CN-122026521-A - Multimode micro-grid-based operation scheduling method, system and device

CN122026521ACN 122026521 ACN122026521 ACN 122026521ACN-122026521-A

Abstract

The invention provides a multimode micro-grid-based operation scheduling method, a multimode micro-grid-based operation scheduling system and a multimode micro-grid-based operation scheduling device, which generate a photovoltaic output prediction curve and a load demand prediction curve in the future prediction domain by acquiring state data and external data of a micro-grid, and calculating a multidimensional state evaluation index, determining a current running mode, calling a corresponding target weight vector and a key constraint set, inputting a multi-target mixed integer programming model for solving, correcting the target weight vector based on a weight dynamic adaptation coefficient, and generating a mode switching execution instruction. The technical scheme of the invention realizes the multi-mode self-adaptive operation scheduling of the micro-grid, and improves the operation economy, low carbon and control flexibility.

Inventors

QIU MINGQUAN
HU FENG
LIN FANG
LIU DONG
Kang Tengda
KANG QI
LI RUNPEI
WANG CHAO
ZHANG YUN
Ding Danlei
SONG ZHIYU
WANG WEI
ZHANG LEI

Assignees

国网北京市电力公司
国家电网有限公司
清华四川能源互联网研究院

Dates

Publication Date: 20260512
Application Date: 20260320

Claims (10)

1. A multimode micro-grid-based operation scheduling method, the method comprising: Generating a photovoltaic output prediction curve and a load demand prediction curve in a future prediction time domain based on the acquired micro-grid state data; Calculating a multi-dimensional state evaluation index based on the state data, the photovoltaic output prediction curve and the load demand prediction curve, wherein the multi-dimensional state evaluation index at least comprises a safety margin index and a comfort degree deviation index of an energy storage charge state; Acquiring a target weight vector and a key constraint set corresponding to the operation mode according to the operation mode, and uniformly inputting the target weight vector and the key constraint set into a multi-target mixed integer programming model to solve to obtain an original control instruction; The method comprises the steps of calculating a weight dynamic adaptation coefficient based on the multi-dimensional state evaluation index in each control period, dynamically correcting an original weight in the target weight vector based on the weight dynamic adaptation coefficient to obtain a corrected target weight vector, and reallocating power proportions of energy storage, power grid interaction and flexible load based on the original control instruction and the corrected target weight vector to generate a mode switching execution instruction.
2. The multi-mode microgrid operation scheduling method according to claim 1, wherein said generating a photovoltaic output prediction curve and a load demand prediction curve in a future prediction time domain based on said state data comprises: The method comprises the steps of constructing an LSTM-FC hybrid prediction model, constructing a photovoltaic input characteristic vector comprising historical power and meteorological factors for photovoltaic output prediction, constructing a load input characteristic vector comprising historical load and date type characteristics for load demand prediction, inputting the input characteristic vector into the encoder to obtain a time context vector, and inputting the time context vector into the decoder to obtain a photovoltaic output prediction curve and a load demand prediction curve.
3. The multi-mode microgrid operation scheduling method according to claim 1, wherein said determining a current operation mode according to said multi-dimensional state evaluation index and a preset priority rule base comprises: and executing sequential scanning on triggering conditions of a safety guarantee mode, a low-carbon operation mode, a comfortable priority mode and an economic optimization mode according to a preset priority order, and locking the operation mode to be used as the current operation mode when the triggering condition of any operation mode is met.
4. The multi-mode microgrid-based operation scheduling method according to claim 1, further comprising updating at least one of the target weight vector and the set of critical constraints according to the current operation mode before the target weight vector and the set of critical constraints corresponding to the operation mode are obtained according to the operation mode, wherein: when the triggering condition of the safety guarantee mode is met, the target weight vector is switched to weight setting corresponding to a safety guarantee target, safety boundary constraint of an energy storage charge state is added to the key constraint set, and an energy storage discharge power instruction is limited; When the triggering condition of the low-carbon operation mode is met, the target weight vector is switched to weight setting corresponding to a low-carbon target, a forbidden reverse power transmission constraint is added to the key constraint set, and an energy storage charging power instruction is updated based on a photovoltaic absorption rate index; When the triggering condition of the comfort priority mode is met, the target weight vector is switched to weight setting corresponding to a comfort level target, and the weight parameter or punishment coefficient corresponding to the comfort level punishment cost is updated.
5. The multi-mode micro-grid operation scheduling method according to claim 1, wherein the objective function of the multi-objective mixed integer programming model at least comprises grid interaction cost, energy storage loss cost, carbon emission cost and comfort penalty cost, and the objective function of the multi-objective mixed integer programming model is as follows: Wherein, the In order to be able to carry out the overall operating costs, For the grid interaction cost at time t, For the energy storage loss cost at time t, For the carbon emission cost at time t, For the comfort penalty cost at time T, λeco is an economic target weight coefficient, λcar is a carbon emission target weight coefficient, λcom is a comfort target weight coefficient, and T is a predicted time domain length.
6. The multimode microgrid operation scheduling method according to claim 1, wherein the weight dynamic adaptation coefficients are obtained by the following formula: Wherein I ω,min and I ω,max are respectively the minimum value and the maximum value of the real-time state evaluation index in a preset statistical interval, Representing the dynamic adaptation coefficients of the weights, A real-time state assessment indicator representing any target dimension; The real-time state evaluation index at least comprises a daily operation cost deviation rate, a photovoltaic digestion rate and a comfort level deviation index.
7. The multi-mode microgrid operation scheduling method according to claim 6, wherein said performing dynamic correction on the original weights in the target weight vector based on the weight dynamic adaptation coefficients to obtain a corrected target weight vector comprises: aiming at an original power instruction of any equipment category in energy storage, power grid interaction and flexible load at a certain moment, an execution power instruction is obtained based on the target weight vector and the weight dynamic adaptation coefficient, and the execution power instruction is obtained by adopting the formula: Wherein, the Representing the original power command of the MILP solving output, Representing the dynamically weighted execution power instruction, Representing the corresponding target weight vector component of the run mode, And (3) representing a weight dynamic adaptation coefficient obtained by normalizing the real-time state evaluation index, wherein omega is a target dimension index.
8. The multi-mode microgrid operation scheduling method according to claim 1, wherein after generating the mode switching execution instruction, the method operates in a closed loop iteration manner, the closed loop iteration comprising: And updating a time index corresponding to the control period after the current control period is finished, returning to the step of executing the acquired state data and external data of the micro-grid, recalculating the weight dynamic adaptation coefficient based on the updated state data, executing dynamic correction on the original weight corresponding to the original control instruction, and updating the mode switching execution instruction.
9. A multi-mode micro-grid based operation scheduling system, the system comprising: the data acquisition module is used for generating a photovoltaic output prediction curve and a load demand prediction curve in the future prediction domain based on the acquired micro-grid state data; The state evaluation module is used for calculating a multi-dimensional state evaluation index based on the state data, the photovoltaic output prediction curve and the load demand prediction curve, wherein the multi-dimensional state evaluation index at least comprises an SOC safety margin index and a comfort level deviation index; The optimization solving module is used for acquiring a target weight vector and a key constraint set corresponding to the operation mode according to the operation mode, and uniformly inputting the target weight vector and the key constraint set into a multi-target mixed integer programming MILP model to solve so as to obtain an original control instruction; The system comprises an instruction generation module, a dynamic weight adjustment module, a mode switching execution instruction generation module and a mode switching execution module, wherein the instruction generation module is used for calculating a weight dynamic adjustment coefficient based on the multidimensional state evaluation index in each control period, dynamically correcting an original weight in the target weight vector based on the weight dynamic adjustment coefficient to obtain a corrected target weight vector, and reallocating power proportions of energy storage, power grid interaction and flexible load based on the original control instruction and the corrected target weight vector to generate the mode switching execution instruction.
10. A multimode microgrid operation scheduling device, characterized in that it comprises a computer device comprising a processor and a memory, the processor having stored therein computer instructions which, when executed, implement a multimode microgrid operation scheduling method according to any one of claims 1 to 9.

Description

Multimode micro-grid-based operation scheduling method, system and device Technical Field The invention belongs to the technical field of power system dispatching automation, and particularly relates to a multimode micro-grid-based operation dispatching method, system and device. Background In the micro-grid application scene taking the distributed photovoltaic device, the energy storage device and the adjustable load as cores, the system operation is obviously influenced by external factors such as meteorological fluctuation, load randomness, electricity price time variability, low-carbon assessment and the like, and meanwhile, the internal states such as energy storage charge state SOC (State of Charge), safety margin, load side comfort degree constraint and the like have strong time sequence coupling characteristics. The micro-grid dispatching needs to uniformly arrange source load output, energy storage charge and discharge, power grid interaction power and load regulation strategies in a prediction time domain so as to meet power balance and operation boundary constraint and keep stable operation under multi-objective requirements of economy, low carbon, safety, comfort and the like. Common practices in the prior art include an energy management strategy based on empirical rules or thresholds, a day-ahead/day-in rolling scheduling strategy based on a single-objective or multi-objective optimization model, and a scheduling scheme in which source load prediction results are input into an optimization solver to obtain control instructions. The partial scheme converts multiple targets into a unified target function by setting penalty items or weight coefficients, or separates strategy decision and equipment control by adopting hierarchical control, so that the basic constraint and scheduling requirements of micro-grid operation are met to a certain extent. However, the general method still has the defects of suitability in the operation of the multi-objective multi-scene dynamic switching micro-grid, high-frequency recalculation or dependence on a large number of parameter settings is often needed for scheduling solving in the aspect of efficiency, real-time performance and global performance are difficult to achieve, calculation burden increase and response lag easily occur in the mode switching or extreme scenes, modeling caliber and control caliber are always inconsistent in the aspect of mechanism consistency between a regular strategy and an optimized strategy, key mechanisms such as energy storage SOC safety boundary, load comfort constraint, grid interaction limitation and the like are difficult to penetrate through a prediction, decision and execution link in an consistent manner, so that the matching between a control command and a physical operation boundary is insufficient, and the micro-grid has the significant cross-period coupling relation in the aspect of cross-time associated characterization, such as SOC dynamic evolution, load adjustment time period continuity, carbon emission or cost accumulation constraint in the time dimension and the like, and the conventional scheme often has the defects of cross-time associated description, and is easy to generate abrupt change in the time domain, the front-back inconsistency or lack of the effect on prediction errors, so that the effective operation stability and the target can be achieved. Therefore, a micro-grid operation scheduling method based on multi-mode adaptive optimization is needed to improve scheduling response efficiency, operation mechanism consistency and cross-time collaborative scheduling capability under multi-objective constraint. Disclosure of Invention According to the scheme provided by the invention, the mode judgment and optimization solution integrated scheduling model taking the micro-grid state data, the external data and the source load prediction curve as inputs is constructed, the operation mode selection mechanism and the weight dynamic adaptation mechanism based on the multi-dimensional state evaluation index are introduced into the scheduling model, and the mode switching execution instruction facing the energy storage, the power grid interaction and the flexible load is output, so that the technical problems that the mode switching in the multi-objective operation scene depends on rigid rules, the optimization model is difficult to uniformly bear the multi-mode switching, and the execution instruction is difficult to adaptively correct along with the working condition in the conventional micro-grid scheduling are solved. In a first aspect of the present invention, there is provided a multimode micro-grid-based operation scheduling method, the method comprising: Acquiring state data and external data of a micro-grid, wherein the state data at least comprises photovoltaic output, load power and energy storage charge state, and the external data at least comprises electricity price information and meteorological infor