Multi-mode adaptive operation control method for intelligent energy storage soft switch under multi-source measurement driving

Abstract

The invention relates to an intelligent energy storage soft switching multi-mode adaptive operation control method under multi-source measurement driving, and the method achieves the data modeling of the operation state of a power distribution network through the measurement data quality identification and processing, avoids the problems of measurement bad data disturbance and the like in a complex operation state, and improves the reliability of the power distribution network. And the multi-mode mapping matrix can construct self-adaptive prediction control models in various modes by using multi-source measurement data according to the control mode requirements of the power distribution network, support is provided for realizing multi-mode operation control of the power distribution network, and on the basis, the electric energy conversion-storage-supply capability of the ESOP is utilized, so that the multi-mode operation control of the power distribution network is realized. The control modes are adaptively switched according to the control requirements of different areas, the power requirements of different areas in the actual operation process are judged and adapted in time, the operation control flexibility of the power distribution network is improved, and the output of the distributed power supply and the rapid load fluctuation are rapidly responded.

Description

Technical Field

[0001] The present invention belongs to the technical field of power distribution system operation control, and in particular to a multi-mode adaptive operation control method for an intelligent energy storage soft switch driven by multi-source measurement. Background Art

[0002] At present, the large-scale access of distributed generators (DG) has become an important means for the development of new power systems, but it has also brought new challenges to the operation and control of distribution networks. The intermittent and random nature of the output of distributed generators has led to increasingly complex and changeable operating conditions of distribution networks, which may cause problems such as power flow reversal, voltage over-limit, and line overload. Single-target control is difficult to meet the actual operation needs of new distribution networks. Therefore, in order to achieve flexible operation and control of distribution networks, it is often necessary to set a variety of control targets, mainly including improving voltage levels, reducing network active power losses, balancing line loads, and improving the absorption capacity of distributed generators.

[0003] Traditional control equipment in distribution networks, such as static VAR compensators and on-load tap-changing transformers, are limited by their operation modes and control methods. Their response speed and control accuracy are relatively low, making it difficult to meet the operation and control requirements of distribution networks under the rapid and frequent fluctuations of distributed power sources. With the rapid development of power electronics technology, new power electronic devices represented by intelligent soft open points (SOPs) are widely used in distribution networks. Through the flexible transformation and control of electric energy, the operation and scheduling potential of various new elements is explored, and the joint support of multiple source and load equipment for the operation objectives of distribution networks is realized. In addition, the energy storage system can be directly connected to the SOP through the DC bus to form an intelligent energy storage integrated soft open points (ESOP). Through the storage and release of electric energy by the energy storage system, the transfer of electric energy in time is realized. The SOP is used to adjust the power in space and control the charging and discharging of the energy storage system, thereby suppressing the impact of frequent fluctuations in the output of distributed power sources on the distribution network, effectively resisting the influence of various uncertain factors, and realizing the economic and safe operation of the distribution network. The flexible control capability of multi-terminal ESOPs lays a physical foundation for the multi-mode operation and control of distribution networks.

[0004] Traditional physical mechanism-based modeling methods are difficult to cope with the complex and changeable operating conditions of the distribution network under the uneven spatial distribution of sources and loads, and are difficult to match the fast and fine adjustment capabilities of ESOP. With the continuous upgrading of the intelligence and informatization of the distribution network, the large-scale deployment of multi-source measurement creates conditions for the application of measurement-driven methods. However, the measurement-driven control method has a high dependence on data quality, and the actual operation process often produces bad data and disturbed data, which will greatly affect the data-driven control effect. Therefore, the measurement-driven method technology that considers data quality can be used to overcome the problem of multi-mode adaptive operation and control of the distribution network under ESOP access. According to the actual operation needs of each region and the measurement feedback data, the control mode of each port of ESOP is adaptively switched to improve the operation flexibility of the distribution network. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a multi-mode adaptive operation control method for an intelligent energy storage soft switch driven by multi-source measurement. By identifying and processing bad data, the method can flexibly switch between economic operation, voltage control, load balancing and other modes online according to the actual state of the distribution network, thereby comprehensively improving the overall flexible operation control level of the distribution network.

[0006] The present invention solves the technical problem by adopting the following technical solutions:

[0007] The multi-mode adaptive operation control method of the intelligent energy storage soft switch driven by multi-source measurement includes the following steps:

[0008] Step 1: According to the selected distribution network, input the distribution network parameters.

[0009] Step 2: Obtain the bad data and measurement disturbance w during the period from t-Δt' to t t The voltage measurement data of each node and the current measurement data of each branch are used to establish a multi-source measurement matrix at time t. right Perform online identification of bad data and measurement disturbance smoothing, and calculate the actual voltage and current mean at time t after bad data processing

[0010] Step 3: Based on the measured data processed in step 2 and the control deviation at time t, the control targets and mapping matrices of each mode are constructed; the mapping prediction matrix is ​​calculated based on the distribution network operation information, historical operation data and control deviation information in the current prediction domain, and the prediction domain T is updated. p and the time step coefficient β;

[0011] Step 4: Set the ESOP coupling operation constraints. According to the mapping prediction matrix in step 3, set the auxiliary decision variables of the voltage control mode of the energy storage link. Auxiliary decision variables of load balancing control mode and build constraints;

[0012] Step 5: Set the mode switching flag for each area according to the deviation between the measurement information and the reference interval of each control target. Adaptively select the distribution network operation control mode and set the optimization target for each mode, and establish the ESOP multi-mode adaptive control model driven by multi-source measurement according to the constraints in step 4;

[0013] Step 6: Use a mathematical solver to solve the ESOP multi-mode adaptive control model driven by multi-source measurement in step 4, and issue and execute the solution strategy result;

[0014] Step 7: Update control time t=t+Δt, and determine t-t0≥βΔT c Is it true: If so, update the prediction domain, that is, let T p =T p -ΔT c , and set β = β + 1; if not, go to the next step;

[0015] Step 8: Determine whether t-t0 is greater than the optimization time T; if not, return to step 2; if so, the iterative calculation ends.

[0016] Moreover, the distribution network parameters in step 1 include the system topology connection relationship, fault location, load, type, access location, capacity and parameters of distributed power supply, capacity and parameters of energy storage intelligent soft switch, access location, distribution network operation voltage constraint and branch current limit, set the starting time to t0, the total fault duration to T, the optimization step length Δt, and the control domain ΔT c .

[0017] Moreover, the specific implementation method of online identification of bad data in step 2 is:

[0018] Set the measurement data acquisition time window Δt', and calculate the multi-source measurement data of the distribution network in the period from t-Δt' to t. Establish a multi-source measurement matrix at time t It is expressed as:

[0019]

[0020] Where z represents the number of measurement data sets uploaded during the period from t-Δt' to t, I represents the dimension of the column vector; M N Represents the number of nodes in the distribution network, M L Indicates the number of lines in the distribution network; represents the ξth group of multi-source measurement data uploaded during the period from t-Δt' to t, D w,Δt',ξ (i) and D w,Δt',ξ (j) represents the ξ-th data on the i-th node and the j-th line in the distribution network during the period from t-Δt' to t;

[0021] Defining a Multi-Source Measurement Matrix The two-dimensional coordinates of the measured data:

[0022]

[0023] In the formula, Represents a multi-source measurement matrix The standard deviation of the ξth group of data in ;

[0024] Construct measurement data D w,Δt',ξ The kth distance neighborhood of (x) is denoted as N k [D w,Δt',ξ (x)], expressed as follows:

[0025]

[0026] In the formula, the measured data D w,Δt',ξ The kth reachable distance of (x) is d k [D w,Δt',ξ (x),D w,Δt',ξ (y)], which can be expressed as follows:

[0027]

[0028] O Δt',x =D w,Δt',ξ (x)-D w,Δt',ξ (y)

[0029] In the formula, d[D w,Δt',ξ (x),D w,Δt',ξ (y)] represents the xth node in the distribution network uploading the ξth data D during the period from t-Δt' to t w,Δt',ξ (x) upload the ξth data D to the yth node w,Δt',ξ The Euclidean distance of (y), O Δt',x Indicates data D w,Δt',ξ (x) and D w,Δt',ξ The difference of (y); d k [D w,Δt',ξ(x) ] represents the kth distance, defined as w,Δt',ξ The kth distance neighborhood N of (x) k [D w,Δt',ξ(x) ]Inside and D w,Δt',ξ (x) Euclidean distance between the farthest data points;

[0030] Then measure the data D w,Δt',ξ The kth local reachability density ρ of (x) k [D w,Δt',ξ (x)], which is expressed as follows:

[0031]

[0032] Get the calculated local outlier factor:

[0033]

[0034] In the formula, F k [D w,Δt',ξ(x) ] indicates measurement data D w,Δt',ξ (x) local outlier factor;

[0035] Calculate the outlier factor threshold for multivariate measurement data mean calculation:

[0036]

[0037] In the formula, F k,max It represents the outlier factor threshold calculated by quartile analysis method, Q1 and Q3 represent the local outlier factor group F k The lower and upper quartiles of F k =[F k [D w,Δt',ξ (1)],…,F k [D w,Δt',ξ (i)],…,F k [D w,Δt',ξ (I)]],F k [D w,Δt',ξ (i)] represent the voltage measurement data D w,Δt',ξ The outlier factor of (i), Denotes the outlier factor group F k The standard deviation of , η1, η2 represent the standard deviation discrimination coefficient;

[0038] Will The data whose local outlier factor is greater than the outlier factor threshold is regarded as bad data and reset to zero to obtain the corrected multi-source measurement matrix. Represents the modified multi-source measurement matrix The ξth group of data;

[0039] According to the modified multi-source measurement matrix Calculate the average measurement data at time t as follows:

[0040]

[0041] In the formula, represents the average true value data at time t, It represents the average measurement disturbance at time t. Since the measurement data still contains measurement disturbance after the online bad data identification processing, in order to reduce the impact of measurement disturbance on the data-driven control process, it is necessary to suppress the measurement disturbance during the control process.

[0042] Moreover, the specific implementation method of the measurement disturbance suppression in step 2 is:

[0043] In view of the measurement disturbance in the actual voltage measurement data, a disturbance suppression method with an attenuation factor is used in the data-driven control scheme considering the quality of the measurement data to suppress the measurement deviation containing the measurement disturbance at time t. Perform measurement disturbance suppression processing to obtain the deviation disturbance suppression data at time t The processing process is shown as follows:

[0044]

[0045] In the formula, e w,t Indicates the node voltage reference value U ref and the average measurement data at time t The difference vector of Represents the average measurement data at time t The average measurement data within t-Δt The difference vector of ζ represents the attenuation factor, β represents the amplification factor, N c =ΔT c / Δt, represents the number of ESOP regulation in each control domain, N w Indicates that at each ΔT c Inside, from |e w,t |≥e s E phase switches to |e w,t | <e s The number of times ESOP has completed regulation in the E stage; E represents the I-dimensional unit column vector; e s is the deviation threshold parameter.

[0046] Moreover, the specific implementation method of step 3 is:

[0047] The control objectives of each mode are:

[0048]

[0049]

[0050] f UL =f U +f L (15)

[0051]

[0052] N=T p / ΔT c (23) In the formula, f U represents the voltage deviation mode control target, f L Indicates the load balancing mode control target, f P represents the control target of economic operation mode, f UL Indicates the joint operation control mode target, Indicates the target reference value of voltage deviation control at time t+Δt, It represents the estimated value of the voltage controlled variable at time t+Δt, represents the load balancing control target reference value at time t+Δt, represents the estimated value of the load balancing controlled quantity at time t+Δt, They represent the reference values ​​of economic operation control targets at time t+Δt, It represents the estimated value of the controlled quantity of network loss at time t+Δt, U t represents the voltage measurement data of each node in the distribution network at time t, O U,REF It represents the reference value of the dedimensionalized voltage deviation. It represents the voltage deviation controlled quantity at time t, represents the voltage control mode mapping matrix at time t, ΔX t Indicates the change in ESOP output at time t, I t It represents the current measurement data of each line in the distribution network at time t, I max Indicates the maximum allowable current of the line, O L,REF Indicates the reference value of load balancing dedimensionalization processing. Indicates the controlled amount of load balancing at time t, represents the load balancing control mode mapping matrix at time t, represents the controlled amount of network loss at time t, represents the mapping matrix of the economic operation control mode at time t, O P,REF It represents the reference value of dedimensionalized processing of the controlled quantity of network loss. It represents the change in the output of port r at time t in the SOP link. They represent the changes in active and reactive output of port r at time t in the SOP link, N represents the number of prediction steps set for the operation constraints of the energy storage link, and T p represents the prediction domain, ΔT c represents the control domain, Δt represents the control time interval;

[0053] The mapping of each mode is:

[0054]

[0055] Ω O={UL,U,L,P} (25)

[0056]

[0057] In the formula, {·} represents the control mode index, Ω O Represents a control mode set, UL, U, L, and P represent the joint operation control mode, voltage deviation control mode, load balancing control mode, and economic operation control mode, respectively. denote the {·} mode mapping matrix of the SOPr port at time t and t-Δt, respectively, It represents the output change of port r of SOP link at time t-Δt. and Represents the weight coefficients related to the mapping matrix solution of the SOP link, represents the {·} mode mapping estimation matrix of the energy storage link at time t, They represent the energy storage link {·} mode mapping matrix at time t, t+nΔT c ,t+(N-1)ΔT c The ESOP energy storage link {·} mode mapping estimation matrix at time, and Represents the weight coefficient related to solving the mapping matrix of the energy storage link, represents the {·} model prediction coefficient matrix at time t, T d represents the prediction coefficient step length, l represents the total number of prediction coefficient steps, α represents the prediction coefficient time index, and Respectively represent the use of t+nΔT c -T d Time and t+nΔT c -lT d The {·} mode mapping matrix is ​​calculated from the processed historical data at time t, and δ represents the weight coefficients related to the solution of the prediction coefficient matrix.

[0058] Moreover, the specific implementation method of step 4 is: setting the SOP link voltage control mode auxiliary decision variable Auxiliary decision variables of load balancing control mode And construct constraints to linearize the control objectives in step 3 based on the auxiliary decision variables;

[0059] The ESOP operation coupling constraints are:

[0060]

[0061] In the formula, It represents the charging and discharging power of the energy storage link at time t, They represent the active and reactive output of the rth port of the SOP link at time t, represents the capacity of the rth port in the SOP link, They represent the upper and lower limits of the active power of the SOP link respectively. Respectively represent the upper and lower limits of the reactive power of the SOP link, Respectively represent t+ΔT c Time and t+(N-1)ΔT c The charging and discharging power difference of the energy storage link at each moment, S ES represents the capacity of the energy storage link, Indicates the upper and lower limits of the state of charge of the energy storage link. Indicates the actual charge state of the energy storage link at time t, They represent the upper and lower limits of the calculated charging and discharging power of the energy storage link, They represent the initial state of the energy storage system and the state of charge after the operation cycle ends;

[0062] The auxiliary decision variable constraints are:

[0063]

[0064]

[0065] In the formula, are the voltage auxiliary decision variables of the energy storage link and the SOP link, Ω b represents the set of nodes in region b, i represents the i-th node in region b, represents the region set, E represents the unit column vector, represents the voltage control change vector of the energy storage link of the ith node in the Nth step prediction at time t, Represents the voltage control change vector of the SOP link of the i-th node at time t+Δt, Respectively represent the upper and lower limits of the voltage control reference interval, They are the auxiliary decision variables of the line load in the energy storage link and the SOP link, represents the line load switching threshold, i and j represent the i-th and j-th nodes in area b, represents the line load rate control change vector of the ijth line energy storage link in the Nth step prediction at time t+Δt, The line load rate control change vector of the SOP link of the ijth line at time t+Δt;

[0066] The objective function is linearized as:

[0067]

[0068] In the formula, represents the control target set of each mode at time t, represents the voltage deviation mode control target at time t, Indicates the load balancing mode control target at time t, represents the control target of the economic operation mode at time t, represents the joint operation control mode target at time t, They respectively represent the switching flags of each control mode at time t.

[0069] Moreover, the specific implementation method of step 5 is:

[0070] The control mode adaptively switches to:

[0071]

[0072] In the formula, The control mode of region b at time t is {·}.

[0073] The advantages and positive effects of the present invention are:

[0074] The present invention uses multi-source measurement to perform data modeling on the operating status of the distribution network through quality identification and processing of measurement data, thereby avoiding problems such as disturbance of bad measurement data under complex operating conditions. In addition, the multi-mode mapping matrix can use multi-source measurement data to construct adaptive predictive control models under various modes according to the control mode requirements of the distribution network, thereby providing support for realizing multi-mode operation control of the distribution network. On this basis, the power transfer-storage-supply capability of ESOP is used to realize adaptive switching of control modes according to the control requirements of different regions, timely judge and adapt to the power requirements of different regions during actual operation, improve the flexibility of distribution network operation control, and quickly respond to rapid fluctuations in distributed power output and load. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 This is a flow chart of the multi-mode adaptive operation control method of the intelligent energy storage soft switch driven by multi-source measurement of the present invention;

[0076] Figure 2 This is a structural diagram of an actual distribution network calculation example including a four-terminal ESOP according to an embodiment of the present invention;

[0077] Figure 3 This is a photovoltaic and wind turbine fluctuation curve diagram of an embodiment of the present invention;

[0078] Figure 4 This is a load fluctuation curve diagram of an embodiment of the present invention;

[0079] Figure 5 It is a SOP active output strategy curve diagram of an embodiment of the present invention;

[0080] Figure 6 It is a SOP reactive power output strategy curve diagram of an embodiment of the present invention;

[0081] Figure 7 This is a curve diagram of the charge state change of the energy storage link in an embodiment of the present invention;

[0082] Figure 8 This is a diagram showing the change of control modes in each area of ​​the distribution network in this embodiment;

[0083] Fig. 9 This is a comparison diagram of load balancing levels of distribution network lines according to an embodiment of the present invention;

[0084] Fig.10 This is a comparison diagram of the operating loss levels of the distribution network according to an embodiment of the present invention;

[0085] Fig.11 This is a comparison diagram of the maximum global voltage results of an embodiment of the present invention;

[0086] Fig.12 This is a comparison diagram of the global voltage minimum value results of an embodiment of the present invention. DETAILED DESCRIPTION

[0087] The present invention is further described in detail below with reference to the accompanying drawings.

[0088] like Figure 1 As shown, the multi-mode adaptive operation control method of the intelligent energy storage soft switch under multi-source measurement drive of the present invention includes the following steps:

[0089] 1) According to the selected distribution network, input the distribution network parameters, including: system topology connection relationship, fault location, load, type of distributed power source, access location, capacity and parameters, energy storage integrated soft open point (ESOP) capacity and parameters, access location, distribution network operation voltage constraint and branch current limit, set the starting time to t0, the total fault duration to T, the optimization step size Δt, and the control domain ΔT c ;

[0090] 2) Obtain the bad data and measurement disturbance w during the period from t-Δt' to t t The voltage measurement data of each node and the current measurement data of each branch are used to establish a multi-source measurement matrix at time t. right Perform online identification of bad data and measurement disturbance smoothing, and calculate the actual voltage and current mean at time t after bad data processing in

[0091] The bad data online identification is as follows:

[0092] Set the measurement data acquisition time window Δt', and calculate the multi-source measurement data of the distribution network in the period from t-Δt' to t. Establish a multi-source measurement matrix at time t It is expressed as:

[0093]

[0094] Where z represents the number of measurement data sets uploaded during the period from t-Δt' to t, I represents the dimension of the column vector; M N Represents the number of nodes in the distribution network, M L Indicates the number of lines in the distribution network; represents the ξth group of multi-source measurement data uploaded during the period from t-Δt' to t, D w,Δt',ξ (i) and D w,Δt',ξ (j) represents the ξ-th data on the i-th node and the j-th line in the distribution network during the period from t-Δt' to t;

[0095] Defining a Multi-Source Measurement Matrix The two-dimensional coordinates of the measured data:

[0096]

[0097] In the formula, Represents a multi-source measurement matrix The standard deviation of the ξth group of data in ;

[0098] Construct measurement data D w,Δt',ξ The kth distance neighborhood of (x) is denoted as N k [D w,Δt',ξ (x)], expressed as follows:

[0099]

[0100] In the formula, the measured data D w,Δt',ξ The kth reachable distance of (x) is d k [D w,Δt',ξ (x),D w,Δt',ξ (y)], which can be expressed as follows:

[0101]

[0102] O Δt',x =D w,Δt',ξ (x)-D w,Δt',ξ (y)

[0103] In the formula, d[D w,Δt',ξ (x),D w,Δt',ξ (y)] represents the xth node in the distribution network uploading the ξth data D during the period from t-Δt' to t w,Δt',ξ (x) upload the ξth data D to the yth node w,Δt',ξ The Euclidean distance of (y), O Δt',x Indicates data Dw,Δt',ξ (x) and D w,Δt',ξ The difference of (y); d k [D w,Δt',ξ (x)] represents the kth distance, defined as w,Δt',ξ The kth distance neighborhood N of (x) k [D w,Δt',ξ (x)]with D w,Δt',ξ (x) Euclidean distance between the farthest data points;

[0104] Then measure the data D w,Δt',ξ The kth local reachability density ρ of (x) k [D w,Δt',ξ (x)], which can be expressed as follows:

[0105]

[0106] Get the calculated local outlier factor:

[0107]

[0108] In the formula, F k [D w,Δt',ξ (x)] represents the measurement data D w,Δt',ξ (x) local outlier factor;

[0109] Calculate the outlier factor threshold for multivariate measurement data mean calculation:

[0110]

[0111] In the formula, F k,max It represents the outlier factor threshold calculated by quartile analysis method, Q1 and Q3 represent the local outlier factor group F k The lower and upper quartiles of F k =[F k [D w,Δt',ξ (1)],…,F k [D w,Δt',ξ (i)],…,F k [D w,Δt',ξ (I)]],F k [D w,Δt',ξ (i)] represent the voltage measurement data D w,Δt',ξ The outlier factor of (i), Denotes the outlier factor group F k The standard deviation of , η1, η2 represent the standard deviation discrimination coefficient;

[0112] Will The data whose local outlier factor is greater than the outlier factor threshold is regarded as bad data and reset to zero to obtain the corrected multi-source measurement matrix. Represents the modified multi-source measurement matrix The first group of data.

[0113] According to the modified multi-source measurement matrix Calculate the average measurement data at time t as follows:

[0114]

[0115] In the formula, represents the average true value data at time t, Represents the average measurement disturbance at time t. Since the measurement data still contains measurement disturbance after online bad data identification, in order to reduce the impact of measurement disturbance on the data-driven control process, it is necessary to suppress the measurement disturbance during the control process;

[0116] The measurement disturbance suppression is as follows:

[0117] In view of the measurement disturbance in the actual voltage measurement data, a disturbance suppression method with an attenuation factor is used in the data-driven control scheme considering the quality of the measurement data to suppress the measurement deviation containing the measurement disturbance at time t. Perform measurement disturbance suppression processing to obtain the deviation disturbance suppression data at time t The processing process is shown as follows:

[0118]

[0119]

[0120] In the formula, e w,t Indicates the node voltage reference value U ref and the average measurement data at time t The difference vector of Represents the average measurement data at time t The average measurement data within t-Δt The difference vector of ζ represents the attenuation factor, β represents the amplification factor, N c =ΔT c / Δt, represents the number of ESOP regulation in each control domain, N w Indicates that at each ΔT c Inside, from |e w,t |≥e s E phase switches to |e w,t | <e s The number of times ESOP has completed regulation in the E stage; E represents the I-dimensional unit column vector; e s is the deviation threshold parameter;

[0121] 3) According to the measured data processed in step 2) and the control deviation at time t, the control targets and mapping matrices of each mode are constructed; the mapping prediction matrix is ​​calculated according to the distribution network operation information, historical operation data and control deviation information in the current prediction domain, and the prediction domain T is updated. p and the time step coefficient β; where,

[0122] The control objectives of each mode are as follows:

[0123]

[0124] f UL =f U +f L (15)

[0125]

[0126] N=T p / ΔT c (23) In the formula, f U represents the voltage deviation mode control target, f L Indicates the load balancing mode control target, f P represents the control target of economic operation mode, f UL Indicates the joint operation control mode target, Indicates the target reference value of voltage deviation control at time t+Δt, It represents the estimated value of the voltage controlled variable at time t+Δt, represents the load balancing control target reference value at time t+Δt, represents the estimated value of the load balancing controlled quantity at time t+Δt, They represent the reference values ​​of economic operation control targets at time t+Δt, It represents the estimated value of the controlled quantity of network loss at time t+Δt, U t represents the voltage measurement data of each node in the distribution network at time t, O U,REF It represents the reference value of the dedimensionalized voltage deviation. It represents the voltage deviation controlled quantity at time t, represents the voltage control mode mapping matrix at time t, ΔX t Indicates the change in ESOP output at time t, I t It represents the current measurement data of each line in the distribution network at time t, I max Indicates the maximum allowable current of the line, O L,REF Indicates the reference value of load balancing dedimensionalization processing. Indicates the controlled amount of load balancing at time t, represents the load balancing control mode mapping matrix at time t, represents the controlled amount of network loss at time t, represents the mapping matrix of the economic operation control mode at time t, O P,REF It represents the reference value of dedimensionalized processing of the controlled quantity of network loss. It represents the change in the output of port r at time t in the SOP link. They represent the changes in active and reactive output of port r at time t in the SOP link, N represents the number of prediction steps set for the operation constraints of the energy storage link, and T p represents the prediction domain, ΔT c represents the control domain, Δt represents the control time interval;

[0127] The mapping of each mode is expressed as:

[0128]

[0129] In the formula, {·} represents the control mode index, Ω O Represents a control mode set, UL, U, L, and P represent the joint operation control mode, voltage deviation control mode, load balancing control mode, and economic operation control mode, respectively. denote the {·} mode mapping matrix of the SOPr port at time t and t-Δt, respectively, It represents the output change of port r of SOP link at time t-Δt. and Represents the weight coefficients related to the mapping matrix solution of the SOP link, represents the {·} mode mapping estimation matrix of the energy storage link at time t, They represent the energy storage link {·} mode mapping matrix at time t, t+nΔT c ,t+(N-1)ΔT c The ESOP energy storage link {·} mode mapping estimation matrix at time, and Represents the weight coefficient related to solving the mapping matrix of the energy storage link, represents the {·} model prediction coefficient matrix at time t, T d represents the prediction coefficient step length, l represents the total number of prediction coefficient steps, α represents the prediction coefficient time index, and Respectively represent the use of t+nΔT c -T d Time and t+nΔT c -lT d The {·} mode mapping matrix is ​​calculated from the processed historical data at time t, and δ represents the weight coefficients related to the solution of the prediction coefficient matrix;

[0130] 4) Set the ESOP coupling operation constraints, and set the auxiliary decision variables of the voltage control mode of the energy storage link according to the mapping prediction matrix described in step 3). Auxiliary decision variables of load balancing control mode And build constraints; set the auxiliary decision variables of the voltage control mode of the SOP link Auxiliary decision variables of load balancing control mode And construct constraints, and linearize the control objectives in step 3) according to the auxiliary decision variables; where,

[0131] The ESOP operation coupling constraint is expressed as:

[0132]

[0133]

[0134] In the formula, It represents the charging and discharging power of the energy storage link at time t, They represent the active and reactive output of the rth port of the SOP link at time t, represents the capacity of the rth port in the SOP link, They represent the upper and lower limits of the active power of the SOP link respectively. Respectively represent the upper and lower limits of the reactive power of the SOP link, Respectively represent t+ΔT c Time and t+(N-1)ΔT c The charging and discharging power difference of the energy storage link at each moment, S ES represents the capacity of the energy storage link, Indicates the upper and lower limits of the state of charge of the energy storage link. Indicates the actual charge state of the energy storage link at time t, They represent the upper and lower limits of the calculated charging and discharging power of the energy storage link, They represent the initial state of the energy storage system and the state of charge after the operation cycle ends;

[0135] The auxiliary decision variable constraints are as follows:

[0136]

[0137]

[0138] In the formula, are the voltage auxiliary decision variables of the energy storage link and the SOP link, Ω b represents the set of nodes in region b, i represents the i-th node in region b, represents the region set, E represents the unit column vector, represents the voltage control change vector of the energy storage link of the ith node in the Nth step prediction at time t, Represents the voltage control change vector of the SOP link of the i-th node at time t+Δt, Respectively represent the upper and lower limits of the voltage control reference interval, They are the auxiliary decision variables of the line load in the energy storage link and the SOP link, represents the line load switching threshold, i and j represent the i-th and j-th nodes in area b, represents the line load rate control change vector of the ijth line energy storage link in the Nth step prediction at time t+Δt, The line load rate control change vector of the SOP link of the ijth line at time t+Δt;

[0139] The objective function is linearized as follows:

[0140]

[0141] In the formula, represents the control target set of each mode at time t, represents the voltage deviation mode control target at time t, Indicates the load balancing mode control target at time t, represents the control target of the economic operation mode at time t, represents the joint operation control mode target at time t, Respectively represent the switch flags of each control mode at time t;

[0142] 5) According to the deviation between the measurement information and the reference interval of each control target, set the mode switching flag of each area Adaptively select the distribution network operation control mode and set the optimization target of each mode, and establish a multi-source measurement-driven ESOP multi-mode adaptive control model according to the constraints described in step 4);

[0143] The control mode adaptive switching is as follows:

[0144]

[0145] In the formula, The control mode of region b at time t is {·}.

[0146] 6) using a mathematical solver to solve the ESOP multi-mode adaptive control model driven by the multi-source measurement in step 4), and issuing and executing the solution strategy result;

[0147] 7) Update control time t = t + Δt, and judge t-t0 ≥ βΔT c Is it true: If so, update the prediction domain, that is, let Tp =T p -ΔT c , and set β = β + 1; if not, go to the next step;

[0148] 8) Determine whether t-t0 is greater than the optimization time T; if not, return to step 2); if so, the iterative calculation ends.

[0149] Here are some specific examples:

[0150] For the example of the present invention, the distribution network includes 44 nodes, and the topological connection is as follows: Figure 2 As shown, the four ends of the SOP are connected to nodes 5, 14, 38, and 43 respectively. The four feeders are interconnected through the four-terminal ESOP. The voltage level is 10.5kV. The total active load and total reactive load of the distribution network are 9.99MW and 7.34Mvar respectively. In the calculation example of the flexible interconnected distribution network with four-terminal ESOP, three photovoltaic systems and three wind turbines are connected respectively, and the penetration rate of distributed power sources reaches 90.11%. Among them, the access locations of the photovoltaic systems are nodes 8, 32, and 36, and the capacity is 1000kWp; the access locations of the wind turbines are nodes 12, 21, and 39, and the capacity is 2000kVA. The daily output curves of distributed power sources and the fluctuation curves of various types of loads are shown as follows. Figure 4 and Figure 5 As shown in the figure, the converter capacity of each port of the four-terminal ESOP is 3MVA, the DC side is connected to 0.5MW / 2MWh energy storage, and the state of charge range is [0.2, 0.8]. The time parameters and mode switching parameters are set as follows: the optimization time T is set to 24 hours, and the prediction domain T p Set to 24 hours, control domain ΔT c Set to 1 hour, control time interval Δt to 5 minutes, measurement data acquisition time window Δt' to 0.5 minutes; voltage control mode switching interval per unit value is set to [0.98,1.02]pu, load balancing control mode switching flag is The measurement data is set to have a measurement disturbance with a deviation of ±2%, and the measurement bad data with a deviation of ±10% is randomly generated. In order to ensure the convergence of the multi-source measurement driven adaptive operation control method, the control parameters are set as follows: δ is set to 1. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement is used for optimization. After the above steps, the ESOP control strategy can be obtained. To verify the effectiveness of this method, the following three scenarios are used for comparison in the distribution system:

[0151] Scenario 1: The initial state of the distribution network is obtained without using any control method;

[0152] Scenario 2: Using the method proposed in the present invention to measure and drive multi-mode operation control of ESOP;

[0153] Scenario 3: Using a centralized operation optimization method based on mechanism modeling to control ESOP.

[0154] The computer hardware environment for performing the optimization calculation is Intel(R) Core(TM) CPU i5-10210U, with a main frequency of 1.6GHz and a memory of 16GB; the software environment is Windows11 operating system. The example test is based on the Matlab 2019a platform, and the above-mentioned ESOP measurement drive multi-mode operation control program is developed, and the CVX commercial solver software package is called for solving. The example topology used in the embodiment of the present invention is as follows Figure 2 The comparison of various indicators of multi-mode operation control is shown in Table 1. The prediction curve changes of distributed power output and load information are shown in Figure 3 and Figure 4 ESOP output strategy is as follows. Figure 5-Figure 7 The control mode changes in each area of ​​the distribution network are shown in Figure 8 As shown in Figure 2. The comparison of load balancing levels of distribution network lines is shown in Figure 2. Fig. 9 As shown in Figure 2. The comparison of distribution network operation loss levels is shown in Figure 2. Fig.10 As shown in the figure. The global voltage comparison is Fig.11 and Fig.12 shown.

[0155] For the multi-mode control effect of multi-source measurement drive, combined with Figure 3 and Figure 4 From the distributed power and load operation curves in Scheme 1, it can be seen that from 1:00 to 3:00 in the morning, due to the fluctuation of distributed power output and the change of load levels in various regions, the node voltage exceeds the limit. Figure 2 From the deployment of distributed power sources, it can be seen that there is no wind turbine system deployed in area 1 and area 4. The load in area 1 is large and the voltage is lower than the lower limit. Therefore, it is in the voltage deviation control mode from 0:00 to 1:05, and at 1:10, it is measured that the voltage is adjusted to the reference range, so it is quickly switched to the economic operation control mode; Figure 3 During the period from 0:00 to 2:00, the photovoltaic system does not produce power but the wind turbine system produces a large power. Figure 8 In (a), area 3 is in a switching state between voltage deviation and load balancing mode because the voltage exceeds the upper limit, and area 4 is in load balancing control mode because the load is small.

[0156] Fig. 9It can be seen that the overall change of the distribution network line load in the three schemes verifies the continuous adjustment ability of ESOP on the line load, indicating that the control effect of the distribution network load level by the method proposed in this patent under the load balancing control mode is close to that of the centralized method. Fig.10 Regarding the level of distribution network operation losses, it can be seen that the network losses of Scheme 2 are significantly reduced compared to Scheme 1 and are close to the optimal control effect of Scheme 3. Fig.11 and Fig.12 It indicates that the method proposed in this patent also has a good voltage control effect.

[0157] According to the data in Table 1, through the regulation of the method proposed in the present invention, compared with scheme 1, the minimum voltage is increased by 0.247pu, the maximum voltage is reduced by 0.426pu, the voltage deviation index is reduced by 45.26%, the load balancing index is reduced by 32.59%, and the network loss is reduced by 45.81%.

[0158] In summary, combined with Table 1 and Figures 5 to 12 It can be seen that the method proposed in the present invention can effectively process and identify bad data, adaptively switch control modes according to the operation requirements of the distribution network, avoid the problem of insufficient flexibility of measurement-driven single-target control, and ensure efficient operation of the distribution network; by comparing with the indicators in Scheme 3, the multi-source measurement-driven ESOP multi-mode operation control method proposed in the present invention can achieve a regulation effect that is close to the global optimality for the operation status of the distribution network.

[0159] Table 1 Comparison of various indicators of multi-mode operation control

[0160]

[0161] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation manner. Any other implementation manners derived by those skilled in the art based on the technical solution of the present invention also fall within the scope of protection of the present invention.

Claims

1. A multi-mode adaptive operation control method for intelligent energy storage soft switch driven by multi-source measurement, characterized in that: The following steps are involved: Step 1: Input the distribution network parameters according to the selected distribution network. Step 2: Obtain the bad data and measurement disturbance w during the period from t-Δt' to t t The voltage measurement data of each node and the current measurement data of each branch are used to establish a multi-source measurement matrix at time t. right Perform online identification of bad data and measurement disturbance smoothing, and calculate the actual voltage and current mean at time t after bad data processing Step 3: Based on the measured data processed in step 2 and the control deviation at time t, the control target and mapping matrix of each mode are constructed; the mapping prediction matrix is ​​calculated based on the distribution network operation information, historical operation data and control deviation information in the current prediction domain, and the prediction domain Tp and the time step coefficient β are updated; Step 4: Set the ESOP coupling operation constraints. According to the mapping prediction matrix in step 3, set the auxiliary decision variables of the voltage control mode of the energy storage link. Auxiliary decision variables of load balancing control mode and build constraints; Step 5: Set the mode switching flag for each area according to the deviation between the measurement information and the reference interval of each control target. Adaptively select the distribution network operation control mode and set the optimization target for each mode, and establish the ESOP multi-mode adaptive control model driven by multi-source measurement according to the constraints in step 4; Step 6: Use a mathematical solver to solve the ESOP multi-mode adaptive control model driven by multi-source measurement in step 4, and issue and execute the solution strategy result; Step 7: Update control time t=t+Δt, and determine t-t0≥βΔT c Is it true: If so, update the prediction domain, that is, let T p =T p -ΔT c , and set β = β + 1; if not, go to the next step; Step 8: Determine whether t-t0 is greater than the optimization time T; if not, return to step 2; if so, the iterative calculation ends.

2. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement according to claim 1 is characterized by: The distribution network parameters in step 1 include the system topology connection relationship, fault location, load, type, access location, capacity and parameters of distributed power sources, capacity and parameters of energy storage intelligent soft switches, access location, distribution network operation voltage constraints and branch current limits, set the starting time to t0, the total fault duration to T, the optimization step length Δt, and the control domain ΔT c .

3. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement according to claim 1 is characterized in that: The specific implementation method of online identification of bad data in step 2 is: Set the measurement data acquisition time window Δt', and calculate the multi-source measurement data of the distribution network in the period from t-Δt' to t. Establish a multi-source measurement matrix at time t It is expressed as: Where z represents the number of measurement data sets uploaded during the period from t-Δt' to t, I represents the dimension of the column vector; M N Represents the number of nodes in the distribution network, M L Indicates the number of lines in the distribution network; represents the ξth group of multi-source measurement data uploaded during the period from t-Δt′ to t, D w,Δt′，ξ (i) and D w,Δt′，ξ (j) represents the ξ-th data on the i-th node and the j-th line in the distribution network during the period from t-Δt′ to t; Defining a Multi-Source Measurement Matrix The two-dimensional coordinates of the measured data: In the formula, Represents a multi-source measurement matrix The standard deviation of the ξth group of data in ; Construct measurement data D w,Δt′，ξ The kth distance neighborhood of (x) is denoted as N k [D w，Δt′，ξ (x)], expressed as follows: In the formula, the measured data D w,Δt′，ξ The kth reachable distance of (x) is d k [D w，Δt′，ξ (x), D w,Δt′，ξ (y)], which can be expressed as follows: In the formula, d[D w，Δt′，ξ (x), D w，Δt′，ξ (y)] represents the xth node in the distribution network uploading the ξth data D during the period from t-Δt′ to t w,Δt′，ξ (x) upload the ξth data D to the yth node w,Δt′，ξ The Euclidean distance of (y), O Δt′，x Indicates data D w,Δt′，ξ (x) and D w,Δt′，ξ The difference of (y); d k [D w，Δt′，ξ (x)] represents the kth distance, defined as w,Δt′，ξ The kth distance neighborhood N of (x) k [D w，Δt′，ξ (x)]with D w,Δt′，ξ (x) Euclidean distance to the farthest data point; Then measure the data D w,Δt′，ξ The kth local reachability density ρ of (x) k [D w，Δt′，ξ (x)], which is expressed as follows: Get the calculated local outlier factor: In the formula, F k [D w，Δt′，ξ (x)] represents the measurement data D w,Δt′，ξ (x) local outlier factor; Calculate the outlier factor threshold for multivariate measurement data mean calculation: In the formula, F k，max It represents the outlier factor threshold calculated by quartile analysis method, Q1 and Q3 represent the local outlier factor group F k The lower and upper quartiles of F k =[F k [D w，Δt′，ξ (1)],…,F k [D w，Δt′，ξ (i)],…,F k [D w，Δt′，ξ (I)]],F k [D w，Δt′，ξ (i)] represent the voltage measurement data D w，Δt′，ξ The outlier factor of (i), Denotes the outlier factor group F k The standard deviation of , η1, η2 represent the standard deviation discrimination coefficient; Will The data whose local outlier factor is greater than the outlier factor threshold is regarded as bad data and reset to zero to obtain the corrected multi-source measurement matrix. Represents the modified multi-source measurement matrix The ξth group of data; According to the modified multi-source measurement matrix Calculate the average measurement data at time t as follows: In the formula, represents the average true value data at time t, Represents the average measurement disturbance at time t. Since the measurement data still contains measurement disturbance after online bad data identification, in order to reduce the impact of measurement disturbance on the data-driven control process, it is necessary to suppress the measurement disturbance during the control process.

4. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement according to claim 3 is characterized by: The specific implementation method of the measurement disturbance suppression in step 2 is: In view of the measurement disturbance in the actual voltage measurement data, a disturbance suppression method with an attenuation factor is used in the data-driven control scheme considering the quality of the measurement data to suppress the measurement deviation containing the measurement disturbance at time t. Perform measurement disturbance suppression processing to obtain the deviation disturbance suppression data at time t The processing process is shown as follows: In the formula, e w，t Indicates the node voltage reference value U ref and the average measurement data at time t The difference vector of Represents the average measurement data at time t The average measurement data within t-Δt The difference vector of represents the attenuation factor, β represents the amplification factor, N c =ΔT c / Δt, represents the number of ESOP regulation in each control domain, N w Indicates that at each ΔT c Inside, from |e w，t |≥e s E phase switches to |e w，t |<e s The number of times ESOP has completed regulation in the E stage; E represents the unit column vector of dimension; e s is the deviation threshold parameter.

5. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement according to claim 4 is characterized in that: The specific implementation method of step 3 is: The control objectives of each mode are: f UL =f U +f L (15) N=T p / ΔT c (23) In the formula, f U represents the voltage deviation mode control target, f L Indicates the load balancing mode control target, f P represents the control target of economic operation mode, f UL Indicates the joint operation control mode target, Indicates the target reference value of voltage deviation control at time t+Δt, It represents the estimated value of the voltage controlled variable at time t+Δt, represents the load balancing control target reference value at time t+Δt, represents the estimated value of the load balancing controlled quantity at time t+Δt, They represent the reference values ​​of economic operation control targets at time t+Δt, It represents the estimated value of the controlled quantity of network loss at time t+Δt, U t represents the voltage measurement data of each node in the distribution network at time t, O U,REF It represents the reference value of the dedimensionalized voltage deviation. It represents the voltage deviation controlled quantity at time t, represents the voltage control mode mapping matrix at time t, ΔX t Indicates the change in ESOP output at time t, I t It represents the current measurement data of each line in the distribution network at time t, I max Indicates the maximum allowable current of the line, O L，REF Indicates the reference value of load balancing dedimensionalization processing. Indicates the controlled amount of load balancing at time t, represents the load balancing control mode mapping matrix at time t, represents the controlled amount of network loss at time t, represents the mapping matrix of the economic operation control mode at time t, O P，REF It represents the reference value of dedimensionalized processing of the controlled quantity of network loss. It represents the change in the output of port r at time t in the SOP link. They represent the changes in active and reactive output of port r at time t in the SOP link, N represents the number of prediction steps set for the operation constraints of the energy storage link, and T p represents the prediction domain, ΔT c represents the control domain, Δt represents the control time interval; The mapping of each mode is: Ω0={UL,U,L,P} (25) In the formula, {·} represents the control mode index, Ω0 represents the control mode set, UL, U, L, and P represent the joint operation control mode, voltage deviation control mode, load balancing control mode, and economic operation control mode, respectively. denote the {·} mode mapping matrix of the SOP r port at time t and t-Δt, respectively, It represents the output change of port r of SOP link at time t-Δt. and Represents the weight coefficients related to the mapping matrix solution of the SOP link, represents the {·} mode mapping estimation matrix of the energy storage link at time t, They represent the energy storage link {·} mode mapping matrix at time t, t+nΔT c ,t+(N-1)ΔT c The ESOP energy storage link {·} mode mapping estimation matrix at time, and Represents the weight coefficient related to solving the mapping matrix of the energy storage link, represents the {·} model prediction coefficient matrix at time t, T d represents the prediction coefficient step length, l represents the total number of prediction coefficient steps, α represents the prediction coefficient time index, and Respectively represent the use of t+nΔT c -T d Time and t+nΔT c -lT d The {·} mode mapping matrix is ​​calculated from the processed historical data at time t, and δ represents the weight coefficients related to the solution of the prediction coefficient matrix.

6. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement according to claim 5 is characterized in that: The specific implementation method of step 4 is: setting the auxiliary decision variable of the voltage control mode of the SOP link Auxiliary decision variables of load balancing control mode And construct constraints to linearize the control objectives in step 3 based on the auxiliary decision variables; The ESOP operation coupling constraints are: In the formula, It represents the charging and discharging power of the energy storage link at time t, They represent the active and reactive output of the rth port of the SOP link at time t, represents the capacity of the rth port in the SOP link, They represent the upper and lower limits of the active power of the SOP link respectively. Respectively represent the upper and lower limits of the reactive power of the SOP link, Respectively represent t+ΔT c Time and t+(N-1)ΔT c The charging and discharging power difference of the energy storage link at each moment, S ES represents the capacity of the energy storage link, Indicates the upper and lower limits of the state of charge of the energy storage link. Indicates the actual charge state of the energy storage link at time t, They represent the upper and lower limits of the calculated charging and discharging power of the energy storage link, They represent the initial state of the energy storage system and the state of charge after the operation cycle ends; The auxiliary decision variable constraints are: In the formula, are the voltage auxiliary decision variables of the energy storage link and the SOP link, Ω b represents the set of nodes in region b, i represents the i-th node in region b, represents the region set, E represents the unit column vector, represents the voltage control change vector of the energy storage link of the ith node in the Nth step prediction at time t, Represents the voltage control change vector of the SOP link of the i-th node at time t+Δt, Respectively represent the upper and lower limits of the voltage control reference interval, They are the auxiliary decision variables of the line load in the energy storage link and the SOP link, represents the line load switching threshold, i and j represent the i-th and j-th nodes in area b, represents the line load rate control change vector of the ijth line energy storage link in the Nth step prediction at time t+Δt, The line load rate control change vector of the SOP link of the ijth line at time t+Δt; The objective function is linearized as: In the formula, represents the control target set of each mode at time t, represents the voltage deviation mode control target at time t, Indicates the load balancing mode control target at time t, represents the control target of the economic operation mode at time t, represents the joint operation control mode target at time t, They respectively represent the switching flags of each control mode at time t.

7. The multi-mode adaptive operation control method of intelligent energy storage soft switch driven by multi-source measurement according to claim 6 is characterized by: The specific implementation method of step 5 is: The control mode adaptively switches to: In the formula, The control mode of region b at time t is {·}.

Images