeding situation. If the number of selected monitoring devices exceeds N o , the corresponding penalty is generated; the penalty function y Penal is defined as follows:

[0048]

[0049] Further, the optimization configuration model of microclimate features is constructed, which is specifically:

[0050] (1) Determine the model decision variables, including installation decision variables and observable decision variables;

[0051] The expression of the installation decision variable z i is:

[0052]

[0053] wherein, the installation decision variable z i is a binary variable, indicating whether to install a monitoring device at photovoltaic node i, z i = 1 indicates that a monitoring device is installed at node i; z i = 0 indicates that it is not installed;

[0054] The expression of the observable decision variable s ij is:

[0055]

[0056] where A ij represents the difference of total radiation between node i and node j; B ij represents the difference of temperature between node i and node j; C ij represents the difference of air quality between node i and node j; r A , r B , r C respectively represent the threshold of total radiation, temperature and air quality; if A ij , B ij , C ij all satisfy the threshold requirement, it indicates that the nodes i and j are similar in microclimate conditions.

[0057] (2) Objective function II optimizes the layout of monitoring devices through microclimate similarity, and considers the growth expectation of future photovoltaic access in the time dimension, and the specific expression is:

[0058] al

[0059] where F2 is the objective function II, ω e is the economic coefficient; N s is the total number of photovoltaic nodes in the whole domain, then represents the total number of distributed photovoltaic in the whole domain; K i is the node criticality coefficient corresponding to each distributed photovoltaic node; T i (t) is the time coefficient, which represents the future access amount expectation of the ith photovoltaic node in the tth year, then represents the criticality coverage of the monitored distributed photovoltaic in the whole domain over time; y Penal as a penalty term, limits the number of monitoring devices in the output optimal solution;

[0060] (3) The constraint conditions include observable constraints and monitoring device number constraints;

[0061] For the observable constraint condition of the node, if a node is covered by the observation range of at least one monitoring device, it indicates that the node is in an observable state, and the specific expression is:

[0062]

[0063] Further, the expression of the observable state of node i is obtained as:

[0064]

[0065] where, represents the total number of monitoring devices that can observe node i in the whole domain; l i1, indicates that node i can be observed by at least one monitoring device, so node i is observable; otherwise i 1, indicates that node i can be observed by at least one monitoring device, so node i is observable; otherwise i 0, indicates that node i is not observable;

[0066] The specific expression for the monitoring device number constraint of the whole system is:

[0067]

[0068] N o is the upper limit value of the number of monitoring devices set;

[0069] The penalty function is introduced to handle the case of exceeding the number of constraints. If the number of selected monitoring devices exceeds N o , a corresponding penalty is generated; the penalty function y Penal is defined as follows:

[0070]

[0071] Further, for the initial period, a genetic optimization algorithm is used to solve the optimization configuration model of geographic space to obtain the optimization configuration result of the micro-meteorological monitoring device.

[0072] For the development period and the mature period, a mathematical programming method is used and the Gurobi solver is used to solve the optimization configuration model of the micro-meteorological characteristics to obtain the optimization configuration result of the micro-meteorological monitoring device.

[0073] Further, the economic rationality and accuracy of the optimization configuration result are tested and judged, specifically:

[0074] The key degree of observability index OI is tested, which is used to describe the global distributed photovoltaic observability under a certain optimization configuration result, analyze the coverage effect of the monitoring device of the optimization configuration result on the global observation, and thus judge the economy of the optimization configuration result, which is specifically represented as:

[0075]

[0076] α OI is the node comprehensive key degree of observability index; is the total number of nodes that can be observed in the whole domain;

[0077] According to the shared micro-meteorological observation data of the optimization configuration result, the distributed photovoltaic output prediction effect is analyzed, the absolute value and relative error of the prediction error are evaluated by RMSE and MAPE, and the fitting degree of the model is evaluated by R 2 , to judge the accuracy and rationality of the configuration result; ​

[0078] Finally, by comparing with the traditional grid division configuration method, the superiority of the experimental method is analyzed.

[0079] Beneficial effects: Compared with the prior art, the significant technical effects of the present application are: the proposed node comprehensive criticality evaluation method takes into account the differences of microclimate in urban environment and the influence on photovoltaic output in addition to single electrical aspect index evaluation, making up for the shortcomings of existing weak node evaluation index of power grid; in addition, the global optimization configuration method based on GA algorithm can describe the node criticality of distributed photovoltaic in power grid, and the proposed OI value and node critical coverage as evaluation criteria can describe the performance of various configuration schemes; according to the cost demand, the upper limit of device quantity and OI value are set, and under the limited conditions, the optimal configuration result of balancing the observability and economic performance is given. It not only overcomes the problem of mismatching between economic performance and technical performance of traditional grid configuration, but also solves the problem of inflexibility of grid configuration. BRIEF DESCRIPTION OF DRAWINGS

[0080] Figure 1 The method flowchart of the present application is shown in the figure;

[0081] Figure 2 The microclimate factor influence schematic diagram in the embodiment of the present application is shown in the figure;

[0082] Figure 3 The global PV node different factor sensitivity index and node criticality index schematic diagram in the embodiment of the present application is shown in the figure;

[0083] Figure 4 The distributed photovoltaic installation user quantity prediction in the embodiment of the present application is shown in the figure;

[0084] Figure 5 The distributed photovoltaic system installation rate prediction in the embodiment of the present application is shown in the figure;

[0085] Figure 6 The convergence trend of individual and overall optimal solution in the genetic algorithm solving process in the embodiment of the present application is shown in the figure;

[0086] Figure 7 The global observability corresponding to different device numbers in the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0087] The present application will be further described below in combination with the figures and specific embodiments.

[0088] In view of the problem that the city containing a high proportion of distributed photovoltaic system cannot meet the comprehensive observability and measurability demand, the present application proposes a distributed photovoltaic microclimate monitoring device optimization configuration method considering the time and space characteristics, mainly including a distributed photovoltaic system scene division method under the time and space variation characteristics of urban distribution network and a monitoring device optimization configuration method under different scenes, such asFigure 1 The method of the present application specifically comprises the following steps:

[0089] Step A, for the roof distributed photovoltaic system with high proportion of access in the city, calculate the distribution network parameters and meteorological information data, and obtain the node comprehensive criticality index of the distributed photovoltaic system by comprehensively considering the electrical and micro-meteorological sensitivities.

[0090] Step B, according to the growth influencing factors of the distributed photovoltaic system, divide the growth stage of the distributed photovoltaic system into initial period, development period and mature period based on the Bass diffusion model; for the initial period, according to the node comprehensive criticality index, take maximizing the total weight of the photovoltaic node covered and minimizing the economic constraint weight as the objective function, and construct the optimization configuration model of the geographic space; for the development period and the mature period, according to the comprehensive node criticality index, optimize the layout of the micro-meteorological monitoring device according to the micro-meteorological similarity, and take the growth expectation of the future photovoltaic access in the time dimension as the objective function, and construct the optimization configuration model of the micro-meteorological characteristics; solve the optimization configuration model to obtain the optimization configuration result of the micro-meteorological monitoring device;

[0091] Step C, according to the optimization configuration result, calculate the observability index and output prediction, and test and judge the economy, rationality and accuracy of the optimization configuration result.

[0092] In the embodiment of the present application, the specific operation of step A is as follows:

[0093] Step A01, the electrical sensitivity analysis method is based on the structure and electrical parameters of the distribution network system, and studies the voltage stability through the differential relationship of the variables in the distribution network system based on the power flow calculation, which can be effectively applied to the weak node determination. The power flow calculation formula is transformed into a matrix, and the voltage sensitivity expression is as follows:

[0094]

[0095] Wherein, S PU and S QU are the active and reactive sensitivity matrices respectively, which represent the change of the node voltage amplitude when the node injects a unit amount of active power and reactive power; S Pδ and S Qδ represent the change amount matrix of the node voltage phase angle when the node injects a unit amount of active power and reactive power, Δδ is the node voltage phase angle change amount matrix, ΔU is the node voltage amplitude change amount matrix, ΔP is the node injected active power change amount matrix, and ΔQ is the reactive power change amount matrix.

[0096] Step A02, determine the micro-meteorological comprehensive sensitivity index according to the micro-meteorological fluctuation influence analysis method, and the micro-meteorological influence index analysis (i.e. the micro-meteorological fluctuation influence analysis method) in the embodiment of the present application is as follows:Figure 2 are shown:

[0097] Assuming that the micro-area is a grid area divided by 10 kilometers in the city, there are obvious differences in micro-meteorological environment between different micro-areas. According to the analysis of the influence of micro-area and various micro-meteorological factors on photovoltaic output power, firstly, for the city micro-area level, due to the mutual influence of urban heat island effect, the density and shading range of building groups and trees, rainfall and snowfall, air quality, wind and cloud cover, the micro-meteorological index values in the micro-area environment are different. Secondly, for the specific level of micro-meteorological factors, there is a strong correlation between PM concentration (i.e. haze concentration) and atmospheric aerosol optical depth AOD. After correction by environmental temperature, aerosol height, air relative humidity and PM concentration factor, the AOD parameter estimation model is obtained. At the same time, according to the influence of temperature, air relative humidity and ground water vapor pressure, a linear fitting model of precipitable water can be proposed. AOD and precipitable water are the two most important parameters affecting atmospheric radiation transmission. Through the REST2 model, the total radiation intensity received by the photovoltaic panel is obtained by calculating the scattered radiation and direct radiation. In addition, the degree of dust accumulation of the photovoltaic panel in the city is mainly affected by three micro-meteorological factors, namely PM cumulative concentration, wind speed and direction, and precipitation, which is reflected in the photovoltaic power attenuation rate and photoelectric conversion efficiency.

[0098] Therefore, in the analysis of the influencing factors of photovoltaic output power, various micro-meteorological factors are finally summarized into three strongly correlated influencing factors, namely the working temperature of the photovoltaic cell (i.e. Figure 2 the working temperature of the photovoltaic panel), radiation intensity and PM concentration.

[0099] Therefore, the expression of the photovoltaic output power engineering model is as follows:

[0100] P S = (1-η)η PV SE[1-k T (t0+T S )] (2)

[0101] In the formula, P S is the output power of photovoltaic power generation; η is the photovoltaic power attenuation rate, and the influence of PM concentration on output power is directly represented by the photovoltaic power attenuation rate; η PV is the photovoltaic cell conversion efficiency; S is the area of the photovoltaic cell panel; E is the total irradiance received by the photovoltaic panel; k T is the temperature correction coefficient, which is 0.005; t0 is the working temperature of the photovoltaic cell, which is specifically expressed as t0=T C +α R E, where T C is the current environmental temperature, α R is the photovoltaic array temperature correction coefficient, which is 0.03; T SFor the standard test case of photovoltaic array surface temperature, generally take 25.

[0102] Step A03, based on the photoelectric conversion model, the partial derivative of power with respect to η, E and T three parameters as the corresponding weight coefficient of microclimate sensitivity factor, the specific expression as follows:

[0103]

[0104] Where, ω PM is the PM concentration coefficient, ω E is the irradiation coefficient, ω T is the temperature coefficient.

[0105] Based on the photovoltaic output power engineering model, according to the Pearson similarity calculation formula, the microclimate sensitivity factor of microclimate data and the same period photovoltaic power generation data is calculated, the specific expression as follows:

[0106]

[0107] Where, r xy is the n-dimensional vector correlation coefficient between photovoltaic power station and microclimate data, as the microclimate sensitivity factor; n is the number of data samples, the number of microclimate data samples and the number of photovoltaic power generation data samples are the same, both n; x i is the same period microclimate data sample, including air quality sensitivity factor C PM,i , total radiation sensitivity factor E i and environmental temperature sensitivity factor T i ; y i is the photovoltaic power station output power P s,i ; is the average value of microclimate data sample, is the average value of photovoltaic power station output power sample

[0108] Step A04, because the microclimate, voltage and photovoltaic output power dimension are different, in order to avoid the interference of the magnitude difference between different data on subsequent model, therefore, the index is normalized to [-1, 1] interval for effective comparison. The normalization formula is:

[0109]

[0110] Where, indicates the normalized output value, x indicates the sample data, x min and x max indicate the minimum and maximum values in the sample data.

[0111] Step A05, comprehensive grid and microclimate sensitivity, the node comprehensive criticality index expression of each node distributed photovoltaic in the system is as follows:

[0112]

[0113] Wherein, K i is the node criticality coefficient corresponding to each distributed photovoltaic node;δ i represents the electrical comprehensive sensitivity index, ρ i represents the microclimate comprehensive sensitivity index, represents the normalized electrical comprehensive sensitivity index, represents the normalized microclimate comprehensive sensitivity index;ω P represents the active weight coefficient;ω Q represents the reactive weight coefficient;s PU represents the active sensitivity factor;s QU represents the reactive sensitivity factor;ω E represents the irradiation coefficient;ω T represents the temperature coefficient;ω PM represents the PM concentration coefficient;E i represents the total radiation sensitivity factor of node i;T i represents the ambient temperature sensitivity factor;C PM,i represents the air quality sensitivity factor.

[0114] In the embodiment of the application, the node criticality is calculated by using a typical IEEE33 node system, and the calculation result is as shown in Figure 3 .

[0115] In the embodiment of the application, the specific operation of step B of the monitoring device optimization configuration method considering the time-space variation characteristics is as follows:

[0116] Step B01, the distributed photovoltaic installation user quantity variation and installation rate variation based on Bass diffusion model are as shown in Figure 4 and Figure 5 .

[0117] The cumulative access amount expression of distributed photovoltaic is:

[0118]

[0119] Wherein, N adoptN is the number of users who install distributed PV; M is the market limit, representing the potential number of users who install in the overall city; p is the innovation coefficient, representing the possibility of users who do not install PV to install distributed PV under the influence of installation promotion by equipment popularization merchants, government policy encouragement and other external factors; q is the imitation coefficient, representing the possibility of users who do not install distributed PV to install distributed PV under the influence of other users; and t is the time variable.

[0120] The expression of market saturation (distributed PV installation rate) is:

[0121]

[0122] Wherein, f(t) is the distributed PV installation rate, used to represent market saturation, N total is the total number of users in the overall city; f min is the lower limit of the installation rate, f max is the upper limit of the installation rate; in the embodiment, the lower limit of the installation rate f min may be 0.1, and the upper limit of the installation rate f max may be 0.9.

[0123] When f(t)≤f min , it is in the initial stage. The access and diffusion of the distributed photovoltaic system are in the initial stage, the growth is relatively slow, and is mainly driven by innovators. At this time, the preliminary enhancement construction of the power grid is carried out, and the change is infrequent. The number of early distributed photovoltaic systems is small, and the distribution is relatively dispersed.

[0124] When f min <f(t)<f max , it is in the development stage, the access and diffusion speed of the distributed photovoltaic system is accelerated, and is mainly driven by imitators, and the growth is rapid. At this time, a large number of new distributed photovoltaic systems are accessed in the distribution network, various different types and scales of photovoltaic projects grow rapidly, and real-time meteorological data becomes more important.

[0125] When f(t)≥f max , it is in the mature stage, at this time, a large number of distributed photovoltaic access transformations have been completed, the structure is stable, the access of the distributed photovoltaic system reaches a stable state, and the growth tends to be saturated. This stage is mainly for maintenance and small-scale upgrading, and the change is less. The photovoltaic system is widely distributed in the city, and the layout is basically fixed.

[0126] In step B02, for the initial stage, a static space scene can be regarded. Since the number of distributed photovoltaic systems in the initial stage is small, the configuration scheme selection is various, and the flexibility of the genetic optimization algorithm used to solve the optimal configuration problem is large. The convergence trend of the individual and the global optimal solution in the solving process of the genetic algorithm in the embodiment of the application is as shown in Figure 6 .

[0127] For the initial stage, an optimal geospatial configuration model is constructed, specifically as follows:

[0128] (1) Determine the model decision variables, including installation decision variables and observable decision variables;

[0129] Install decision variable z i The expression is:

[0130]

[0131] Among them, the installation decision variable z i Z is a binary variable representing whether a monitoring device is installed at photovoltaic node i. i =1 indicates that a monitoring device is installed at node i; z i =0 indicates that it will not be installed.

[0132] Observable decision variable c ij The observation radius of the monitoring device is used as the geospatial threshold, expressed as:

[0133]

[0134] Among them, the observable decision variable c ij ρ is a binary variable representing whether the j-th monitoring device covers the i-th distributed photovoltaic node; ρ is the observation range radius of the monitoring device; D s,ij Let c be the distance between nodes i and j. ij =1 indicates that node i can be observed by the monitoring device configured on node j; c ij =0 indicates that it cannot be observed.

[0135] (2) The objective function is to maximize the total weight of the photovoltaic nodes and minimize the economic constraint weight. The specific expression is as follows:

[0136]

[0137] Where F1 is the objective function I, ω e N is the economic coefficient; S If the total number of distributed photovoltaic systems in the entire region is [number], then [number] K represents the total number of distributed photovoltaic systems in the entire region. i For each distributed photovoltaic node, the node criticality coefficient is then... Represents the critical coverage status of distributed photovoltaic systems monitored across the entire region, m i Represents the observable state of node i; y Penal As a penalty, the number of monitoring devices that output the optimal solution is limited.

[0138] (3) Constraints include observability constraints and the number of monitoring devices;

[0139] For the observable constraint condition of nodes, if a node is covered by the observation range of at least one monitoring device, it means that the node is in an observable state, and the specific expression is:

[0140]

[0141] Further, the expression of the observable state of node i is:

[0142]

[0143] wherein, represents the total number of monitoring devices that can observe node i in the whole region; m i represents the observable state of node i, when m i takes the value of 1, indicating that node i can be observed by at least one monitoring device, so node i is observable; otherwise, m i takes the value of 0, indicating that node i is not observable;

[0144] For the monitoring device number constraint of the whole system, the specific expression is:

[0145]

[0146] wherein, N o is the upper limit value of the number of monitoring devices;

[0147] In addition, a penalty function is introduced to handle the case where the number of constraints exceeds, and if the number of selected monitoring devices exceeds N o , a corresponding penalty is generated; the penalty function y Penal is defined as follows:

[0148]

[0149] At this time, the expression of the optimization configuration model of geographic space is:

[0150]

[0151] wherein, is the economic cost of installing monitoring devices; ω e is the economic coefficient; N S is the total number of photovoltaic nodes in the whole region, then represents the total number of distributed photovoltaic in the whole region; K i is the node criticality coefficient corresponding to each distributed photovoltaic node, then represents the criticality coverage of the monitored distributed photovoltaic in the whole region; m i represents the observable state of node i; N o is the upper limit value of the number of monitoring devices; y PenalThe penalty function handles the case where the number of constraints exceeds.

[0152] In the embodiments of the present application, a genetic optimization algorithm is used to solve the optimization configuration model of geographic space, and an optimization configuration result is obtained. Figure 7

[0153] Step B03, for the development stage, can be regarded as a dynamic spatial scene of long-term change. Distributed photovoltaic access and diffusion grow rapidly, and in the optimization configuration model of this stage, microclimate characteristics are introduced as the radius index for dividing similarity. Due to the expansion of the system, a mathematical programming method is used and a Gurobi solver is used to solve, so as to obtain an accurate and stable configuration scheme result.

[0154] At this time, the optimization configuration model of microclimate characteristics is constructed, specifically:

[0155] (1) initialize the Gurobi parameter, determine the model decision variable, including the installation decision variable and the observable decision variable;

[0156] The installation decision variable z i is expressed as:

[0157]

[0158] Among them, the installation decision variable z i is a binary variable, indicating whether to install a monitoring device at the photovoltaic node i, z i = 1 indicates that a monitoring device is installed at the node i; z i = 0 indicates that it is not installed.

[0159] The observable decision variable s ij is based on the microclimate characteristics, and the microclimate difference between nodes is compared with the microclimate similarity threshold, which is expressed as:

[0160]

[0161] Among them, A ij represents the total radiation difference between node i and node j; B ij represents the temperature difference between node i and node j; C ij represents the air quality difference between node i and node j (A ij = |A j -A i |, B ij = |B j -B i |, C ij = |C j -C i |), wherein A i and A​j B represents the total radiation of node i and node j, respectively. i and B j The temperatures of nodes i and j are C, respectively. i and C j The air quality between node i and node j are respectively; r A r B r C These represent the threshold values ​​for total radiation, temperature, and air quality, respectively. If A ij B ij C ij If all three conditions meet the threshold requirements, it indicates that nodes i and j are similar in terms of micrometeorological conditions.

[0162] (2) The objective function optimizes the layout of monitoring devices through micro-meteorological similarity and considers the expected growth of future photovoltaic access in the time dimension. The specific expression is as follows:

[0163]

[0164] Where F2 is the objective function II, ω e N is the economic coefficient; S If the total number of photovoltaic nodes in the entire region is , then K represents the total number of distributed photovoltaic systems in the entire region. i The node criticality coefficient for each distributed photovoltaic node; T i (t) is a time coefficient, representing the expected future grid connection volume of the i-th photovoltaic node in year t. This represents the critical coverage status of distributed photovoltaic systems monitored across the entire region over time; y Penal As a penalty, the number of monitoring devices that output the optimal solution is limited;

[0165] (3) Constraints include observability constraints and the number of monitoring devices;

[0166] Regarding the observability constraint for a node, if a node is covered by the observation range of at least one monitoring device, then the node is considered observable. The specific expression is as follows:

[0167]

[0168] Furthermore, the expression for the observable state of node i is obtained as follows:

[0169]

[0170] in, This represents the total number of monitoring devices capable of observing node i across the entire region; l i Represents the observable state of node i, when time li Value 1, indicating that node i can be observed by at least one monitoring device, so node i is observable. i Value 0, indicating that node i is not observable.

[0171] For the number constraint of the monitoring device of the whole system, the specific expression is:

[0172]

[0173] Wherein, N o is the upper limit value of the number of monitoring devices;

[0174] In addition, the penalty function is introduced to handle the constraint number exceeding situation, and if the number of selected monitoring devices exceeds N o , the corresponding penalty is generated; the penalty function y Penal is defined as follows:

[0175]

[0176] At this time, the expression of the optimization configuration model of the micro-meteorological characteristics is:

[0177]

[0178] Wherein, is the economic cost of installing monitoring devices; ω e is the economic coefficient; N S is the total number of photovoltaic nodes in the whole domain, so represents the total number of distributed photovoltaic in the whole domain; T i (t) is the time coefficient, indicating the expected future access amount of the ith photovoltaic node in the tth year, so represents the coverage of the monitored distributed photovoltaic key degree in the whole domain changing with time; l i represents the observable state of node i; N o is the upper limit value of the number of monitoring devices; y Penal is the penalty function, which handles the constraint number exceeding situation.

[0179] In the embodiment of the application, the optimization configuration model of the micro-meteorological characteristics is solved by using the mathematical programming method and the Gurobi solver, and the optimization configuration result of the micro-meteorological monitoring device is obtained.

[0180] For the mature stage scenario, it can be regarded as a static space scenario finally installed saturated. The growth of distributed photovoltaic tends to be saturated, a large number of distributed photovoltaic access transformation is completed, the system reaches a stable state, and the overall layout is basically fixed. In this stage, the above-mentioned optimization configuration model of the micro-meteorological characteristics is also constructed, and the mathematical programming method and the Gurobi solver are still used to solve the optimization configuration model.

[0181] In the embodiment of the present application, the specific operation of step C is as follows:

[0182] Step C01, the test and judgment method of the effect of the optimization configuration result:

[0183] Through the key degree observability index (OI) test, the global distributed photovoltaic observability under a certain optimization configuration result is described, the coverage effect of the monitoring device of the optimization configuration result on the global observation is analyzed, and the economic performance of the configuration result is judged, which is specifically expressed as:

[0184]

[0185] Wherein, α OI is the node comprehensive key degree observability index; is the total number of nodes that can be observed in the global.

[0186] Step C02, according to the shared microclimate observation data of the configuration result, the prediction effect of the distributed photovoltaic output is analyzed, the absolute value and relative error of the prediction error are evaluated through RMSE and MAPE, and the fitting degree of the model is evaluated through R 2 The accuracy and rationality of the configuration result are judged.

[0187] Step C03, finally, through the comparison with the traditional grid division configuration method, the superiority of the experimental method is compared and analyzed.

[0188] Therefore, combined with the definition of the key link of the power grid in the city distribution network and the microclimate state, and taking the maximum observability, the most optimal economy, the accurate and flexible adjustment of the monitoring and prediction effect as the goal, the optimization configuration method of different levels of monitoring device is proposed in stages, which has important significance for realizing that the photovoltaic detection equipment meets the observable and measurable demand of the high proportion of distributed photovoltaic operation state in the scale growth change.

[0189] The above only describes the preferred embodiments of the present application, and it should be pointed out that for ordinary skilled persons in the technical field, several improvements and modifications can be made without departing from the technical principles of the present application, and these improvements and modifications should be regarded as the protection scope of the present application.

Claims

1.A method for optimizing configuration of a distributed photovoltaic microclimate monitoring device considering spatiotemporal characteristics, characterized in that, Comprising the following steps: For the high proportion of access to the roof of the city distributed photovoltaic system, the distribution network parameters and microclimate information data are calculated, and the node comprehensive key index of the distributed photovoltaic system is obtained by comprehensively considering the electrical and microclimate sensitivity, and the expression is: wherein, K i is the node criticality coefficient corresponding to each distributed photovoltaic node; δ i represents the electrical comprehensive sensitivity index, p i represents the micro-meteorological comprehensive sensitivity index, represents the normalized electrical comprehensive sensitivity index, represents the normalized micro-meteorological comprehensive sensitivity index; ω P represents the active weight coefficient; ω Q represents the reactive weight coefficient; s PU represents the active sensitivity factor; s QU represents the reactive sensitivity factor; ω E represents the irradiance coefficient; ω T represents the temperature coefficient; ω PM represents the PM concentration coefficient; E i represents the total radiation sensitivity factor of node i; T i represents the ambient temperature sensitivity factor; C PM,i represents the air quality sensitivity factor; According to the influence analysis method of microclimate fluctuation on photovoltaic output, the microclimate comprehensive sensitivity index is determined, including: The expression of the photovoltaic output power engineering model is: P S = (1 - η)η PV SE[1 - k T (T C + α R E + T S )] Wherein, P S is the output power of photovoltaic power generation; η is the photovoltaic power generation power attenuation rate, the influence of PM concentration on the output power is directly represented by the photovoltaic power attenuation rate; η PV is the conversion efficiency of the photovoltaic cell; S is the area of the photovoltaic cell panel; E is the total irradiance received by the photovoltaic cell; k T is the temperature correction coefficient; T C is the current ambient temperature; T C + α R E represents the working temperature of the photovoltaic cell, α R is the temperature correction coefficient of the photovoltaic array; T S is the surface temperature of the photovoltaic array under standard test conditions; Based on the photoelectric conversion model, the partial derivative of power with respect to η, E and T is taken as the weight coefficient corresponding to the microclimate sensitivity factor; Based on the photovoltaic output power engineering model, the microclimate sensitivity factor of the microclimate data and the photovoltaic power generation data at the same period is calculated according to the Pearson similarity calculation formula; The microclimate, voltage and photovoltaic output are normalized; According to the growth influencing factors of the distributed photovoltaic system, the growth stages of the distributed photovoltaic system are divided into initial period, development period and mature period based on Bass diffusion model; for the initial period, according to the node comprehensive key index, the total weight of the maximum coverage photovoltaic node and the minimum economic constraint weight are taken as the objective function I, and the optimization configuration model of geographic space is constructed; for the development period and mature period, according to the comprehensive node key index, the layout of the microclimate monitoring device is optimized, and the growth expectation of future photovoltaic access in time dimension is considered as the objective function II, and the optimization configuration model of microclimate characteristics is constructed; the optimization configuration model is solved to obtain the optimization configuration result of the microclimate monitoring device; The constructed optimization configuration model of geographic space is: (1) determine the model decision variable, including installation decision variable and observation decision variable; (2) the objective function I is to maximize the total weight of the covered photovoltaic node and minimize the economic constraint weight, and the specific expression is: Wherein, F1 is the objective function I, ω e is the economic coefficient; N s is the total number of distributed photovoltaic in the whole region, then represent the total number of distributed photovoltaic in the whole region; K i is the node key degree coefficient corresponding to each distributed photovoltaic node, then represent the key degree coverage of the monitored distributed photovoltaic in the whole region; y Penal As a penalty term, limit the number of monitoring devices of the output optimal solution; (3) the constraint conditions include observation constraint and monitoring device quantity constraint; For the observation constraint condition of the node, if a node is covered by at least one monitoring device, it means that the node is in the observable state, and the specific expression is: Further, the expression of the observable state of node i is: where, represents the total number of monitoring devices that can observe node i in the global domain; m i represents the observable state of node i, when m i takes the value of 1, indicating that node i is observable by at least one monitoring device, so node i is observable; otherwise m i takes the value of 0, indicating that node i is not observable; For the monitoring device quantity constraint of the whole system, the specific expression is: N o is an upper limit value for setting the number of monitoring devices; The penalty function is introduced to deal with the case that the number of constraints exceeds, if the number of selected monitoring devices exceeds N o The corresponding penalty is generated; the penalty function y Penal is defined as follows: The optimization configuration model of microclimate characteristics is constructed, which is: (1) determine the model decision variable, including installation decision variable and observation decision variable; (2) the objective function II optimizes the layout of the monitoring device through the microclimate similarity, and considers the growth expectation of future photovoltaic access in time dimension, and the specific expression is: Wherein, F2 is the objective function II, ω e is the economic coefficient; N s is the total number of photovoltaic nodes in the whole region, then represent the total number of distributed photovoltaic in the whole region; K i is the node key degree coefficient corresponding to each distributed photovoltaic node; T i (t) is the time coefficient, which represents the expected future access amount of the ith photovoltaic node in the tth year, then represent the key degree coverage of the monitored distributed photovoltaic in the whole region over time; y Penal As a penalty term, limit the number of monitoring devices of the output optimal solution; (3) the constraint conditions include observation constraint and monitoring device quantity constraint; For the observation constraint condition of the node, if a node is covered by at least one monitoring device, it means that the node is in the observable state, and the specific expression is: Further, the expression of the observable state of node i is: where, represents the total number of monitoring devices that can observe node i in the global domain; l i represents the observable state of node i, when l i takes the value of 1, indicating that node i is observable by at least one monitoring device, so node i is observable; otherwise l i takes the value of 0, indicating that node i is not observable; For the monitoring device quantity constraint of the whole system, the specific expression is: N o is an upper limit value for setting the number of monitoring devices; The penalty function is introduced to deal with the case that the number of constraints exceeds, if the number of selected monitoring devices exceeds N o A corresponding penalty is generated; the penalty function y Penal is defined as follows: For the optimization configuration result, the observability index and output prediction calculation are carried out, and the economy, rationality and accuracy of the optimization configuration result are tested and judged. 2.The method for optimizing configuration of distributed photovoltaic microclimate monitoring devices considering space-time characteristics according to claim 1, characterized in that, According to the Bass diffusion model, the growth stages of the distributed photovoltaic system are divided into initial period, development period and mature period, which are: The initial stage, development stage and mature stage of distributed photovoltaic installation degree are divided according to the cumulative access amount and market potential of distributed photovoltaic. The initial stage and mature stage of Bass diffusion model are defined as static spatial scenarios, and the development stage is defined as a dynamic spatial scenario, which can better describe and manage different development stages of high proportion of distributed photovoltaic access distribution network system. The factors in static spatial scenarios are relatively fixed, which is helpful for the planning and system design of short-term stable scenarios. The factors in dynamic spatial scenarios change frequently, which need real-time monitoring and management to ensure the efficient and stable operation of the system, and are suitable for long-term flexible planning. 3.The method for optimizing configuration of distributed photovoltaic microclimate monitoring devices considering space-time characteristics according to claim 2, characterized in that, The cumulative access amount expression of distributed photovoltaic is: where N adopt is the number of users installing distributed PV; M is the market limit, representing the number of potential users installing PV in the city; p is the innovation coefficient, representing the possibility of users not installing PV being influenced by external factors to choose to install distributed PV; q is the imitation coefficient, representing the possibility of users not installing distributed PV being influenced by other users to choose to install distributed PV; and t is the time variable; The market saturation is expressed by the installation rate of distributed PV, and its expression is: where f(t) is the distributed PV installation rate, used to represent market saturation, N total is the total number of users in the entire city; let f min be the lower limit of the installation rate, f max is the upper limit of the installation rate; when f(t)≤f min , it is in the initial stage; when f min < f(t) < f max , it is in the development stage; when f(t)≥f max , it is in the mature stage. 4.The method of claim 1, wherein, Geospatial optimization placement model installation decision variable z i The expression is: where the installation decision variable z i is a binary variable indicating whether a monitoring device is installed at photovoltaic node i, z i = 1 indicates that a monitoring device is installed at node i; z i = 0 indicates that it is not installed; observable decision variable c ij The observation radius of the monitoring device is taken as the geospatial threshold, expressed as: where c is the observable decision variable ij is a binary variable indicating whether the jthmonitoring device covers the ithdistributed PV node; p is the monitoring device's observation range radius; D s,ij is the distance between nodes i and j; c ij = 1 indicates that node i can be observed by the monitoring device configured at node j; c ij = 0 indicates that it cannot be observed. 5.The method for optimizing configuration of distributed photovoltaic microclimate monitoring devices considering space-time characteristics according to claim 1, characterized in that, Optimal configuration model of microclimate features with installation decision variable z i The expression is: where the installation decision variable z i is a binary variable indicating whether a monitoring device is installed at photovoltaic node i, z i = 1 indicates that a monitoring device is installed at node i; z i = 0 indicates that it is not installed; Observable decision variable s ij Based on the microclimate characteristics, the microclimate difference between nodes is compared with the microclimate similarity threshold, and the expression is: where A ij represents the difference of total radiation between node i and node j; B ij represents the difference of temperature between node i and node j; C ij represents the difference of air quality between node i and node j; r A , r B , r C respectively represent the threshold of total radiation, temperature and air quality; if A ij , B ij , C ij all satisfy the threshold requirement, it means that the similarity of microclimate between node i and j. 6.The method for optimizing configuration of distributed photovoltaic microclimate monitoring devices considering space-time characteristics according to claim 1, characterized in that, For the initial stage, the genetic optimization algorithm is used to solve the optimization configuration model of geographical space, and the optimization configuration results of micro-meteorological monitoring devices are obtained. For the development stage and mature stage, the optimization configuration model of micro-meteorological characteristics is solved by using mathematical programming method and Gurobi solver, and the optimization configuration results of micro-meteorological monitoring devices are obtained. 7.The method for optimizing configuration of distributed photovoltaic microclimate monitoring devices considering space-time characteristics according to claim 1, characterized in that, The economic rationality and accuracy of the optimization configuration results are tested and judged, which are as follows: The key degree of observability index OI is used to test the global distributed photovoltaic observation under a certain optimization configuration result, analyze the coverage effect of the monitoring devices of the optimization configuration result on the global observation, and judge the economy of the optimization configuration result, which is specifically expressed as: wherein, α OI is the integrated key degree of node observability index; is the total number of nodes observable within the universe; According to the sharing microclimate observation data of the optimization configuration result, the distributed photovoltaic output prediction effect is analyzed, the absolute value and relative error of the prediction error are evaluated through RMSE and MAPE, and the fitting degree of the evaluation model is evaluated, and the accuracy and rationality of the configuration result are judged. 2 The fitting degree of the evaluation model is evaluated, and the accuracy and rationality of the configuration result are judged. Finally, the superiority of the experimental method is compared and analyzed by comparing with the traditional grid division configuration method.

Images