A novel power distribution system rolling optimization scheduling method, system, device and medium considering multi-time scale coupling

Abstract

The application provides a novel power distribution system rolling optimization scheduling method, system, equipment and medium considering multi-time scale coupling, the method comprises the following steps: collecting power distribution system related parameters, and performing multi-time scale load and new energy output prediction to obtain multi-time scale load prediction value and new energy output prediction value; based on the power distribution system related parameters, the multi-time scale load prediction value and the new energy output prediction value, under the constraint of the cross-time scale coupling constraint model, a distributed solving framework based on the improved alternating direction multiplier method is used to solve the multi-objective rolling optimization scheduling model to obtain a novel power distribution system scheduling scheme; the novel power distribution system scheduling scheme is executed in a rolling time domain, and a scheduling control instruction is output; the application successfully constructs a multi-stage coupling rolling optimization architecture in view of the weak coupling of the scheduling models at all levels in the existing power distribution network scheduling, introduces a cross-time scale coupling constraint, and effectively suppresses the problems of instruction disconnection and deviation accumulation.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching and operation control technology, specifically to a novel rolling optimization dispatching method, system, equipment, and medium for distribution systems that takes into account multi-timescale coupling. Background Technology

[0002] The high proportion of new power sources such as distributed photovoltaics, wind power, and electric vehicles integrated into the distribution network is driving the transformation of traditional distribution networks into new distribution systems with deep coupling of "source-grid-load-storage". However, the randomness of new energy output and the uncertainty of load pose serious challenges to the safe and economical operation of the distribution network. Multi-timescale coordinated scheduling has become the core technical framework for addressing this challenge, typically divided into three scales: day-ahead, intraday, and real-time, using progressive refinement and rolling optimization to smooth out fluctuations.

[0003] The existing technical methods have the following shortcomings: (1) The coupling mechanism between multiple time scales is weak, and the scheduling instructions are distorted at each level. Existing methods mostly adopt an architecture of independent modeling and serial execution of "day-to-day-real-time", with each scale only connected by simple data transmission, lacking systematic and quantifiable coupling constraints. This results in day-to-day plans being unable to effectively guide intraday adjustments, and intraday corrections being unable to accurately support real-time responses, ultimately causing scheduling instructions to become disconnected at each level, the accumulation of actual execution deviations, and even the system running beyond its limits.

[0004] (2) Insufficient coordination capacity across time scales and lack of differentiated regulation mechanisms. The response characteristics and regulation costs of regulation resources with different response speeds differ significantly at time scales such as day-ahead, intraday, and real-time. Existing methods have failed to establish a hierarchical coordination architecture based on resource response speed, often incorporating all resources into a unified optimization model, leading to frequent adjustments of slow resources, accelerated aging, and shortened lifespan, while fast resources over-reserve capacity at long time scales, resulting in wasted or insufficient regulation capacity.

[0005] (3) Limited multi-objective optimization capabilities make it difficult to balance economy, safety, and equipment lifespan. Most current scheduling methods focus on a single safety indicator as the core optimization objective, neglecting the coordinated optimization of key indicators such as system operating economy, energy storage cycle life loss, and tie-line power smoothness. This results in scheduling results that, while meeting safety constraints, may lead to excessively high operating costs, accelerated energy storage aging, or severe power fluctuations between the grid and the upstream grid.

[0006] (4) There is a prominent contradiction between solution efficiency and privacy protection. The scheduling problem for new power distribution systems usually needs to deal with discrete or non-convex constraints of a large number of distributed devices. The centralized solution model is large in scale and faces communication bandwidth bottlenecks and computational pressure on the central node. Although existing distributed optimization methods can alleviate the communication pressure to some extent, they have slow convergence speed, are sensitive to parameter settings, and still have the risk of privacy leakage in scenarios with multiple participants. How to achieve strict information privacy protection while ensuring computational efficiency remains an urgent technical problem to be solved.

[0007] Therefore, there is an urgent need for a new rolling optimization scheduling method that fully considers multi-timescale coupling, achieves resource coordination, and balances economy and security. Summary of the Invention

[0008] To address the problems of weak coupling between scheduling models at different levels, insufficient resource coordination, lack of multi-objective optimization, and low solution efficiency in existing distribution network scheduling technologies, this invention proposes a novel rolling optimization scheduling method for distribution systems that considers multi-timescale coupling, including: Collect relevant parameters of the power distribution system and perform load and renewable energy output forecasts at multiple time scales to obtain load forecasts and renewable energy output forecasts at multiple time scales. Based on the relevant parameters of the power distribution system, the load forecast values ​​at multiple time scales, and the power output forecast values ​​of new energy sources, under the constraints of the cross-time scale coupled constraint model, a distributed solution framework based on the improved alternating direction multiplier method is used to solve the multi-objective rolling optimization scheduling model, resulting in a new power distribution system scheduling scheme. The novel power distribution system scheduling scheme is executed in the rolling time domain, and scheduling control commands are output. Among them, the cross-timescale coupled constraint model couples a novel power distribution system scheduling scheme with different timescales; the multi-objective rolling optimization scheduling model achieves adaptive adjustment of weight coefficients through the entropy weight method; the improved alternating direction multiplier method adaptively adjusts the penalty parameter according to the original residual and the dual residual and introduces the super-relaxation technique in the iteration process.

[0009] Preferably, the multiple time scales include day-ahead scale, intraday scale, and real-time scale. The process of collecting relevant parameters of the power distribution system and performing multi-time scale load and renewable energy output forecasts to obtain multi-time scale load forecasts and renewable energy output forecasts includes: Collect distribution network topology parameters, historical data of distributed power sources and loads, meteorological data, and energy storage status charge data as relevant parameters of the distribution system; Based on the Long Short-Term Memory Network, the system takes into account the weather forecast for the first hour of the future and historical data for the same period, and outputs the day-ahead point prediction curves of new energy sources and base load as the day-ahead scale load prediction values ​​and new energy output prediction values. Using probability box theory and combining the latest measured meteorological and load data, the uncertainty intervals of new energy output and load for the second time period in the future are generated as intraday load forecasts and new energy output forecasts. Every preset period, based on the most recent third-longest high-frequency data, time series decomposition combined with conformal quantile regression is used to output the power prediction values ​​of ultra-short-term load and renewable energy output for a set future time period, which are used as the real-time load prediction values ​​and renewable energy output prediction values. The lengths of the first duration, the second duration, and the third duration decrease sequentially.

[0010] Preferably, the cross-timescale coupling constraint model includes: energy storage state of charge time coupling constraint, slow regulation device action number constraint, tie line power coupling constraint, and three-level decision variable coupling relationship constraint; The calculation formula for the time coupling constraint of the energy storage state of charge is as follows:

[0011] Among them, SOC It is the state of charge at time t. It is the state of charge at time t-1. It's about charging efficiency. It is the charging power during time period t. It is the discharge power during time period t. It is the discharge efficiency. It refers to the duration of the time period; The formula for calculating the constraint on the number of actions of the slow-speed adjustment device is as follows:

[0012] Among them, u The equipment status is adjusted slowly during time period t. It is a slow adjustment of the equipment status at time t-1. This is the maximum number of actions allowed. The formula for calculating the tie-line power coupling constraint is as follows:

[0013] in, It is the actual power of the tie line in time period t. The planned power output for the connecting line is t days prior. It is the proportional coefficient for allowable deviation; The calculation formula for the coupling relationship constraint of the three-level decision variables is as follows:

[0014] in, It is the planned value of the decision variable for the day. It is the intraday rolling plan value of the decision variable. It is the real-time planned value of the decision variable. It is the day-ahead forecast information of the decision variables. It is the intraday forecast bias of the decision variable. (·) is therefore and Let δ(·) be the intraday correction function of the independent variable. and is a real-time correction function for the independent variable, where the decision variable is the variable constituting the new power distribution system dispatch scheme.

[0015] Preferably, the multi-objective rolling optimization scheduling model includes: a day-ahead scheduling optimization model, an intraday rolling optimization model, and a real-time tracking control model; The objective function of the day-ahead scheduling optimization model is calculated as follows:

[0016] in,

[0017]

[0018]

[0019]

[0020] In the formula, Total operating cost as of today; It is the cost of purchasing electricity. It is the electricity purchase price for time period t. It represents the active power purchased from the upstream power grid during time period t. t is the duration of the time period; It is the cost of energy storage charging and discharging losses. It is the unit cost of energy storage charging and discharging losses. It is the charging power during time period t. It is the discharge power during time period t; It is the cost of penalties for abandoning wind and solar power. It is the penalty coefficient for wind / solar curtailment. It represents the amount of photovoltaic power curtailment during time period t. The amount of wind power curtailment during time period t; It is the cost of standby capacity. It is the unit cost of standby capacity. The increased reserve capacity provided during time period t The reduced reserve capacity provided during time period t; The objective function of the intraday rolling optimization model is calculated as follows:

[0021] in, Total rolling operating cost for the day; is the length of the intraday control cycle; k is the time period index; The actual tie-line power of k during the intraday period. Planned power of the connecting line during the current period (k); The electricity purchase cost for period k during the day; The energy storage charging and discharging loss cost during the k-period of the day; This refers to the cost of wind and solar power curtailment penalties during the intraday period (k). The state of charge of energy storage during the intraday period k. The planned energy storage state of charge for the period k before the date; This is the cost weighting coefficient. This is the tie-line power weighting factor. These are the weighting coefficients for the state of charge. ; The objective function of the real-time tracking control model is calculated as follows:

[0022] in, It is the real-time total operating cost; ) represents the change in the control variable over time period k. ) represents the tie-line power deviation during time period k. It is the energy storage charging power during time period k. θ1 is the energy storage discharge power during time period k; θ2 is the control change penalty coefficient; θ3 is the tie line deviation penalty coefficient; θ4 is the energy storage charging and discharging cost coefficient. It is the control time domain at the real-time scale; The cost weighting coefficient, tie-line power weighting coefficient, state of charge weighting coefficient, control change penalty coefficient, tie-line deviation penalty coefficient, and energy storage charging and discharging cost coefficient are dynamically adjusted using the entropy weighting method.

[0023] Preferably, based on the relevant parameters of the power distribution system, load forecasts at multiple time scales, and renewable energy output forecasts, under the constraints of a cross-time scale coupled constraint model, a distributed solution framework based on an improved alternating direction multiplier method is used to solve the multi-objective rolling optimization scheduling model, resulting in a novel power distribution system scheduling scheme, including: Based on relevant parameters of the power distribution system, the power distribution system is divided into multiple autonomous regions according to electrical distance and regional controllable resources. The global objective of the power distribution network is decomposed into each autonomous region to obtain the sub-objective of each autonomous region. Boundary variables and multiplier information are exchanged between autonomous regions. The boundary variables include tie line power and boundary node voltage amplitude. Under the constraints of the cross-timescale coupled constraint model and the boundary conditions between autonomous regions, the sub-objectives of each autonomous region are iteratively solved based on the improved alternating direction multiplier method according to the load forecast and new energy output forecast values ​​of multiple timescales until convergence, and a new power distribution system scheduling scheme is obtained. During the iteration process, the penalty parameter is adaptively adjusted based on the original residual and the dual residual, and a relaxation factor is introduced to adjust the decision variable after iteration and update the dual variable.

[0024] Preferably, the calculation formula for the sub-targets of the autonomous region is as follows;

[0025] in, i This is the autonomous region index, where N is the set of autonomous regions. It is an autonomous region i Local decision variables, It is an autonomous region i The local cost function; It is an autonomous region i The equality constraint coefficient matrix, It is an autonomous region i The right-hand side of the equality constraint, It is an autonomous region i The local feasible region; The iterative calculation formula is as follows:

[0026] in, This is the penalty parameter, and k is the number of iterations. It is the autonomous region in the k-th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i Local decision variables; It is the autonomous region adjusted by introducing a relaxation factor in the (k+1)th iteration. i Decision variables; The calculation formula is as follows:

[0027] in, It is a relaxation factor; Penalty parameters The adjustment is dynamically made based on the relative magnitudes of the original residual and the dual residual during the iteration process.

[0028] Preferably, the penalty parameter The adjustment method is as follows: If the global original residual is greater than the product of the global dual residual and the adjustment coefficient, the penalty parameter is increased according to the preset rules. If the global original residual is less than the product of the global dual residual and the adjustment coefficient, the penalty parameter is reduced according to the preset rules. Otherwise, keep the penalty parameter unchanged.

[0029] Preferably, the formula for calculating the global original residual is as follows:

[0030] The formula for calculating the global dual residual is as follows:

[0031] In the formula, This represents the global original residual at the (k+1)th iteration. This represents the global dual residual at the (k+1)th iteration. Describes the autonomous region at the (k+1)th iteration. i The original residuals, Describes the autonomous region at the (k+1)th iteration. i The dual residual; The calculation formula is as follows:

[0032] The calculation formula is as follows:

[0033] In the formula This represents the penalty parameter at the k-th iteration.

[0034] Based on the same inventive concept, the present invention also provides a novel rolling optimization scheduling system for power distribution systems that takes into account multi-timescale coupling, including: a data acquisition and prediction module, a solution module, and a control and scheduling module; The acquisition and prediction module is used to acquire relevant parameters of the power distribution system and perform load and renewable energy output prediction at multiple time scales to obtain load prediction values ​​and renewable energy output prediction values ​​at multiple time scales. The solution module is used to solve a multi-objective rolling optimization scheduling model based on the relevant parameters of the power distribution system, load forecasts at multiple time scales, and new energy output forecasts, under the constraints of a cross-time scale coupled constraint model, and using a distributed solution framework based on an improved alternating direction multiplier method to obtain a new power distribution system scheduling scheme. The control and scheduling module is used to execute the new power distribution system scheduling scheme in the rolling time domain and output scheduling control commands; Among them, the cross-timescale coupled constraint model couples a novel power distribution system scheduling scheme with different timescales; the multi-objective rolling optimization scheduling model achieves adaptive adjustment of weight coefficients through the entropy weight method; the improved alternating direction multiplier method adaptively adjusts the penalty parameter according to the original residual and the dual residual and introduces the super-relaxation technique in the iteration process.

[0035] Preferably, the multiple time scales include day-ahead scale, intraday scale, and real-time scale, and the acquisition and prediction module is specifically used for: Collect distribution network topology parameters, historical data of distributed power sources and loads, meteorological data, and energy storage status charge data as relevant parameters of the distribution system; Based on the Long Short-Term Memory Network, the system takes into account the weather forecast for the first hour of the future and historical data for the same period, and outputs the day-ahead point prediction curves of new energy sources and base load as the day-ahead scale load prediction values ​​and new energy output prediction values. Using probability box theory and combining the latest measured meteorological and load data, the uncertainty intervals of new energy output and load for the second time period in the future are generated as intraday load forecasts and new energy output forecasts. Every preset period, based on the most recent third-longest high-frequency data, time series decomposition combined with conformal quantile regression is used to output the power prediction values ​​of ultra-short-term load and renewable energy output for a set future time period, which are used as the real-time load prediction values ​​and renewable energy output prediction values. The lengths of the first duration, the second duration, and the third duration decrease sequentially.

[0036] Preferably, the cross-timescale coupling constraint model in the solution module includes: energy storage state of charge time coupling constraint, slow regulation device action number constraint, tie line power coupling constraint, and third-level decision variable coupling relationship constraint; The calculation formula for the time coupling constraint of the energy storage state of charge is as follows:

[0037] Among them, SOC It is the state of charge at time t. It is the state of charge at time t-1. It's about charging efficiency. It is the charging power during time period t. It is the discharge power during time period t. It is the discharge efficiency. It refers to the duration of the time period; The formula for calculating the constraint on the number of actions of the slow-speed adjustment device is as follows:

[0038] Among them, u The equipment status is adjusted slowly during time period t. It is a slow adjustment of the equipment status at time t-1. This is the maximum number of actions allowed. The formula for calculating the tie-line power coupling constraint is as follows:

[0039] in, It is the actual power of the tie line in time period t. The planned power output for the connecting line is t days prior. It is the proportional coefficient for allowable deviation; The calculation formula for the coupling relationship constraint of the three-level decision variables is as follows:

[0040] in, It is the planned value of the decision variable for the day. It is the intraday rolling plan value of the decision variable. It is the real-time planned value of the decision variable. It is the day-ahead forecast information of the decision variables. It is the intraday forecast bias of the decision variable. (·) is therefore and Let δ(·) be the intraday correction function of the independent variable. and is a real-time correction function for the independent variable, where the decision variable is the variable constituting the new power distribution system dispatch scheme.

[0041] Preferably, the multi-objective rolling optimization scheduling model in the solution module includes: a day-ahead scheduling optimization model, an intraday rolling optimization model, and a real-time tracking control model; The objective function of the day-ahead scheduling optimization model is calculated as follows:

[0042] in,

[0043]

[0044]

[0045]

[0046] In the formula, Total operating cost as of today; It is the cost of purchasing electricity. It is the electricity purchase price for time period t. It represents the active power purchased from the upstream power grid during time period t. t is the duration of the time period; It is the cost of energy storage charging and discharging losses. It is the unit cost of energy storage charging and discharging losses. It is the charging power during time period t. It is the discharge power during time period t; It is the cost of penalties for abandoning wind and solar power. It is the penalty coefficient for wind / solar curtailment. It represents the amount of photovoltaic power curtailment during time period t. The amount of wind power curtailment during time period t; It is the cost of standby capacity. It is the unit cost of standby capacity. The increased reserve capacity provided during time period t The reduced reserve capacity provided during time period t; The objective function of the intraday rolling optimization model is calculated as follows:

[0047] in, Total rolling operating cost for the day; is the length of the intraday control cycle; k is the time period index; The actual tie-line power of k during the intraday period. Planned power of the connecting line during the current period (k); The electricity purchase cost for period k during the day; The energy storage charging and discharging loss cost during the k-period of the day; This refers to the cost of wind and solar power curtailment penalties during the intraday period (k). The state of charge of energy storage during the intraday period k. The planned energy storage state of charge for the period k before the date; This is the cost weighting coefficient. This is the tie-line power weighting factor. These are the weighting coefficients for the state of charge. ; The objective function of the real-time tracking control model is calculated as follows:

[0048] in, It is the real-time total operating cost; ) represents the change in the control variable over time period k. ) represents the tie-line power deviation during time period k. It is the energy storage charging power during time period k. θ1 is the energy storage discharge power during time period k; θ2 is the control change penalty coefficient; θ3 is the tie line deviation penalty coefficient; θ4 is the energy storage charging and discharging cost coefficient. It is the control time domain at the real-time scale; The cost weighting coefficient, tie-line power weighting coefficient, state of charge weighting coefficient, control change penalty coefficient, tie-line deviation penalty coefficient, and energy storage charging and discharging cost coefficient are dynamically adjusted using the entropy weighting method.

[0049] Preferably, the solution module is specifically used for: Based on relevant parameters of the power distribution system, the power distribution system is divided into multiple autonomous regions according to electrical distance and regional controllable resources. The global objective of the power distribution network is decomposed into each autonomous region to obtain the sub-objective of each autonomous region. Boundary variables and multiplier information are exchanged between autonomous regions. The boundary variables include tie line power and boundary node voltage amplitude. Under the constraints of the cross-timescale coupled constraint model and the boundary conditions between autonomous regions, the sub-objectives of each autonomous region are iteratively solved based on the improved alternating direction multiplier method according to the load forecast and new energy output forecast values ​​of multiple timescales until convergence, and a new power distribution system scheduling scheme is obtained. During the iteration process, the penalty parameter is adaptively adjusted based on the original residual and the dual residual, and a relaxation factor is introduced to adjust the decision variable after iteration and update the dual variable.

[0050] Preferably, the calculation formula for the sub-objectives of the autonomous region in the solution module is as follows;

[0051] in, i This is the autonomous region index, where N is the set of autonomous regions. It is an autonomous region i Local decision variables, It is an autonomous region i The local cost function; It is an autonomous region i The equality constraint coefficient matrix, It is an autonomous region i The right-hand side of the equality constraint, It is an autonomous region i The local feasible region; The iterative calculation formula is as follows:

[0052] in, This is the penalty parameter, and k is the number of iterations. It is the autonomous region in the k-th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i Local decision variables; It is the autonomous region adjusted by introducing a relaxation factor in the (k+1)th iteration. i Decision variables; The calculation formula is as follows:

[0053] in, It is a relaxation factor; Penalty parameters The adjustment is dynamically based on the relative magnitudes of the original residual and the dual residual during the iteration process.

[0054] Preferably, the penalty parameter in the solution module The adjustment method is as follows: If the global original residual is greater than the product of the global dual residual and the adjustment coefficient, the penalty parameter is increased according to the preset rules. If the global original residual is less than the product of the global dual residual and the adjustment coefficient, the penalty parameter is reduced according to the preset rules. Otherwise, keep the penalty parameter unchanged.

[0055] Preferably, the formula for calculating the global original residual in the solution module is as follows:

[0056] The formula for calculating the global dual residual is as follows:

[0057] In the formula, This represents the global original residual at the (k+1)th iteration. This represents the global dual residual at the (k+1)th iteration. Describes the autonomous region at the (k+1)th iteration. i The original residuals, Describes the autonomous region at the (k+1)th iteration. i The dual residual; The calculation formula is as follows:

[0058] The calculation formula is as follows:

[0059] In the formula This represents the penalty parameter at the k-th iteration.

[0060] In another aspect, the present invention also provides an electronic device, comprising: at least one processor and a memory; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, a novel rolling optimization scheduling method for power distribution systems that takes into account multi-timescale coupling is implemented as described above.

[0061] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, it implements a novel rolling optimization scheduling method for power distribution systems that takes into account multi-time-scale coupling, as described above.

[0062] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides a novel rolling optimization scheduling method, system, device, and medium for power distribution systems that considers multi-timescale coupling. The method includes: collecting relevant parameters of the power distribution system and performing multi-timescale load and renewable energy output predictions to obtain multi-timescale load and renewable energy output predictions; based on the relevant parameters of the power distribution system, the multi-timescale load and renewable energy output predictions, and under the constraints of a cross-timescale coupling constraint model, solving a multi-objective rolling optimization scheduling model using a distributed solution framework based on an improved alternating direction multiplier method to obtain a novel power distribution system scheduling scheme; executing the novel power distribution system scheduling scheme in the rolling time domain and outputting scheduling control commands; wherein, the cross-timescale coupling constraint model is coupled... This invention presents novel power distribution system scheduling schemes at different time scales; a multi-objective rolling optimization scheduling model that adaptively adjusts weight coefficients using the entropy weight method; an improved alternating direction multiplier method that adaptively adjusts penalty parameters based on the original residual and dual residual, and introduces over-relaxation techniques during the iteration process; addressing the problem of weak coupling between scheduling models at different levels in existing power distribution network scheduling, this invention successfully constructs a multi-level coupled rolling optimization architecture, introduces cross-time scale coupling constraints, and effectively suppresses the problem of step-by-step command decoupling and deviation accumulation; addressing the problem of low solution efficiency in existing power distribution network scheduling, this invention adopts an improved distributed solution framework using the alternating direction multiplier method, which significantly shortens the solution time while ensuring global optimality and effectively protects the data privacy of all participating entities. Attached Figure Description

[0063] Figure 1 A flowchart of a novel rolling optimization scheduling method for power distribution systems that takes into account multi-time-scale coupling is provided by the present invention. Figure 2 A schematic diagram illustrating the overall process of a novel rolling optimization scheduling method for a power distribution system that takes into account multi-time-scale coupling, provided by the present invention. Figure 3 The three-level coupled rolling optimization scheduling architecture diagram provided by this invention; Figure 4 A schematic diagram of a novel rolling optimization scheduling system for power distribution systems that takes into account multi-time-scale coupling is provided by the present invention. Figure 5 This is a schematic diagram of an electronic device structure provided by the present invention. Detailed Implementation

[0064] To address the problems of weak coupling between scheduling models at different levels, insufficient resource coordination, lack of multi-objective optimization, and low solution efficiency in existing distribution network scheduling, this invention proposes a novel rolling optimization scheduling method, system, equipment, and medium for distribution systems that considers multi-timescale coupling. Based on the coupling constraints of three timescales—day-ahead, intraday, and real-time—and the differentiated adjustment capabilities of slow and fast resources, a three-level coupled rolling optimization architecture and resource coordination mechanism are constructed. Through this architecture, this invention can achieve collaborative optimization across timescales. The innovation of this invention lies in its ability to comprehensively consider multiple optimization objectives such as economy, security, and renewable energy absorption rate, and the introduction of a distributed solution framework based on the improved alternating direction multiplier method (ADMM). This ensures global optimality and computational efficiency while also protecting data privacy, adapting to the dynamically changing power grid operating environment. This method not only provides power grid companies with scientific multi-scale resource scheduling schemes but also significantly improves the operational economy, security, and renewable energy absorption level of the distribution network.

[0065] Example 1: A novel rolling optimization scheduling method for power distribution systems that considers multi-timescale coupling, such as... Figure 1 As shown, it includes: Step 1: Collect relevant parameters of the power distribution system and perform load and renewable energy output forecasts at multiple time scales to obtain load forecasts and renewable energy output forecasts at multiple time scales; Step 2: Based on the relevant parameters of the power distribution system, the load forecast values ​​at multiple time scales and the power output forecast values ​​of new energy sources, under the constraints of the cross-time scale coupled constraint model, the multi-objective rolling optimization scheduling model is solved using a distributed solution framework based on the improved alternating direction multiplier method, and a new power distribution system scheduling scheme is obtained. Step 3: Execute the new power distribution system dispatching scheme in the rolling time domain and output dispatching control commands; Among them, a novel power distribution system scheduling scheme is coupled with a cross-timescale coupling constraint model to couple different timescales; a multi-objective rolling optimization scheduling model achieves adaptive adjustment of weight coefficients through the entropy weight method; and an improved alternating direction multiplier method adaptively adjusts the penalty parameters based on the original residual and dual residual and introduces the super-relaxation technique during the iteration process.

[0066] Step 1 specifically includes: Step S1: Data Acquisition and Multi-Time Scale Prediction Collect distribution network topology parameters, historical data of distributed power sources / loads, meteorological data, and energy storage SOC data. Construct a three-level prediction model, as shown in Table 1 below.

[0067] Table 1: Three-level prediction model

[0068] On a day-ahead scale, the day-ahead forecast curves of new energy sources and base loads are predicted using LSTM (Long Short-Term Memory) network, weather forecasts, and historical data from the same period.

[0069] On an intraday scale, probability box theory is used, combined with the latest measured meteorological and load data, to generate uncertainty intervals for future renewable energy output and load.

[0070] At the real-time scale, based on the most recent high-frequency data, time series decomposition combined with conformal quantile regression is used to output the power prediction values ​​of ultra-short-term load and renewable energy output for a set future duration.

[0071] Step 2 specifically includes: Step S2: Construct a cross-timescale coupled constraint model (1) Time-coupling constraints of energy storage SOC (State of Charge)

[0072] Among them, SOC It is the state of charge at time t. It is the state of charge at time t-1. It's about charging efficiency. It is the charging power during time period t. It is the discharge power during time period t. It refers to discharge efficiency.

[0073] The energy storage plan determined by the dispatching authority is used as a boundary constraint for intraday optimization, and the intraday adjustment amount must be within the allowable range of energy storage capacity.

[0074] (2) Constraint on the number of actions of slow-speed adjustment equipment Among them, u The time period t is the slow adjustment of the equipment status (1 is put into operation / running, 0 is cut off / shutdown). It is a slow adjustment of the equipment status at time t-1. It represents the maximum number of actions allowed.

[0075] (3) Power coupling constraints of tie lines

[0076] in, It is the actual power of the tie line in time period t. The planned power output for the connecting line is t days prior. It is the proportional coefficient for allowable deviation; .

[0077] (4) Coupling relationship of three-level decision variables

[0078] in, It is the planned value of the decision variable for the day. It is the intraday rolling plan value of the decision variable. It is the real-time planned value of the decision variable. It is the day-ahead forecast information of the decision variables. It is the intraday forecast bias of the decision variable. (·) is therefore and Let δ(·) be the intraday correction function of the independent variable. and is a real-time correction function for the independent variable, where the decision variable is the variable constituting the new power distribution system dispatch scheme.

[0079] Step S3: Construct a multi-objective rolling optimization scheduling model S3.1 Day-ahead Scheduling Optimization Model

[0080] in,

[0081]

[0082]

[0083]

[0084] In the formula, Total operating cost as of today; It is the cost of purchasing electricity. It is the electricity purchase price for time period t. It represents the active power purchased from the upstream power grid during time period t. t is the duration of the time period; It is the cost of energy storage charging and discharging losses. It is the unit cost of energy storage charging and discharging losses. It is the charging power during time period t. It is the discharge power during time period t; It is the cost of penalties for abandoning wind and solar power. It is the penalty coefficient for wind / solar curtailment. The amount of curtailed photovoltaic power during time period t. The amount of wind power curtailment during time period t; It is the cost of standby capacity. It is the unit cost of standby capacity. The increased reserve capacity provided during time period t This refers to the reduced reserve capacity provided during time period t.

[0085] Constraints: Power balance, node voltage Line capacity, distributed power output, energy storage SOC ( And the beginning and end balance), OLTC (on-load tap changer) / capacitor operation count.

[0086] S3.2 Intraday Rolling Optimization Model (Model Predictive Control (MPC) Framework)

[0087] in, Total rolling operating cost for the day; For the length of the intraday control cycle, Hour; k is the time period index; The actual tie-line power of k during the intraday period. Planned power of the connecting line during the current period (k); The electricity purchase cost for period k during the day; The energy storage charging and discharging loss cost during the k-hour period of the day; This refers to the cost of wind and solar power curtailment penalties during the intraday period (k). The state of charge of energy storage during the intraday period k. The planned energy storage state of charge for the period k before the date; This is the cost weighting coefficient. This is the tie-line power weighting factor. These are the weighting coefficients for the state of charge. .

[0088] S3.3 Real-time tracking control model

[0089] in, It is the real-time total operating cost; ) represents the change in the control variable over time period k. ) represents the tie-line power deviation during time period k. It is the energy storage charging power during time period k. θ1 is the energy storage discharge power during time period k; θ2 is the control change penalty coefficient; θ3 is the tie line deviation penalty coefficient; θ4 is the energy storage charging and discharging cost coefficient. It is the control time domain at the real-time scale. minutes, solution cycle minute.

[0090] S3.4 Multi-objective coordination mechanism The entropy weight method is used to dynamically adjust the weights: the information entropy of each target is calculated in real time based on the current voltage limit exceedance, tie line power fluctuation, and renewable energy absorption rate, and the weights are automatically allocated.

[0091] Step S4: Distributed solution method based on improved ADMM Divide the distribution network into The global problem is decomposed into several autonomous regions:

[0092] i This is the autonomous region index, where N is the set of autonomous regions. It is an autonomous region i Local decision variables, It is an autonomous region i The local cost function; It is an autonomous region i The equality constraint coefficient matrix, It is an autonomous region i The right-hand side of the equality constraint, It is an autonomous region i The local feasible region.

[0093] Improved ADMM iteration format:

[0094] in, This is the penalty parameter, where k is the number of iterations; It is the autonomous region in the k-th iteration. i The dual variable, i.e., the one in the k-th iteration that is related to the equality constraint. The corresponding dual variable is used to modify the subproblem objective and accelerate convergence; It is the autonomous region in the (k+1)th iteration. i Local decision variables.

[0095] Improvement measures: Adaptive penalty parameter adjustment: Dynamically adjust based on the relative magnitudes of the original residual and the dual residual. ;

[0096]

[0097] Describes the autonomous region at the (k+1)th iteration. i The original residuals, Describes the autonomous region at the (k+1)th iteration. i The dual residuals.

[0098]

[0099]

[0100] It is the global original residual at the (k+1)th iteration. It is the global dual residual at the (k+1)th iteration.

[0101] if Then increase .

[0102] if Then decrease .

[0103] Otherwise, keep constant.

[0104] in, It is an adjustment factor.

[0105] Super-relaxation technique: Introducing relaxation factors Accelerate convergence.

[0106] In calculation Then, instead of directly passing it to the next step, we calculate the relaxed value:

[0107] Then use replace Used for subsequent dual variable updates, i.e.

[0108] Step 3 specifically includes: Step S5: Rolling time-domain optimized scheduling execution Rolling optimization is performed according to a three-level time sequence of day-day-intraday-real-time, updating forecast data and revising scheduling plans in real time.

[0109] Step S6: Output scheduling control instructions It outputs commands for the active / reactive power output of each distributed power source, the charging and discharging power of energy storage, the response of flexible loads, the OLTC / capacitor level, and the charging and discharging power of electric vehicles.

[0110] This invention addresses the problem of weak coupling between various levels of dispatching models in existing distribution network dispatching, and successfully constructs a day-to-day dispatching system. Within days The real-time three-level coupled rolling optimization architecture introduces cross-timescale coupling constraints to effectively suppress the problem of instruction decoupling and deviation accumulation. This invention addresses the problem of insufficient resource coordination in existing power distribution network scheduling by establishing a hierarchical coordination mechanism in which slow resources are responsible for baseline scheduling and fast resources are responsible for deviation compensation, thereby improving the efficiency of resource utilization across multiple time scales. This invention addresses the lack of multi-objective optimization in existing distribution network dispatching, simultaneously optimizing objectives such as operating costs, voltage quality, and renewable energy absorption rate, while taking into account the economy, safety, and system operation stability of the new distribution network. This invention addresses the problem of low solution efficiency in existing power distribution network scheduling by adopting an improved distributed solution framework using the alternating direction multiplier method. This significantly shortens the solution time while ensuring global optimality and effectively protects the data privacy of all participating entities.

[0111] In summary, this invention provides a scientific scheduling strategy that can accurately express coupling constraints across time scales while taking into account economy, safety, and renewable energy absorption rate. It is particularly suitable for distribution network operation and management in complex scenarios such as high-penetration distributed power sources, electric vehicles, energy storage systems, and large-scale flexible user-side resources. It helps to promote the evolution of distribution networks towards high elasticity and high intelligence, and supports the sustainable use of energy and the construction of new power systems.

[0112] Example 2: The following is a specific example of a novel rolling optimization scheduling method for power distribution systems that takes into account multi-timescale coupling.

[0113] Scenario: A new 33-node standard topology distribution network, comprising: 5MW distributed photovoltaic (multiple access points, some with reactive power regulation capabilities), 3MW distributed wind power (two 1.5MW units), a 2MW / 4MWh battery energy storage system (rated power 2MW, capacity 4MWh), and 500 electric vehicle charging piles (200 of which have V2G capabilities, with a total adjustable power of ±1.5MW). The distribution network is connected to the upstream grid via a 10kV tie line. Electricity purchase prices are high during peak hours and low during off-peak hours. The main challenges include fluctuations in renewable energy output, voltage fluctuations caused by the randomness of electric vehicle charging, and the issue of tie line power exceeding limits.

[0114] Implementation steps Step S1: Data Acquisition and Multi-Time Scale Prediction Collect distribution network topology parameters (33-node line impedance, transformer parameters), historical data of distributed power sources / loads (PV / wind power output and charging pile power over the past 3 years), meteorological data (irradiance, wind speed, temperature, humidity) and energy storage SOC data.

[0115] Day-ahead forecast (24 hours): Based on an LSTM neural network, it takes as input the weather forecast for the next 24 hours (1-hour resolution) and historical data for the same period, and outputs day-ahead forecast curves for photovoltaic, wind power, and base load, with a time resolution of 15 minutes. The day-ahead forecast error is <8% (measured by root mean square error).

[0116] Intraday forecast (4-6 hour rolling): Rolled every 4 hours, using the Probability Box theory, combined with the latest measured meteorological and load data, to generate the uncertainty range (90% confidence level) of new energy output for the next 6 hours, which is used for intraday robust optimization scheduling.

[0117] Real-time forecast (15 minutes - 1 hour): Executed every 15 minutes, based on high-frequency data from the most recent hour (sampling frequency 1 minute), using time series decomposition (STL, which decomposes the sequence into trend, period, and residual) combined with conformal quantile regression, to output ultra-short-term power forecasts for the next 15 minutes to 1 hour (providing the median and 90% prediction range).

[0118] Step S2: Construct a cross-timescale coupled constraint model Energy storage SOC time coupling constraints Energy storage model:

[0119] in , minute.

[0120] Operating scope: The SOC constraint is as follows: the day-ahead scheduling requires the SOC to be 0.5 at both 00:00 and 24:00, and the deviation between the actual SOC and the day-ahead planned value during intraday optimization shall not exceed ±0.05.

[0121] The intraday adjustment amount (change in charging and discharging power) must be within the allowable range of the energy storage capacity, that is, it must meet the following requirements. , And they cannot be charged and discharged at the same time.

[0122] Slow-speed adjustment equipment operation frequency constraint The distribution network is equipped with one on-load tap changing transformer (OLTC, tap range ±8×1.25%) and three sets of switchable capacitor banks (each set 0.5MVar).

[0123] The number of daily operations for OLTC is ≤8, and the number of daily operations for each capacitor bank is ≤5.

[0124] In intraday rolling optimization, a maximum of one OLTC action or one capacitor switching is allowed within any 15-minute period.

[0125] Tie line power coupling constraints The planned value for the switching power of the tie line is as follows: .

[0126] Actual switching power during the day Must meet:

[0127] in .

[0128] Third-level decision variable coupling relationship

[0129]

[0130] Ensure a smooth transition from day-to-day to real-time decision-making to avoid drastic fluctuations in equipment commands.

[0131] Step S3: Construct a multi-objective rolling optimization scheduling model S3.1 Day-ahead Scheduling Optimization Model Objective function:

[0132] Electricity purchase cost: Electricity prices are based on time-of-use pricing (peak price 1.2 yuan / kWh, flat price 0.7 yuan / kWh, valley price 0.3 yuan / kWh).

[0133] Energy storage loss cost:

[0134] Punishment for abandoning wind and light:

[0135] Backup capacity cost:

[0136] Constraints: Power balance, node voltage ( pu, PU), line capacity (not exceeding thermal stability limit), upper limit of distributed power output, energy storage SOC range and start-end balance, and OLTC / capacitor operation limit.

[0137] Solution method: Mixed Integer Linear Programming (MILP), using the commercial solver Gurobi, with a time resolution of 15 minutes.

[0138] S3.2 Intraday Rolling Optimization Model (MPC Framework) Prediction Time Domain The time domain is controlled for 4 hours, and the solution is obtained by rolling every 1 hour.

[0139] Objective function:

[0140] Weighting coefficients ( )satisfy The system employs an entropy weighting method for dynamic adjustment: it calculates the information entropy of each target in real time based on the current voltage limit exceedance, tie-line power fluctuations, and renewable energy absorption rate. For example, when the renewable energy absorption rate is below 90%, it automatically increases... When the tie line power deviation exceeds 5%, automatically increase... .

[0141] S3.3 Real-time tracking control model Control Time Domain The solution time is 10 minutes (adjustable from 15 to 60 minutes).

[0142] Objective function:

[0143] : Deviation of each node voltage from the reference value (1.0 pu).

[0144] : Deviation between the power of the connecting line and the planned value for the day.

[0145] The third item is the sum of the energy storage charging and discharging power (the smaller the value, the better for extending the lifespan).

[0146] Weight , , When a voltage exceeding the limit (<0.95 or >1.05) is detected, Automatically multiply by 10.

[0147] S3.4 Multi-objective coordination mechanism The entropy weight method is used to dynamically adjust the values ​​in S3.2. and in S3.3 .

[0148] Implementation details: The evaluation index vector is calculated every 15 minutes, including voltage qualification rate, tie-line volatility, renewable energy absorption rate, and energy storage lifespan loss rate. The information entropy of each index is calculated; a higher entropy value indicates greater uncertainty / importance, and thus a higher weight. For example, during periods of sharp drop in photovoltaic output, the entropy value of the renewable energy absorption rate increases. The weight is automatically increased.

[0149] Step S4: Distributed solution method based on improved ADMM Based on electrical distance and regionally controllable resources, the 33-node distribution network is divided into 3 autonomous regions: Area 1: Nodes 1-12, including 2MW of photovoltaic power, 1MW of wind power, and 200 charging piles.

[0150] Area 2: Nodes 13-23, including 1.5MW photovoltaic, 1MW wind power, 1MW / 2MWh energy storage (part of the energy storage), and 150 charging piles.

[0151] Area 3: Nodes 24-33, including 1.5MW photovoltaic, 1MW wind power, 1MW / 2MWh energy storage (another part), and 150 charging piles (100 of which are V2G charging piles).

[0152] Inter-region coupling variables are achieved through tie-line power and boundary node voltage magnitude.

[0153] Improved ADMM iteration format: The global problem is decomposed into:

[0154] Iterative updates:

[0155]

[0156] Adaptive penalty parameter adjustment: based on the original residual and dual residuals Relative size adjustment .like ,but ;like ,but Otherwise, keep Unchanged. Initially .

[0157] Hyperrelaxation technique: introducing relaxation factors In updating Execution after:

[0158] Accelerate convergence.

[0159] Convergence criterion: Both the original residual and the dual residual are less than 1. (The per-unit base value is 10 MVA). It usually converges within 20 iterations.

[0160] Distributed solution: each region solves its subproblems independently, exchanging only boundary variables and multiplier information, thus protecting data privacy within each region.

[0161] Step S5: Rolling time-domain optimized scheduling execution Execute as follows: Daily 00:00 executes day-ahead scheduling (future 24 hours, 15-minute resolution).

[0162] Intraday rolling forecasts: executed daily at 06:00, 10:00, 14:00, 18:00, and 22:00, each time rolling optimizations are made for the next 6 hours based on the latest forecast. ), control time domain 4 hours.

[0163] Real-time control: Executes every 10 minutes, with a control time range of 30 minutes.

[0164] Critical event simulation: Around 12:00 noon (12th hour), due to a sudden increase in cloud cover, the photovoltaic output plummeted from 4.2MW to 1.7MW within 5 seconds (a 60% decrease). The system response is as follows: t=12:00:00: The real-time control module detects a sharp drop in photovoltaic power and immediately triggers an emergency adjustment (no need to wait for the next 10-minute cycle, event-driven).

[0165] t=12:00:05: The energy storage system increases the discharge power from 0.2MW to 1.2MW within 5 seconds (using its rapid response capability), while the flexible load (15% of the 500 charging piles, approximately 75 piles) receives the instruction to reduce the charging power from 0.3MW to 0.1MW, making up for a total power deficit of approximately 1.2MW.

[0166] t=12:00:10: Intraday rolling optimization (the previous solution is still valid, but the event triggers a re-solution) completes local re-optimization within 10 seconds, adjusting the energy storage charging and discharging plan and tie-line power plan for subsequent periods.

[0167] t=12:00:30: System power balance is restored, and the deviation between the actual power of the tie line and the planned value of the day is controlled within 5%.

[0168] During the rolling execution process, scheduling instructions for each time period are sent to field equipment, such as energy storage PCS, photovoltaic inverters, charging pile controllers, and OLTC controllers, through the SCADA system.

[0169] Step S6: Output scheduling control instructions An instruction set is output every 10 minutes (real-time control cycle), containing the following: Distributed power generation active / reactive power output: Photovoltaic inverter 1 (node ​​5) active power output setpoint 0.85 pu (per unit value), reactive power output - 0.1 pu (reactive power absorption); wind power (node ​​18) active power output 0.92 pu.

[0170] Energy storage charging and discharging power: The power sequence of the energy storage system every 5 minutes over the next 30 minutes: +0.5MW (charging), +0.3MW, -0.2MW (discharging), -0.5MW, -0.6MW, -0.4MW.

[0171] Flexible load response: The total power limit of the charging pile group is adjusted from 1.2MW to 1.0MW (reduction of 0.2MW), lasting for 15 minutes.

[0172] OLTC setting: Current setting + 2 (boost 2 × 1.25%), next allowed action time 15 minutes later.

[0173] Capacitor bank switching status: Group 1 is in, Group 2 is out, and Group 3 is out.

[0174] Electric vehicle charging and discharging power (V2G): 50 V2G vehicles can provide a maximum discharge power of 0.3MW in the next 30 minutes, with a discharge electricity price compensation of 0.6 yuan / kWh.

[0175] All instructions are timestamped (current time + next 30 minutes, 5-minute resolution) and executed with priority (high: energy storage / flexible load; medium: photovoltaic / wind power; low: OLTC / capacitor).

[0176] After the command is executed, the system collects actual feedback data (SCADA measurements, PMU data) for the next cycle of prediction model correction and rolling optimization initialization.

[0177] Implementation effect The total scheduling cost was 12,800 yuan (a 17.2% reduction compared to no optimization plan).

[0178] The peak-to-valley difference of the tie line was reduced by 28%, effectively smoothing out power fluctuations.

[0179] The renewable energy consumption rate reached 96.5% (the wind and solar curtailment rate was only 3.5%).

[0180] In the event of a sudden 60% drop in photovoltaic power, the system re-optimizes and restores power balance within 5 seconds, keeping tie-line deviations within 5%.

[0181] Real-time control is executed every 10 minutes, and the voltage deviation is always less than ±3%.

[0182] The energy storage system is charged and discharged an average of 4.8 times per day (equivalent cycle count), which is within a reasonable range and does not excessively shorten its lifespan.

[0183] A specific example of the flow of a novel rolling optimization scheduling method for power distribution systems that considers multi-timescale coupling is as follows: Figure 2 As shown in the diagram, the three-level coupled rolling optimization scheduling architecture is as follows: Figure 3 As shown.

[0184] Figure 2 In the traditional rolling adjustment mechanism, based on the day-ahead plan, updated forecast data is used to continuously adjust the power generation plan for the next few hours. It prioritizes optimal economics and gradually corrects deviations. Event-driven rapid re-optimization, on the other hand, must be initiated immediately when disturbances occur that traditional adjustments cannot handle. Its primary goal is to ensure safe system operation; economic efficiency is secondary. Traditional rolling adjustment is a closed-loop rolling optimization, while event-driven rapid re-optimization is an immediate, trigger-based intervention. Once triggered, instructions are immediately generated and executed until the system returns to normal, at which point it switches back to traditional rolling adjustment mode.

[0185] Figure 3 In the loop containing day-ahead scheduling and intraday rolling correction, the corrected reference boundary refers to the process during the intraday rolling correction phase where the boundary conditions are reasonably adjusted based on the day-ahead scheduling plan to make the corrected plan more consistent with the current actual operating state, and then returned to the day-ahead scheduling. The corrected plan / boundary state in the loop containing intraday rolling correction and implementation tracking control is the complete solution output to real-time control after the intraday correction is completed. The planned trajectory obtained after intraday optimization, plus the boundary state, is input to real-time tracking control.

[0186] Example 3: Based on the same inventive concept, this invention also provides a novel rolling optimization scheduling system for power distribution systems that considers multi-time-scale coupling, such as... Figure 4 As shown, it includes: The module comprises a data acquisition and prediction module, a solution module, and a control and scheduling module. The acquisition and prediction module is used to acquire relevant parameters of the power distribution system and perform load and renewable energy output prediction at multiple time scales to obtain load prediction values ​​and renewable energy output prediction values ​​at multiple time scales. The solution module is used to solve a multi-objective rolling optimization scheduling model based on the relevant parameters of the power distribution system, load forecasts at multiple time scales, and new energy output forecasts, under the constraints of a cross-time scale coupled constraint model, and using a distributed solution framework based on an improved alternating direction multiplier method to obtain a new power distribution system scheduling scheme. The control and scheduling module is used to execute the new power distribution system scheduling scheme in the rolling time domain and output scheduling control commands; Among them, the cross-timescale coupled constraint model couples a novel power distribution system scheduling scheme with different timescales; the multi-objective rolling optimization scheduling model achieves adaptive adjustment of weight coefficients through the entropy weight method; the improved alternating direction multiplier method adaptively adjusts the penalty parameter according to the original residual and the dual residual and introduces the super-relaxation technique in the iteration process.

[0187] Preferably, the multiple time scales include day-ahead scale, intraday scale, and real-time scale, and the acquisition and prediction module is specifically used for: Collect distribution network topology parameters, historical data of distributed power sources and loads, meteorological data, and energy storage status charge data as relevant parameters of the distribution system; Based on the Long Short-Term Memory Network, the system takes into account the weather forecast for the first hour of the future and historical data for the same period, and outputs the day-ahead point prediction curves of new energy sources and base load as the day-ahead scale load prediction values ​​and new energy output prediction values. Using probability box theory and combining the latest measured meteorological and load data, the uncertainty intervals of new energy output and load for the second time period in the future are generated as intraday load forecasts and new energy output forecasts. Every preset period, based on the most recent third-longest high-frequency data, time series decomposition combined with conformal quantile regression is used to output the power prediction values ​​of ultra-short-term load and renewable energy output for a set future time period, which are used as the real-time load prediction values ​​and renewable energy output prediction values. The lengths of the first duration, the second duration, and the third duration decrease sequentially.

[0188] Preferably, the cross-timescale coupling constraint model in the solution module includes: energy storage state of charge time coupling constraint, slow regulation device action number constraint, tie line power coupling constraint, and third-level decision variable coupling relationship constraint; The calculation formula for the time coupling constraint of the energy storage state of charge is as follows:

[0189] Among them, SOC It is the state of charge at time t. It is the state of charge at time t-1. It's about charging efficiency. It is the charging power during time period t. It is the discharge power during time period t. It is the discharge efficiency. It refers to the duration of the time period; The formula for calculating the constraint on the number of actions of the slow-speed regulating device is as follows:

[0190] Among them, u The equipment status is adjusted slowly during time period t. It is a slow adjustment of the equipment status at time t-1. This is the maximum number of actions allowed. The formula for calculating the tie-line power coupling constraint is as follows:

[0191] in, It is the actual power of the tie line in time period t. The planned power output for the connecting line is t days prior. It is the proportional coefficient for allowable deviation; The calculation formula for the coupling relationship constraint of the three-level decision variables is as follows:

[0192] in, It is the planned value of the decision variable for the day. It is the intraday rolling plan value of the decision variable. It is the real-time planned value of the decision variable. It is the day-ahead forecast information of the decision variables. It is the intraday forecast bias of the decision variable. (·) is therefore and Let δ(·) be the intraday correction function of the independent variable. and is a real-time correction function for the independent variable, where the decision variable is the variable constituting the new power distribution system dispatch scheme.

[0193] Preferably, the multi-objective rolling optimization scheduling model in the solution module includes: a day-ahead scheduling optimization model, an intraday rolling optimization model, and a real-time tracking control model; The objective function of the day-ahead scheduling optimization model is calculated as follows:

[0194] in,

[0195]

[0196]

[0197]

[0198] In the formula, Total operating cost as of today; It is the cost of purchasing electricity. It is the electricity purchase price for time period t. It represents the active power purchased from the upstream power grid during time period t. t is the duration of the time period; It is the cost of energy storage charging and discharging losses. It is the unit cost of energy storage charging and discharging losses. It is the charging power during time period t. It is the discharge power during time period t; It is the cost of penalties for abandoning wind and solar power. It is the penalty coefficient for wind / solar curtailment. The amount of curtailed photovoltaic power during time period t. The amount of wind power curtailment during time period t; It is the cost of standby capacity. It is the unit cost of standby capacity. The increased reserve capacity provided during time period t The reduced reserve capacity provided during time period t; The objective function of the intraday rolling optimization model is calculated as follows:

[0199] in, Total rolling operating cost for the day; is the length of the intraday control cycle; k is the time period index; The actual tie-line power of k during the intraday period. Planned power of the connecting line during the current period (k); The electricity purchase cost for period k during the day; The energy storage charging and discharging loss cost during the k-hour period of the day; This refers to the cost of wind and solar power curtailment penalties during the intraday period (k). The state of charge of energy storage during the intraday period k. The planned energy storage state of charge for the period k before the date; This is the cost weighting coefficient. This is the tie-line power weighting factor. These are the weighting coefficients for the state of charge. ; The objective function of the real-time tracking control model is calculated as follows:

[0200] in, It is the real-time total operating cost; ) represents the change in the control variable over time period k. ) represents the tie-line power deviation during time period k. It is the energy storage charging power during time period k. θ1 is the energy storage discharge power during time period k; θ2 is the control change penalty coefficient; θ3 is the tie line deviation penalty coefficient; θ4 is the energy storage charging and discharging cost coefficient. It is the control time domain at the real-time scale; The cost weighting coefficient, tie-line power weighting coefficient, state of charge weighting coefficient, control change penalty coefficient, tie-line deviation penalty coefficient, and energy storage charging and discharging cost coefficient are dynamically adjusted using the entropy weighting method.

[0201] Preferably, the solution module is specifically used for: Based on relevant parameters of the power distribution system, the power distribution system is divided into multiple autonomous regions according to electrical distance and regional controllable resources. The global objective of the power distribution network is decomposed into each autonomous region to obtain the sub-objective of each autonomous region. Boundary variables and multiplier information are exchanged between autonomous regions. The boundary variables include tie line power and boundary node voltage amplitude. Under the constraints of the cross-timescale coupled constraint model and the boundary conditions between autonomous regions, the sub-objectives of each autonomous region are iteratively solved based on the improved alternating direction multiplier method according to the load forecast and new energy output forecast values ​​of multiple timescales until convergence, and a new power distribution system scheduling scheme is obtained. During the iteration process, the penalty parameter is adaptively adjusted based on the original residual and the dual residual, and a relaxation factor is introduced to adjust the decision variable after iteration and update the dual variable.

[0202] Preferably, the calculation formula for the sub-objectives of the autonomous region in the solution module is as follows;

[0203] in, i This is the autonomous region index, where N is the set of autonomous regions. It is an autonomous region i Local decision variables, It is an autonomous region i The local cost function; It is an autonomous region iThe equality constraint coefficient matrix, It is an autonomous region i The right-hand side of the equality constraint, It is an autonomous region i The local feasible region; The iterative calculation formula is as follows:

[0204] in, This is the penalty parameter, and k is the number of iterations. It is the autonomous region in the k-th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i Local decision variables; It is the autonomous region adjusted by introducing a relaxation factor in the (k+1)th iteration. i Decision variables; The calculation formula is as follows:

[0205] in, It is a relaxation factor; Penalty parameters The adjustment is dynamically based on the relative magnitudes of the original residual and the dual residual during the iteration process.

[0206] Preferably, the penalty parameter in the solution module The adjustment method is as follows: If the global original residual is greater than the product of the global dual residual and the adjustment coefficient, the penalty parameter is increased according to the preset rules. If the global original residual is less than the product of the global dual residual and the adjustment coefficient, the penalty parameter is reduced according to the preset rules. Otherwise, keep the penalty parameter unchanged.

[0207] Preferably, the formula for calculating the global original residual in the solution module is as follows:

[0208] The formula for calculating the global dual residual is as follows:

[0209] In the formula, This represents the global original residual at the (k+1)th iteration. This represents the global dual residual at the (k+1)th iteration. Describes the autonomous region at the (k+1)th iteration. i The original residuals, Describes the autonomous region at the (k+1)th iteration. i The dual residual; The calculation formula is as follows:

[0210] The calculation formula is as follows:

[0211] In the formula This represents the penalty parameter at the k-th iteration.

[0212] Example 4 like Figure 5 As shown, the present invention also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0213] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to realize the corresponding method flow or corresponding function, so as to realize the steps of a novel rolling optimization scheduling method for power distribution systems that takes into account multi-time scale coupling in the above embodiments.

[0214] Example 6 Based on the same inventive concept, this invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the storage medium here can include both built-in storage media within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. Loading and executing one or more instructions stored in the storage medium by the processor can implement the steps of a novel rolling optimization scheduling method for power distribution systems considering multi-timescale coupling in the above embodiments.

[0215] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0216] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0217] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The function specified in one or more boxes.

[0218] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0219] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A novel rolling optimization scheduling method for power distribution systems that considers multi-timescale coupling, characterized in that, include: Collect relevant parameters of the power distribution system and perform load and renewable energy output forecasts at multiple time scales to obtain load forecasts and renewable energy output forecasts at multiple time scales. Based on the relevant parameters of the power distribution system, the load forecast values ​​at multiple time scales, and the power output forecast values ​​of new energy sources, under the constraints of the cross-time scale coupled constraint model, a distributed solution framework based on the improved alternating direction multiplier method is used to solve the multi-objective rolling optimization scheduling model, resulting in a new power distribution system scheduling scheme. The novel power distribution system scheduling scheme is executed in the rolling time domain, and scheduling control commands are output. Among them, the cross-timescale coupled constraint model couples a novel power distribution system scheduling scheme with different timescales; the multi-objective rolling optimization scheduling model achieves adaptive adjustment of weight coefficients through the entropy weight method; the improved alternating direction multiplier method adaptively adjusts the penalty parameter according to the original residual and the dual residual and introduces the super-relaxation technique in the iteration process.

2. The novel rolling optimization scheduling method for power distribution systems as described in claim 1, characterized in that, The multiple time scales include day-ahead, intraday, and real-time scales. The process involves collecting relevant parameters of the power distribution system and performing load and renewable energy output forecasts at multiple time scales to obtain multi-time scale load forecasts and renewable energy output forecasts, including: Collect distribution network topology parameters, historical data of distributed power sources and loads, meteorological data, and energy storage status charge data as relevant parameters of the distribution system; Based on the Long Short-Term Memory Network, the system takes into account the weather forecast for the first hour of the future and historical data for the same period, and outputs the day-ahead point prediction curves of new energy sources and base load as the day-ahead scale load prediction values ​​and new energy output prediction values. Using probability box theory and combining the latest measured meteorological and load data, the uncertainty intervals of new energy output and load for the second time period in the future are generated as intraday load forecasts and new energy output forecasts. Every preset period, based on the most recent third-longest high-frequency data, time series decomposition combined with conformal quantile regression is used to output the power prediction values ​​of ultra-short-term load and renewable energy output for a set future time period, which are used as the real-time load prediction values ​​and renewable energy output prediction values. The lengths of the first duration, the second duration, and the third duration decrease sequentially.

3. The novel rolling optimization scheduling method for power distribution systems as described in claim 2, characterized in that, The cross-timescale coupling constraint model includes: energy storage state of charge time coupling constraint, slow regulation device action number constraint, tie line power coupling constraint, and third-level decision variable coupling relationship constraint. The calculation formula for the time coupling constraint of the energy storage state of charge is as follows: Among them, SOC It is the state of charge at time t. It is the state of charge at time t-1. It's about charging efficiency. It is the charging power during time period t. It is the discharge power during time period t. It is the discharge efficiency. It refers to the duration of the time period; The formula for calculating the constraint on the number of actions of the slow-speed regulating device is as follows: Among them, u The equipment status is adjusted slowly during time period t. It is a slow adjustment of the equipment status at time t-1. This is the maximum number of actions allowed. The formula for calculating the tie-line power coupling constraint is as follows: in, It is the actual power of the tie line in time period t. The planned power output for the connecting line is t days prior. It is the proportional coefficient for allowable deviation; The calculation formula for the coupling relationship constraint of the three-level decision variables is as follows: in, It is the planned value of the decision variable for the day. It is the intraday rolling plan value of the decision variable. It is the real-time planned value of the decision variable. It is the day-ahead forecast information of the decision variables. It is the intraday forecast bias of the decision variable. (·) is therefore and Let δ(·) be the intraday correction function of the independent variable. and is a real-time correction function for the independent variable, where the decision variable is the variable constituting the new power distribution system dispatch scheme.

4. The novel rolling optimization scheduling method for power distribution systems as described in claim 2, characterized in that, The multi-objective rolling optimization scheduling model includes: a day-ahead scheduling optimization model, an intraday rolling optimization model, and a real-time tracking control model; The objective function of the day-ahead scheduling optimization model is calculated as follows: in, In the formula, Total operating cost as of today; It is the cost of purchasing electricity. It is the electricity purchase price for time period t. It represents the active power purchased from the upstream power grid during time period t. t is the duration of the time period; It is the cost of energy storage charging and discharging losses. It is the unit cost of energy storage charging and discharging losses. It is the charging power during time period t. It is the discharge power during time period t; It is the cost of penalties for abandoning wind and solar power. It is the penalty coefficient for wind / solar curtailment. The amount of curtailed photovoltaic power during time period t. The amount of wind power curtailment during time period t; It is the cost of standby capacity. It is the unit cost of standby capacity. The increased reserve capacity provided during time period t The reduced reserve capacity provided during time period t; The objective function of the intraday rolling optimization model is calculated as follows: in, Total rolling operating cost for the day; is the length of the intraday control cycle; k is the time period index; The actual tie-line power of k during the intraday period. Planned power of the connecting line during the current period (k); The electricity purchase cost for period k during the day; The energy storage charging and discharging loss cost during the k-hour period of the day; This refers to the cost of wind and solar power curtailment penalties during the intraday period (k). The state of charge of energy storage during the intraday period k. The planned energy storage state of charge for the period k before the date; This is the cost weighting coefficient. This is the tie-line power weighting factor. These are the weighting coefficients for the state of charge. ; The objective function of the real-time tracking control model is calculated as follows: in, It is the real-time total operating cost; ) represents the change in the control variable over time period k. ) represents the tie-line power deviation during time period k. It is the energy storage charging power during time period k. θ1 is the energy storage discharge power during time period k; θ2 is the control change penalty coefficient; θ3 is the tie line deviation penalty coefficient; θ4 is the energy storage charging and discharging cost coefficient. It is the control time domain at the real-time scale; The cost weighting coefficient, tie-line power weighting coefficient, state of charge weighting coefficient, control change penalty coefficient, tie-line deviation penalty coefficient, and energy storage charging and discharging cost coefficient are dynamically adjusted using the entropy weighting method.

5. The novel rolling optimization scheduling method for power distribution systems as described in claim 1, characterized in that, Based on the relevant parameters of the power distribution system, load forecasts at multiple time scales, and renewable energy output forecasts, and under the constraints of a cross-time scale coupled constraint model, a distributed solution framework based on an improved alternating direction multiplier method is used to solve the multi-objective rolling optimization scheduling model, resulting in a novel power distribution system scheduling scheme, including: Based on relevant parameters of the power distribution system, the power distribution system is divided into multiple autonomous regions according to electrical distance and regional controllable resources. The global objective of the power distribution network is decomposed into each autonomous region to obtain the sub-objective of each autonomous region. Boundary variables and multiplier information are exchanged between autonomous regions. The boundary variables include tie line power and boundary node voltage amplitude. Under the constraints of the cross-timescale coupled constraint model and the boundary conditions between autonomous regions, the sub-objectives of each autonomous region are iteratively solved based on the improved alternating direction multiplier method according to the load forecast and new energy output forecast values ​​of multiple timescales until convergence, and a new power distribution system scheduling scheme is obtained. During the iteration process, the penalty parameter is adaptively adjusted based on the original residual and the dual residual, and a relaxation factor is introduced to adjust the decision variable after iteration and update the dual variable.

6. The novel rolling optimization scheduling method for power distribution systems as described in claim 5, characterized in that, The calculation formula for the sub-objectives of the autonomous region is as follows; in, i This is the autonomous region index, where N is the set of autonomous regions. It is an autonomous region i Local decision variables, It is an autonomous region i The local cost function; It is an autonomous region i The equality constraint coefficient matrix, It is an autonomous region i The right-hand side of the equality constraint, It is an autonomous region i The local feasible region; The iterative calculation formula is as follows: in, This is the penalty parameter, and k is the number of iterations. It is the autonomous region in the k-th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i Local decision variables; It is the autonomous region adjusted by introducing a relaxation factor in the (k+1)th iteration. i Decision variables; The calculation formula is as follows: in, It is a relaxation factor; Penalty parameters The adjustment is dynamically based on the relative magnitudes of the original residual and the dual residual during the iteration process.

7. The novel rolling optimization scheduling method for power distribution systems as described in claim 6, characterized in that, The penalty parameter The adjustment method is as follows: If the global original residual is greater than the product of the global dual residual and the adjustment coefficient, the penalty parameter is increased according to the preset rules. If the global original residual is less than the product of the global dual residual and the adjustment coefficient, the penalty parameter is reduced according to the preset rules. Otherwise, keep the penalty parameter unchanged.

8. The novel rolling optimization scheduling method for power distribution systems as described in claim 7, characterized in that, The formula for calculating the global original residual is as follows: The formula for calculating the global dual residual is as follows: In the formula, This represents the global original residual at the (k+1)th iteration. This represents the global dual residual at the (k+1)th iteration. Describes the autonomous region at the (k+1)th iteration. i The original residuals, Describes the autonomous region at the (k+1)th iteration. i The dual residual; The calculation formula is as follows: The calculation formula is as follows: In the formula This represents the penalty parameter at the k-th iteration.

9. A novel rolling optimization scheduling system for power distribution systems that considers multi-timescale coupling, characterized in that, include: The module comprises a data acquisition and prediction module, a solution module, and a control and scheduling module. The acquisition and prediction module is used to acquire relevant parameters of the power distribution system and perform load and renewable energy output prediction at multiple time scales to obtain load prediction values ​​and renewable energy output prediction values ​​at multiple time scales. The solution module is used to solve a multi-objective rolling optimization scheduling model based on the relevant parameters of the power distribution system, load forecasts at multiple time scales, and new energy output forecasts, under the constraints of a cross-time scale coupled constraint model, and using a distributed solution framework based on an improved alternating direction multiplier method to obtain a new power distribution system scheduling scheme. The control and scheduling module is used to execute the new power distribution system scheduling scheme in the rolling time domain and output scheduling control commands; Among them, the cross-timescale coupled constraint model couples a novel power distribution system scheduling scheme with different timescales; the multi-objective rolling optimization scheduling model achieves adaptive adjustment of weight coefficients through the entropy weight method; the improved alternating direction multiplier method adaptively adjusts the penalty parameter according to the original residual and the dual residual and introduces the super-relaxation technique in the iteration process.

10. The novel power distribution system rolling optimization scheduling system as described in claim 9, characterized in that, The multiple time scales include day-ahead scale, intraday scale, and real-time scale, and the acquisition and prediction module is specifically used for: Collect distribution network topology parameters, historical data of distributed power sources and loads, meteorological data, and energy storage status charge data as relevant parameters of the distribution system; Based on the Long Short-Term Memory Network, the system takes into account the weather forecast for the first hour of the future and historical data for the same period, and outputs the day-ahead point prediction curves of new energy sources and base load as the day-ahead scale load prediction values ​​and new energy output prediction values. Using probability box theory and combining the latest measured meteorological and load data, the uncertainty intervals of new energy output and load for the second time period in the future are generated as intraday load forecasts and new energy output forecasts. Every preset period, based on the most recent third-longest high-frequency data, time series decomposition combined with conformal quantile regression is used to output the power prediction values ​​of ultra-short-term load and renewable energy output for a set future time period, which are used as the real-time load prediction values ​​and renewable energy output prediction values. The lengths of the first duration, the second duration, and the third duration decrease sequentially.

11. The novel power distribution system rolling optimization scheduling system as described in claim 10, characterized in that, The cross-timescale coupling constraint model in the solution module includes: energy storage state of charge time coupling constraint, slow regulation device action number constraint, tie line power coupling constraint, and third-level decision variable coupling relationship constraint. The calculation formula for the time coupling constraint of the energy storage state of charge is as follows: Among them, SOC It is the state of charge at time t. It is the state of charge at time t-1. It's about charging efficiency. It is the charging power during time period t. It is the discharge power during time period t. It is the discharge efficiency. It refers to the duration of the time period; The formula for calculating the constraint on the number of actions of the slow-speed regulating device is as follows: Among them, u The equipment status is adjusted slowly during time period t. It is a slow adjustment of the equipment status at time t-1. This is the maximum number of actions allowed. The formula for calculating the tie-line power coupling constraint is as follows: in, It is the actual power of the tie line in time period t. The planned power output for the connecting line is t days prior. It is the proportional coefficient for allowable deviation; The calculation formula for the coupling relationship constraint of the three-level decision variables is as follows: in, It is the planned value of the decision variable for the day. It is the intraday rolling plan value of the decision variable. It is the real-time planned value of the decision variable. It is the day-ahead forecast information of the decision variables. It is the intraday forecast bias of the decision variable. (·) is therefore and Let δ(·) be the intraday correction function of the independent variable. and is a real-time correction function for the independent variable, where the decision variable is the variable constituting the new power distribution system dispatch scheme.

12. The novel power distribution system rolling optimization scheduling system as described in claim 10, characterized in that, The multi-objective rolling optimization scheduling model in the solution module includes: a day-ahead scheduling optimization model, an intraday rolling optimization model, and a real-time tracking control model; The objective function of the day-ahead scheduling optimization model is calculated as follows: in, In the formula, Total operating cost as of today; It is the cost of purchasing electricity. It is the electricity purchase price for time period t. It represents the active power purchased from the upstream power grid during time period t. t is the duration of the time period; It is the cost of energy storage charging and discharging losses. It is the unit cost of energy storage charging and discharging losses. It is the charging power during time period t. It is the discharge power during time period t; It is the cost of penalties for abandoning wind and solar power. It is the penalty coefficient for wind / solar curtailment. The amount of curtailed photovoltaic power during time period t. The amount of wind power curtailment during time period t; It is the cost of standby capacity. It is the unit cost of standby capacity. The increased reserve capacity provided during time period t The reduced reserve capacity provided during time period t; The objective function of the intraday rolling optimization model is calculated as follows: in, Total rolling operating cost for the day; is the length of the intraday control cycle; k is the time period index; The actual tie-line power of k during the intraday period. Planned power of the connecting line during the current period (k); The electricity purchase cost for period k during the day; The energy storage charging and discharging loss cost during the k-hour period of the day; This refers to the cost of wind and solar power curtailment penalties during the intraday period (k). The state of charge of energy storage during the intraday period k. The planned energy storage state of charge for the period k before the date; This is the cost weighting coefficient. This is the tie-line power weighting factor. These are the weighting coefficients for the state of charge. ; The objective function of the real-time tracking control model is calculated as follows: in, It is the real-time total operating cost; ) represents the change in the control variable over time period k. ) represents the tie-line power deviation during time period k. It is the energy storage charging power during time period k. θ1 is the energy storage discharge power during time period k; θ2 is the control change penalty coefficient; θ3 is the tie line deviation penalty coefficient; θ4 is the energy storage charging and discharging cost coefficient. It is the control time domain at the real-time scale; The cost weighting coefficient, tie-line power weighting coefficient, state of charge weighting coefficient, control change penalty coefficient, tie-line deviation penalty coefficient, and energy storage charging and discharging cost coefficient are dynamically adjusted using the entropy weighting method.

13. The novel power distribution system rolling optimization scheduling system as described in claim 9, characterized in that, The solution module is specifically used for: Based on relevant parameters of the power distribution system, the power distribution system is divided into multiple autonomous regions according to electrical distance and regional controllable resources. The global objective of the power distribution network is decomposed into each autonomous region to obtain the sub-objective of each autonomous region. Boundary variables and multiplier information are exchanged between autonomous regions. The boundary variables include tie line power and boundary node voltage amplitude. Under the constraints of the cross-timescale coupled constraint model and the boundary conditions between autonomous regions, the sub-objectives of each autonomous region are iteratively solved based on the improved alternating direction multiplier method according to the load forecast and new energy output forecast values ​​of multiple timescales until convergence, and a new power distribution system scheduling scheme is obtained. During the iteration process, the penalty parameter is adaptively adjusted based on the original residual and the dual residual, and a relaxation factor is introduced to adjust the decision variable after iteration and update the dual variable.

14. The novel power distribution system rolling optimization scheduling system as described in claim 13, characterized in that, The calculation formulas for the sub-objectives of the autonomous region in the solution module are as follows; in, i This is the autonomous region index, where N is the set of autonomous regions. It is an autonomous region i Local decision variables, It is an autonomous region i The local cost function; It is an autonomous region i The equality constraint coefficient matrix, It is an autonomous region i The right-hand side of the equality constraint, It is an autonomous region i The local feasible region; The iterative calculation formula is as follows: in, This is the penalty parameter, and k is the number of iterations. It is the autonomous region in the k-th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i dual variables, It is the autonomous region in the (k+1)th iteration. i Local decision variables; It is the autonomous region adjusted by introducing a relaxation factor in the (k+1)th iteration. i Decision variables; The calculation formula is as follows: in, It is a relaxation factor; Penalty parameters The adjustment is dynamically based on the relative magnitudes of the original residual and the dual residual during the iteration process.

15. The novel power distribution system rolling optimization scheduling system as described in claim 14, characterized in that, The penalty parameter in the solution module The adjustment method is as follows: If the global original residual is greater than the product of the global dual residual and the adjustment coefficient, the penalty parameter is increased according to the preset rules. If the global original residual is less than the product of the global dual residual and the adjustment coefficient, the penalty parameter is reduced according to the preset rules. Otherwise, keep the penalty parameter unchanged.

16. The novel power distribution system rolling optimization scheduling system as described in claim 15, characterized in that, The formula for calculating the global original residual in the solution module is as follows: The formula for calculating the global dual residual is as follows: In the formula, This represents the global original residual at the (k+1)th iteration. This represents the global dual residual at the (k+1)th iteration. Describes the autonomous region at the (k+1)th iteration. i The original residuals, Describes the autonomous region at the (k+1)th iteration. i The dual residual; The calculation formula is as follows: The calculation formula is as follows: In the formula This represents the penalty parameter at the k-th iteration.

17. An electronic device, characterized in that, include: At least one processor and memory; The memory and processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the novel rolling optimization scheduling method for power distribution systems that takes into account multi-time-scale coupling as described in any one of claims 1 to 8 is implemented.

18. A computer-readable storage medium, characterized in that, It contains an execution program, which, when executed, implements the novel rolling optimization scheduling method for power distribution systems that takes into account multi-time-scale coupling as described in any one of claims 1 to 8.

Images