CN-121979125-A - EMS system data acquisition and dynamic regulation method and device

CN121979125ACN 121979125 ACN121979125 ACN 121979125ACN-121979125-A

Abstract

The application relates to a method and a device for data acquisition and dynamic regulation of an EMS (energy management system) system, belongs to the technical field of energy cooperative control, and aims to solve the problems that the existing data processing dimension is single, a multi-dimensional characteristic system is lacked, a dynamic regulation strategy is not available and the system compatibility is poor. The method comprises the steps of collecting multi-source data, converting the multi-source data into a unified format, carrying out space-time synchronization on the multi-source data after the unified format through time sequence processing, purifying the multi-source data and extracting characteristics, constructing a hybrid prediction model, combining a multi-target optimization algorithm to construct a dynamic regulation strategy, and developing a protocol analysis module when the system is newly added with energy equipment. The method fuses the multi-source data and outputs the dynamic adjustment strategy, thereby improving the compatibility of the system.

Inventors

WANG ZHEN
XIE XINGCHANG
LI HAO
Cao Yangcui
WANG JIE

Assignees

山东浪潮智慧能源科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251217

Claims (10)

1. The data acquisition and dynamic regulation method of the EMS system is characterized by comprising the following steps of: Step S1, multi-source data acquisition and space-time synchronization are carried out, multi-source data are acquired and converted into a unified format, and the space-time synchronization is carried out on the multi-source data after the unified format through time sequence processing; Step S2, multi-source data purification and feature extraction, wherein the multi-source data after time-space synchronization is subjected to data purification, and feature data is extracted through double-stage features; and step S3, a step of dynamic regulation and control, in which a hybrid prediction model is constructed based on the characteristic data extracted in the step S2, and a dynamic regulation and control strategy is constructed by combining a multi-objective optimization algorithm.
2. The method for collecting and dynamically controlling data of an EMS system according to claim 1, wherein the multi-source data includes output data of a photovoltaic device, charge and discharge amount data of an energy storage device, battery remaining power state data of the energy storage device, energy consumption data and operation state data of a charging pile, and energy consumption value data of a water and electricity combustion calorimeter; converting the collected multi-source data into a unified JSON-LD semantic format; The time-space synchronization of the multi-source data through the time sequence processing comprises the time-space synchronization of the multi-source data after the unified format through a millisecond time stamp algorithm.
3. The method for data acquisition and dynamic regulation of an EMS system according to claim 2, wherein the millisecond time stamp algorithm includes complement hierarchical interpolation, timing calibration, error correction; The method comprises the following steps of selecting an interpolation mode according to equipment types based on an energy consumption numerical data rule of a water-electricity combustion calorimeter, adopting linear interpolation by stationary equipment, and adopting adjacent interpolation and peak value correction by pulse equipment; Constructing 1 ms/time stamp grids by taking EMS server time as a reference, mapping original time stamps of high-frequency data to the time stamp grids, calculating deviation delta t, directly matching the nearest grid points when delta t is less than or equal to 5ms, correcting communication delay or marking equipment faults through a historical mean value when delta t is more than 5ms, and adopting sliding window time sequence calibration with a default window of 100 ms; the correction error comprises checking time sequence association by the time sequence association degree R, wherein the mathematical expression of the time sequence association R is as follows: wherein R is the time sequence association degree, n sampling total times, For the i-th sampling instant of time, Is the first The photovoltaic output of the moment of time, Is the sampled average value of the photovoltaic output, For the purpose of storing the charge power, Is the sampled average value of the stored charge power, Is a preset delay time.
4. The method for data acquisition and dynamic regulation of an EMS system according to claim 1, wherein the data cleansing includes noise filtering, missing value filling; The noise filtering adopts wavelet noise reduction to conduct layered decomposition on the synchronized multi-source data frequency, wherein second-level data is decomposed into 3-4 layers, and millisecond-level data is decomposed into 5-6 layers; For the missing part of the multi-source data after wavelet noise reduction, adopting the data average value filling at adjacent moments for the missing time length < the preset time part of the multi-source data, and for the missing time length of the multi-source data not less than the preset time part, completing filling by combining the energy consumption numerical data rule of the similar equipment in the multi-source data and the real-time operation data of the related energy equipment; The similar equipment comprises photovoltaic equipment, energy storage equipment, a charging pile and a water-electricity combustion heat meter, wherein the associated energy equipment is determined based on the cooperative operation relation of the photovoltaic equipment, the energy storage equipment, the charging pile and the water-electricity combustion heat meter.
5. The method for data collection and dynamic adjustment and control of an EMS system according to claim 3 or 4, wherein the dual-stage features include basic feature extraction and hidden feature extraction; The basic feature extraction comprises the steps of generating basic features based on energy domain rules; The basic characteristics comprise a photovoltaic output fluctuation rate, an energy storage equipment SOC dynamic coefficient and a charging pile peak-valley energy consumption label, wherein the SOC dynamic coefficient is a dynamic change coefficient of the state of the residual electric quantity of the energy storage battery; the hidden feature extraction comprises the steps of carrying out weight distribution on multi-source data based on an attention mechanism, and extracting hidden features; The method comprises the steps of carrying out weight distribution on multi-source data based on an attention mechanism, carrying out normalization preprocessing on the multi-source data based on time sequence calibration, dividing the time sequence data by adopting a sliding time window of 100ms, weighting and calculating the weight of each data and each feature by a time sequence association factor, a device cooperation contribution factor and a feature differentiation factor three-level factor, dynamically updating the weight every sliding time window, and reserving multi-source data association with the weight more than or equal to x; and x is a preset weight threshold, the time sequence correlation factor quantifies the time linkage compactness between the data based on the time sequence correlation degree R, the equipment cooperative contribution factor is set according to the contribution degree of the photovoltaic equipment, the energy storage equipment, the charging pile and the water and electricity combustion heat meter to the multi-energy cooperative regulation, and the characteristic discrimination factor is quantified according to the time sequence variance of the multi-source data.
6. The method for data acquisition and dynamic regulation and control of an EMS system according to claim 1, wherein constructing a hybrid prediction model includes taking purified multi-source time sequence data as an input of an LSTM network, constructing a 3-layer hidden layer to extract long-term and short-term dependence of the data in a specific time scale, and outputting a time sequence feature vector; and carrying out dimension unification treatment on the time sequence feature vector output by the LSTM and the double-stage feature, then fusing to generate a fused feature matrix, inputting the fused feature matrix into the nonlinear association among learning features of the gradient lifting tree model, and outputting a photovoltaic output short-term predicted value, a charging pile energy load predicted value and an energy storage charging and discharging demand predicted value.
7. The method for data acquisition and dynamic regulation and control of an EMS system according to claim 1, wherein the step of constructing a dynamic regulation and control strategy by combining a multi-objective optimization algorithm includes multi-objective quantization, constraint condition setting, and optimal strategy solving; the multi-objective quantization is to quantize the regulation objective into a green electricity consumption maximization objective function, an energy cost minimization objective function and a carbon emission intensity minimization objective function, wherein the mathematical expression of the green electricity consumption maximization objective function is as follows: maximizing the objective function for green electricity consumption; The energy cost minimization objective function mathematical expression is: minimizing an objective function for energy cost; the mathematical expression of the carbon emission intensity minimization objective function is: the objective function is minimized for carbon emission intensity.
8. The method for data acquisition and dynamic regulation and control of an EMS system according to claim 7, wherein the setting of the constraint conditions comprises setting a device constraint and a supply-demand balance constraint, wherein the device safety constraint is that the SOC dynamic coefficient of the energy storage device is in a preset range, and the supply-demand balance constraint is that photovoltaic output + grid power supply + energy storage discharge = charging pile energy + traditional load energy; The method comprises the steps of respectively solving an optimal solution of a green electricity consumption maximization objective function, an energy consumption cost minimization objective function and a carbon emission intensity minimization objective function in a constraint range through an optimization algorithm, outputting an energy storage charging and discharging strategy, a charging pile ordered charging scheme and a traditional energy distribution logic; The energy storage charging and discharging strategy comprises photovoltaic peak period charging and electricity price peak period discharging, and the ordered charging scheme of the charging pile comprises peak staggering charging.
9. The data acquisition and dynamic regulation device of the EMS system is characterized by comprising an acquisition multi-source data and space-time synchronization module, a multi-source data purification and feature extraction module and a dynamic regulation module; the multi-source data acquisition and space-time synchronization module acquires multi-source data, converts the multi-source data into a unified format and performs space-time synchronization on the multi-source data after the unified format through time sequence processing; The multi-source data purification and feature extraction module is used for carrying out data purification on the multi-source data after time-space synchronization and extracting feature data through double-stage features; the dynamic regulation and control module constructs a hybrid prediction model based on the characteristic data extracted by the multi-source data purification and characteristic extraction module, and constructs a dynamic regulation and control strategy by combining a multi-target optimization algorithm.
10. The device for data collection and dynamic adjustment and control of an EMS system according to claim 9, wherein the development protocol analysis module is adapted to collect multi-source data and space-time synchronization module, multi-source data purification and feature extraction module, and dynamic adjustment and control module when the energy device is newly added.

Description

EMS system data acquisition and dynamic regulation method and device Technical Field The invention belongs to the technical field of energy cooperative control, and particularly relates to a method and a device for data acquisition and dynamic regulation of an EMS (energy management system). Background The current EMS (energy management system) and related energy management method and equipment have the problems that a photovoltaic and energy storage equipment protocol lacks a unified analysis adaptation scheme, the cycle difference between power data second-level acquisition and gas etc. daily-level statistics is large, time sequence alignment is difficult to realize, each energy module independently operates, global linkage of green electricity consumption-energy storage peak clipping and valley filling-ordered charging of a charging pile is lacking, photovoltaic output fluctuation and electricity load mutation scene cannot be adapted, system suitability is weak, and newly-increased energy equipment needs to re-develop an interface module. The invention patent with the publication number of CN120855452A discloses a charge-discharge intelligent management system based on linkage of a BMS and an EMS, which is characterized by comprising a data acquisition module, a state vector generation module, an instruction generation module, a verification module and an optimization module, wherein the data acquisition module acquires energy cooperative control data through the BMS and the EMS and performs preprocessing, the state vector generation module constructs a combined impedance model, inputs the preprocessed energy cooperative control data into the combined impedance model to generate a combined state vector containing optimal matching frequency, total harmonic distortion rate, battery state-of-charge data and harmonic suppression weight, the instruction generation module constructs a charge-discharge instruction generation model, inputs the combined state vector into the charge-discharge instruction generation model, generates an optimal charge-discharge strategy instruction through a strategy generation layer and a decision optimization layer, analyzes and executes the optimal charge-discharge strategy instruction, monitors energy cooperative control data change in real time and calculates an optimal charge-discharge strategy instruction execution effect score to generate a verification report, and the optimization module updates parameters of the charge-discharge instruction generation model according to the verification report. The prior art has the defects of single data processing dimension, lack of a multidimensional characteristic system, no dynamic adjustment strategy and poor system compatibility. This is a disadvantage of the prior art. In view of the above, the present invention provides a method and apparatus for data acquisition and dynamic adjustment of an EMS system to solve the above-mentioned drawbacks of the prior art. Disclosure of Invention Aiming at the technical problems of single data processing dimension, lack of a multidimensional characteristic system, no dynamic regulation strategy and poor system compatibility in the prior art, the invention provides a method and a device for data acquisition and dynamic regulation of an EMS system, so as to solve the technical problems. In a first aspect, the present invention provides a method for data acquisition and dynamic regulation of an EMS system, including: Step S1, multi-source data acquisition and space-time synchronization are carried out, multi-source data are acquired and converted into a unified format, and the space-time synchronization is carried out on the multi-source data after the unified format through time sequence processing; The multi-source data comprise output data of the photovoltaic equipment, charge and discharge amount data of the energy storage equipment, battery residual electric quantity state data of the energy storage equipment, energy consumption data and running state data of the charging pile and energy consumption numerical data of the water-electricity combustion heat meter; The output data of the photovoltaic equipment comprises real-time power generation power, accumulated power generation capacity, power generation power fluctuation rate and irradiance related output data of the photovoltaic equipment. Converting the collected multi-source data into a unified JSON-LD semantic format; performing space-time synchronization on the multi-source data through time sequence processing comprises performing space-time synchronization on the multi-source data with a unified format by adopting a millisecond time stamp algorithm; the millisecond time stamp algorithm comprises complementing hierarchical interpolation, time sequence calibration and error correction; The method comprises the following steps of selecting an interpolation mode according to equipment types based on an energy consumption numerical data rul