CN119543421B

CN119543421B - An energy management system and method based on cloud-edge integrated photovoltaic storage system

Info

Publication number: CN119543421B
Application number: CN202411590741.XA
Authority: CN
Inventors: 郭树星; 范琳祥; 沈尉伟; 龚林; 刘凤梁; 李敏
Original assignee: Shanghai Chengchuan Energy Storage Technology Co ltd
Current assignee: Shanghai Chengchuan Energy Storage Technology Co ltd
Priority date: 2024-11-08
Filing date: 2024-11-08
Publication date: 2025-08-22
Anticipated expiration: 2044-11-08
Also published as: CN119543421A

Abstract

The present application provides an energy management system based on a cloud-edge integrated photovoltaic storage system, including a cloud module and an edge module; an EMS controller for collecting information from the photovoltaic storage device module and the load module, and sending the collected data to the cloud module; the cloud module receives the data collected by the EMS controller, and predicts the future photovoltaic power generation power/electricity through the photovoltaic power generation prediction model; predicts the future load through the load prediction model; analyzes the peak and valley periods of electricity prices through the electricity price management model, and sends the predicted data to the edge module; the edge module performs energy management according to the battery SOC, rechargeable capacity, photovoltaic predicted power and power generation, superimposed peak and valley arbitrage and safety anti-backflow logic through the control logic of energy management. Through the above-mentioned settings, the present application makes the entire energy management system highly intelligent and automated, realizes the optimization of energy distribution and use, and improves energy utilization efficiency.

Description

Energy management system and method based on cloud-edge integrated optical storage system

Technical Field

The application relates to the technical field of photovoltaic energy storage, in particular to an energy management system and method based on a cloud-edge integrated light storage system.

Background

With the deep implementation of global energy structure transformation and sustainable development strategy, the development and utilization of renewable energy have become important ways to relieve energy pressure and reduce carbon emissions. Photovoltaic power generation and wind power generation are main forms of renewable energy sources, and have been rapidly developed in recent years due to advantages such as cleanliness, reproducibility, wide distribution, and the like. However, these renewable energy power generation have intermittent and uncertain characteristics, and the power generation amount and the power generation time are difficult to accurately predict, which brings challenges to the stable operation of the power system.

In the prior art, china patent with the application number of CN109473999A discloses a light storage power distribution method and a light storage power distribution device, the method is applied to a photovoltaic energy storage system, a peak clipping and valley filling mode in which the photovoltaic energy storage system is located is firstly obtained, and the current energy storage state of an energy storage unit is determined based on the peak clipping and valley filling mode. And then determining the discharging steady-state power and the charging steady-state power of the energy storage unit based on the peak clipping and valley filling mode, the photovoltaic power generation power, the load power, the preset minimum value of the exchange power with the mains supply and the preset maximum steady-state value of the exchange power with the mains supply. The first distributed power of the energy storage unit is determined based on a peak-to-valley fill mode, a current energy storage state, a rated power, a discharge steady-state power or a charge steady-state power of the energy storage unit, and the patent is developed around peak-to-valley arbitrage, but the peak time period is charged based on a real-time state, the flat time period is discharged and the like, so that the division is too mechanical, and the comprehensive benefit is poor.

The patent application No. CN117762744A discloses a large data-based optical energy storage management system, which particularly relates to the technical field of photovoltaic energy storage, and comprises an optical storage device data collection module, a data arrangement module, a neural network module, a database, an evaluation module and an optical storage device management module, wherein the optical storage device management module is respectively connected with the optical storage device data collection module, the data arrangement module, the neural network module, the database and the evaluation module, and the optical storage device data collection module is used for collecting body operation information, charging circuit information, discharging circuit information and external environment information of the optical storage device and transmitting the body operation information, the charging circuit information, the discharging circuit information and the external environment information of the optical storage device to the data arrangement module.

Disclosure of Invention

Aiming at the defects in the prior art, the application aims to provide an energy management system and method based on a cloud-edge integrated light storage system, and aiming at a photovoltaic energy storage system, a complete cloud-edge integrated design scheme is adopted, so that scheme design is carried out from the angle of benefit maximization, technical innovation and project practicality are considered, the effect that photovoltaic can be fully utilized and power grid electricity is taken as little as possible is realized.

The application provides an energy management system based on a cloud-edge integrated optical storage system, which comprises a cloud module and an edge module;

The cloud module comprises a photovoltaic power generation prediction model, an electricity price management model and a load prediction model;

The edge end module comprises an EMS controller with energy management control logic and an information acquisition function;

the EMS controller is used for collecting information of the light storage equipment module and the load module and uniformly transmitting the collected data to the cloud module;

The cloud module predicts future photovoltaic power generation power and power generation capacity through the photovoltaic power generation prediction model by receiving data acquired by the EMS controller, predicts future load through the load prediction model, analyzes electricity price peak-valley period through the electricity price management model and sends predicted data to the edge module;

and the edge module is used for carrying out energy management according to the battery SOC, chargeable quantity, future photovoltaic power generation power and power generation quantity, superposition peak Gu Tao and safety anti-backflow logic through the EMS controller.

Further, the photovoltaic power generation prediction model includes:

the first data collection unit is used for collecting the meteorological data of the place where the equipment is located, the historical power generation data of the equipment, the state data of the equipment and the geographic information of the equipment;

the first data preprocessing unit is used for cleaning the collected data and preprocessing the characteristic engineering data;

and the first prediction unit adopts a linear regression equation, inputs the preprocessed data into the photovoltaic power generation prediction model, and predicts future photovoltaic power generation power and power generation capacity.

Further, the linear regression equation is:

p＝β+β₁T+β₂H+β₃S+β₄C+β₅S_pv+β₆S_inv+β₇L_lat+β₈L_lon+β₉A+∈;

,β、β₁、β₂、β₃、β₄、β₅、β₆、β₇、β₈、β₈ is a coefficient to be fitted, which is estimated through training data, epsilon is correction deviation, p is photovoltaic power generation prediction data, S _pv is a photovoltaic panel state, S _inv is an inverter state, L _lat is latitude of a geographic position, L _lon is longitude of the geographic position, C is cloud cover, S is irradiation intensity, T is temperature, H is humidity, and A is altitude.

Further, the load prediction model includes:

The second data collection unit is used for collecting historical electricity utilization data and electricity utilization plans;

the second data preprocessing unit is used for carrying out normalization processing on the collected data;

And the second prediction unit is used for constructing a state equation and an observation equation to predict the future load according to the data processed by the second data preprocessing unit and the state space model.

Further, the state equation is:

the observation equation is:

Where X _t is the state vector at time t, Y _t is the predicted vector at time t, u _t is the control input at time t, A, B, C, D is the system matrix, and w _t and v _t are the process noise and the observation noise.

Further, the future load prediction specifically comprises the steps of estimating parameters of a state space model by using a Bayesian method and carrying out state estimation and prediction by using a Kalman filter.

Further, the electricity price management model includes:

A third data collection unit for acquiring electricity price data from an electric power company or an electric power market operator website;

The third data preprocessing unit fills or deletes the missing value, ensures that the electricity price data has a consistent time stamp format and aligns the data;

the data visualization unit is used for displaying the change of electricity price along with time;

and the peak-valley period identification unit automatically identifies the peak period, the normal period and the valley period in the electrical value by using a K-means clustering algorithm.

Further, the EMS controller includes:

The battery management module is used for monitoring the SOC of the battery and the residual capacity of the battery which can be charged currently;

the photovoltaic power generation management module is used for monitoring the photovoltaic power generation amount in real time and reporting the photovoltaic power generation amount to the cloud module to help the cloud module to correct the predicted value of the photovoltaic power generation power and the power generation amount in real time;

the charging management module calculates the SOC value of the battery pre-charge at night according to the peak-valley time period of the power consumption, the photovoltaic power generation amount and the power price of the load in the future day, and controls the battery to charge;

the discharging management module is used for preferentially discharging the battery or reducing power grid electricity purchasing according to the electricity price time period and the battery SOC;

and the anti-reflux strategy module monitors the anti-reflux meter in real time, adjusts the photovoltaic power generation or battery charging and discharging strategy, and ensures that no redundant power flows back to the power grid.

In a second aspect of the present application, there is provided an energy management method based on a cloud-edge integrated optical storage system, the method comprising:

collecting information of the optical storage equipment module and the load module through the EMS controller, and sending data to the cloud module;

the cloud module predicts future photovoltaic power generation power and power generation capacity through a photovoltaic power generation prediction model, predicts future load through a load prediction model, and analyzes electricity price peak-valley time periods through an electricity price management model;

Control logic for performing energy management in the EMS controller according to battery SOC, chargeable amount, future photovoltaic power generation power and power generation, superposition peak Gu Tao li, and safe anti-reflux logic;

the EMS controller monitors photovoltaic power generation capacity, battery SOC and anti-backflow conditions in real time, and adjusts photovoltaic power generation and battery charging and discharging strategies according to instructions of the cloud module;

the cloud module calculates benefits based on the real-time data and optimizes an energy management strategy according to the benefits.

Further, the photovoltaic power generation prediction model adopts a linear regression mode, the load prediction model uses a state space model, and the electricity price management model uses a K-means clustering algorithm;

the control logic of the energy management comprises battery management, photovoltaic power generation management, charge management, discharge management and anti-reflux strategy;

And the EMS controller dynamically adjusts photovoltaic power generation and battery charging and discharging strategies according to the cloud module prediction data and the real-time data so as to realize energy optimization management.

Compared with the prior art, the application has at least one of the following beneficial effects:

1. According to the photovoltaic power generation prediction model, the electricity price management model and the load prediction model of the cloud module, future photovoltaic power generation capacity, electricity price change and load demand can be accurately predicted, the electricity price management model can analyze electricity price peak-valley time periods, so that an energy management system can charge when electricity price is low, discharge is carried out when electricity price is high, optimization of power cost is achieved, safety anti-backflow logic is considered by an edge end module in energy management, safe operation of the system under the condition of power grid faults or abnormal conditions is guaranteed, impact caused by backflow to a power grid is prevented, and the cloud module and the edge module work cooperatively, so that the whole energy management system has high intelligent and automatic levels, can make fast and accurate decisions according to real-time data, optimize energy distribution and use, and improve energy utilization efficiency.

2. The application integrates data collection, preprocessing, visualization and peak-valley period automatic identification functions, effectively integrates and optimizes the electricity price data obtained from an electric company or a market operator, ensures the consistency and the integrity of the data, reduces the possibility of manual intervention and errors, simultaneously presents the change trend of the electricity price in an intuitive way, accurately identifies the peak period, the normal period and the valley period in the electricity price through a K-means clustering algorithm, optimizes a reasonable power scheduling scheme through the change rule of the electricity price, and realizes optimal power resource allocation and planning of economic benefit

Drawings

Other features, objects and advantages of the present application will become more apparent upon reading of the detailed description of non-limiting embodiments, given with reference to the accompanying drawings in which:

FIG. 1 is a block diagram of an energy management system based on a cloud-edge integrated optical storage system according to an embodiment of the present application.

Fig. 2 is a schematic structural diagram of an energy management system based on a cloud-edge integrated optical storage system according to an embodiment of the present application.

FIG. 3 is a schematic diagram of an embodiment of an optical storage system.

Fig. 4 is a flowchart of an energy management method based on a cloud-edge integrated optical storage system according to an embodiment of the present application.

Detailed Description

The present application will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the present application, but are not intended to limit the application in any way. It should be noted that variations and modifications could be made by those skilled in the art without departing from the inventive concept. These are all within the scope of the present application.

Referring to fig. 1, an energy management system based on a cloud-edge integrated optical storage system according to an embodiment of the application includes a cloud module and an edge module.

The cloud module comprises a photovoltaic power generation prediction model, an electricity price management model and a load prediction model, and the edge end module comprises an EMS controller with energy management control logic and an information acquisition function.

The EMS controller is used for collecting information of the light storage equipment module and the load module and sending the collected data to the cloud module, the cloud module predicts future photovoltaic power generation power and power generation capacity through the photovoltaic power generation prediction model by receiving the data collected by the EMS controller, predicts future load through the load prediction model, analyzes electricity price peak-valley time periods through the electricity price management model and sends the predicted data to the edge module, and the edge module carries out energy management according to the battery SOC, chargeable capacity, future photovoltaic power generation power and power generation capacity, superposition peak Gu Tao and safe anti-backflow logic through the control logic of energy management.

According to the photovoltaic power generation prediction model, the electricity price management model and the load prediction model of the cloud module, future photovoltaic power generation capacity, electricity price change and load demand can be accurately predicted, the electricity price management model can analyze electricity price peak-valley time periods, so that an energy management system can charge when electricity price is low, discharge is carried out when electricity price is high, optimization of power cost is achieved, safety anti-backflow logic is considered by an edge end module in energy management, safe operation of the system under the condition of power grid faults or abnormal conditions is guaranteed, impact caused by backflow to a power grid is prevented, and the cloud module and the edge module work cooperatively, so that the whole energy management system has high intelligent and automatic levels, can make fast and accurate decisions according to real-time data, optimize energy distribution and use, and improve energy utilization efficiency.

The cloud-edge integrated method comprises the steps of deploying a photovoltaic power generation prediction model, a wind power generation prediction model, an electricity price management model and a load prediction model to a cloud end module, deploying control logic and an information acquisition function for specific energy management to an edge end module, using an EMS (energy management system) controller as a data collection module and a management module, collecting information of a photovoltaic storage device module and a load module by the EMS, uniformly transmitting the collected data to the cloud end, combining information of battery SOC, chargeable capacity, future photovoltaic power generation power, generated energy and the like and safe anti-backflow logic according to the received prediction data by the edge module, performing energy management by means of control logic of energy management, including establishment and execution of strategies such as charging, discharging and energy storage, and supplementing electric quantity difference by using trough electricity price and predicting load, and enabling photovoltaic power to be fully utilized and electric network electricity taking to be as small as possible.

In some possible embodiments, the photovoltaic power generation prediction model comprises a first data collection unit, a first data preprocessing unit and a first prediction unit, wherein the first data collection unit is used for collecting equipment location meteorological data, equipment historical power generation data, equipment state data and equipment geographic information, the first data preprocessing unit is used for cleaning collected data and preprocessing characteristic engineering data, and the first prediction unit adopts a linear regression equation to input the preprocessed data into the photovoltaic power generation prediction model to predict future photovoltaic power generation power and power generation capacity.

The photovoltaic power generation prediction module firstly collects meteorological data such as illumination intensity, temperature, humidity and wind speed of a place where equipment is located through a first data collection unit, collects historical power generation data of the equipment, including power generation power, power generation capacity and the like in a period of time, collects state data of the equipment, such as working state of the equipment, fault records and the like, collects geographic information of the equipment, such as longitude, latitude and altitude, and the like, takes the data as input of the power generation prediction module, then cleans the collected data through a first data preprocessing unit, preprocesses characteristic engineering data to obtain characteristics useful for photovoltaic power generation prediction, improves prediction accuracy and stability of the model, inputs the data into a first prediction unit of a photovoltaic power generation prediction model, and finally predicts future photovoltaic power generation power and future photovoltaic power generation power by a linear regression equation through the first prediction unit.

The method comprises the steps of cleaning data, namely removing abnormal values, missing values or repeated values, guaranteeing data quality, preprocessing characteristic engineering data, and extracting characteristics useful for photovoltaic power generation prediction, such as time sequence characteristics of illumination intensity, relationship characteristics of temperature and power generation power and the like.

The method has the advantages that the characteristics useful for photovoltaic power generation prediction are extracted by collecting abundant data and preprocessing, the linear regression equation is used as a basic prediction model, the accuracy of the prediction model is improved, the support is provided for the energy management decision of the system by predicting the future photovoltaic power generation power and the power generation amount, the system is helped to better arrange the charging and discharging plans of the energy storage equipment, the energy utilization is optimized, the power generation and energy storage strategies are adjusted according to the prediction result, and the cost is reduced and the economic benefit is improved.

In the above embodiment, the linear regression equation is:

,β、β₁、β₂、β₃、β₄、β₅、β₆、β₇、β₈、β₈ is a coefficient to be fitted and needs to be estimated through training data, epsilon is correction deviation, p is photovoltaic power generation predicted power, S _pv is photovoltaic panel state, S _inv is inverter state, L _lat is latitude of a geographic position, L _lon is longitude of the geographic position, C is cloud cover, S is irradiation intensity, T is temperature, H is humidity, and A is altitude.

Specifically, the power generation power P predicts the photovoltaic p＝β+β₁T+β₂H+β_3s+β₄C+β₅S_pv+β₆S_inv+β₇L_lat+β₈L_lon+β₉A+∈,, β、β₁、β₂、β₃、β₄、β₅、β₆、β₇、β₈、β₉ is a coefficient to be fitted, the coefficient to be fitted is estimated through training data, epsilon is correction deviation, P is historical power generation data, the photovoltaic panel state S _pv, the inverter state S _inv, the geographic position L can be obtained through latitude L _lat, longitude L _lon, cloud cover C, irradiation intensity S, temperature T, humidity H and altitude a, the irradiation coefficient is an irradiation amount normalization processing result, each coefficient can be estimated through regression analysis of the historical data, and the power generation amount of 24 hours in the future can be estimated in a time-by-time mode by using the trained model.

In some possible embodiments, the load prediction model comprises a second data collection unit, a second data preprocessing unit and a second prediction unit, wherein the second data collection unit is used for collecting historical electricity utilization data and electricity utilization plans, the second data preprocessing unit is used for carrying out normalization processing on the collected data, and the second prediction unit is used for constructing a state equation and an observation equation to predict future loads according to the data processed by the second data preprocessing unit and the state space model.

As shown in fig. 3, the core-most part is a cloud platform and a site end EMS, which form an energy management system of a cloud-edge integrated optical storage system, so as to facilitate understanding of the object-oriented system and the application scope of the patent. In fig. 3, the system is mainly divided into a time-varying energy system such as a photovoltaic and inverter, a fan and a converter, a storage energy system such as a lithium battery/sodium battery and battery management system BMS, a bi-directional inverter PCS and a super capacitor, a supplementary energy system such as a gasoline/diesel generator and a controller ECU, a fuel cell, an energy storage support system such as a thermal management system and a fire protection system, a detection execution system such as an electric energy meter, a switch component and the like, and a load system such as an adjustable load and a necessary load.

The method comprises the steps of collecting historical electricity utilization data and electricity utilization plans through a second data collecting unit, carrying out normalization processing on the collected data through a second data preprocessing unit to achieve consistency and comparability of the data, and finally constructing a state equation and an observation equation by a second prediction unit according to the processed data and a preset state space model to predict future loads. The load prediction model is used for predicting future load by integrating a data collection, preprocessing and prediction unit based on a state space model, and the future power consumption is predicted by importing a power consumption plan (long-term and temporary) and historical power consumption information, the data is normalized, and the load prediction model is input to predict the future load, so that the historical power consumption data and the power consumption plan can be efficiently utilized, and an accurate state equation and an observation equation are constructed after normalization processing, so that the accurate prediction of the future load is realized.

In the above embodiment, the state equation is: The observation equation is:

The load is predicted using a state space model that describes the change in state of the system by using state equations, and the change between the values and the state equations is observed using observation equations.

Specifically, the load is predicted using a state space model that describes the change in state of the system by using state equations, which are the changes between the observed values and the state equationsThe observation equation isX _t is the state vector at time t, Y _t is the predicted vector at time t (e.g., predicted load), u _t is the control input at time t (historical power usage data), A, B, C, D is the system matrix, and w _t and v _t are the process noise and the observation noise. The dynamic behavior of the system is described by selecting the appropriate state variables.

One embodiment is to select the trend, seasonal component, and random fluctuation of the load as state variables. And constructing a state equation describing the evolution of the state variable according to the historical electricity utilization data and the electricity utilization plan.

The trend and seasonal components of the load are T _t、S_t respectively, then the state equation can be written as T _t＝T_t-1+b_t+w_1t and S _t＝S_t-1+w_2t;

Where b _t is the trend variance, w _1t and w _2t are added noise, describing the relationship between observations (i.e., load data) and state variables, y _t＝T_t+S_t+D_t+v_t;D_t is the impact of the power usage plan, v _t is the observed noise, and finally predicting future loads.

In the above embodiment, the seasonal components are strongly related, for example, the electricity consumption of the air conditioner in summer is obviously higher than that in autumn, and the load trend is that the load in a certain season changes in a period of time, for example, a day, even a week, etc.;

The method specifically comprises the steps of estimating parameters of a state space model by using a Bayesian method, and finally carrying out state estimation and prediction by using a Kalman filter (KALMAN FILTER). The Kalman filter is deleted for state estimation and prediction, and only parameters of the state space model are reserved for estimation by using a Bayesian method, which is an existing and universal method.

In some possible embodiments, the electricity price management model comprises a third data collection unit, a third data preprocessing unit, a data visualization unit and a peak-valley period identification unit, wherein the third data collection unit is used for obtaining electricity price data from an electric company or electric market operator website, the third data preprocessing unit is used for filling or deleting missing values, ensuring that the electricity price data has consistent time stamp formats and is aligned with the data, the data visualization unit is used for displaying the change of the electricity price along with time, and the peak-valley period identification unit is used for automatically identifying peak periods, average periods and valley periods in the electricity price by using a K-means clustering algorithm.

The method comprises the steps of data collection, data preprocessing, data visualization and electricity price peak-valley period identification, wherein the data format obtained from an electric company or an electric market operator website can be CSV file, JSON file or through an API interface, the data preprocessing comprises the steps of filling or deleting missing values, guaranteeing that all electricity price data have consistent timestamp formats, finally guaranteeing data alignment if the data come from a plurality of sources, the data visualization comprises the step of using a time sequence diagram to show the change of electricity price along with time, and the fourth step of peak-valley period identification comprises the step of using a K-means clustering unsupervised learning algorithm to automatically identify three periods, namely a peak period, a flat period and a valley period in the electricity price.

Through integrating data collection, preprocessing, visualization and peak-valley period automatic identification functions, the electric power price data acquired from an electric power company or a market operator are effectively integrated and optimized, the consistency and the integrity of the data are ensured, the possibility of manual intervention and errors is reduced, meanwhile, the electric power price change trend is displayed in an intuitive mode, the peak period, the normal period and the valley period in the electric power price are accurately identified through a K-means clustering algorithm, and a reasonable electric power scheduling scheme is optimized through the change rule of the electric power price, so that the optimal electric power resource configuration and the planning of economic benefits are realized.

As shown in fig. 2, in some possible embodiments, the EMS controller includes a battery management module for monitoring the SOC of the battery and the remaining capacity of the battery that is currently chargeable.

The method comprises the steps of setting maximum and minimum SOC thresholds of a battery to be 95% and 15% respectively through a cloud end issuing instruction to an EMS, predicting peak-to-valley time periods of electricity price by a cloud end electricity price prediction model, and predicting power consumption in the future day by a cloud end load prediction model.

The photovoltaic power generation management module is used for monitoring photovoltaic power generation capacity in real time and reporting the photovoltaic power generation capacity to the cloud end, and helping the cloud end to correct the predicted value of the photovoltaic power generation capacity in real time.

The cloud photovoltaic power generation prediction model predicts photovoltaic power generation, predicts photovoltaic power generation amount in a future day according to historical data and weather forecast, the EMS end monitors the photovoltaic power generation amount in real time and reports the photovoltaic power generation amount to the cloud to help the cloud to correct predicted values of the photovoltaic power generation amount in real time, and battery management is divided into battery charging management and discharging management, and SOC monitoring monitors the SOC of the battery in real time.

And the charging management module calculates the SOC value of the battery pre-charge at night according to the peak-valley time period of the power consumption, the photovoltaic power generation and the power price of the load in the future day, and controls the battery to charge.

The method comprises the steps of predicting the power consumption of a load in the future day, calculating the SOC value of the battery pre-charge in the evening by an energy management system of the cloud through a peak-valley time period of the photovoltaic power generation amount in the future day and the power price in the future day, transmitting data to an EMS, controlling the battery to be charged to the value of the SOC appointed by the cloud by the EMS, and charging the battery preferentially when the photovoltaic power generation amount in the day is larger than the current load demand and the SOC of the battery is lower than the maximum threshold. When the electricity price is in the off-peak period and the battery SOC is below the maximum threshold, charging from the grid.

And the discharging management module is used for preferentially discharging the battery or reducing power grid electricity purchasing according to the electricity price time period and the battery SOC.

When the electricity price is in a peak period and the battery SOC is higher than a minimum threshold value, the battery is preferably used for discharging so as to reduce power grid electricity purchasing. When the photovoltaic power generation is insufficient to meet the load demand and the battery SOC is above a minimum threshold, the battery is used for discharging.

As shown in FIG. 2, the EMS end monitors the anti-reflux table in real time, if the electric power is monitored to be refluxed to the power grid, the charging and discharging strategy of the photovoltaic power generation or the battery is adjusted to ensure that no redundant electric power is refluxed to the power grid, data are uploaded to the cloud end, the cloud end calculates benefits based on the real-time data, meanwhile, under the condition that the photovoltaic power and the battery SOC are higher than the lowest threshold value and can meet the load requirement, the change value of the photovoltaic power is input into the fuzzy logic control model, the output of the model is used as the output power of the inverter, the quick adjustment of the output power of the inverter is realized, the power balance of the photovoltaic system and the load is ensured, and the power taking from the power grid is prevented.

In the process of inputting the change value of the photovoltaic power into the fuzzy logic control model, fuzzy logic is removed, a classical PID algorithm is used for adjusting, namely, a proportional, integral and differential mode is adopted, namely, output is input into the PID algorithm,

The control output e _t＝P_b,t+1-P_b,t,K_pK_iK_d is proportional, integral and differential gain, wherein the PID is an incremental PID, and the use of the incremental PID has the advantages of preventing excessive system oscillation and reducing the phenomena of power grid electricity taking and countercurrent under the condition that energy storage and photovoltaic can meet the load.

According to the application, through the energy management logic, chargeable quantity, photovoltaic predicted power and generated energy are comprehensively considered, peak Gu Tao benefit and safe anti-backflow logic are overlapped, firstly, battery State (SOC) and current chargeable residual capacity of a battery are obtained through a battery management module, the photovoltaic predicted power is predicted power of photovoltaic power generation in a future period, the current generated energy of the photovoltaic is actual generated energy of a current photovoltaic system, local peak valley electricity prices are fluctuation of grid electricity prices in different periods, and the anti-backflow strategy is used for preventing redundant electric power from flowing back to a grid and protecting the grid and equipment.

Referring to fig. 4, in a second aspect of the present application, there is provided an energy management method based on a cloud-edge integrated optical storage system, the method comprising:

s1, collecting information of the light storage device module and the load module through an EMS controller, and sending data to the cloud module.

S2, the cloud module predicts future photovoltaic power generation power and power generation capacity through a photovoltaic power generation prediction model, predicts future load through a load prediction model, and analyzes electricity price peak-valley time periods through an electricity price management model.

S3, executing control logic for energy management in the EMS controller according to the battery SOC, chargeable quantity, future photovoltaic power generation power and power generation quantity, superposition peak Gu Tao and safety anti-backflow logic.

And S4, monitoring the photovoltaic power generation amount, the battery SOC and the anti-backflow condition in real time by the EMS controller, and adjusting the photovoltaic power generation and battery charging and discharging strategies according to the instruction of the cloud module.

The method comprises the steps of deploying a photovoltaic power generation prediction model, a power price management model and a load prediction model on a cloud end module, deploying an EMS (energy management system) controller on an edge end module, collecting information of a photovoltaic storage device module and a load module, sending the information to the cloud end, predicting future photovoltaic power generation power/electricity quantity through the photovoltaic power generation prediction model, predicting future load through the load prediction model, analyzing a power peak-valley period through the power price management model, executing control logic of energy management in the EMS controller according to a battery SOC (state of charge), chargeable capacity, photovoltaic predicted power and electricity generation capacity, a superposition peak Gu Tao and safe anti-reflux logic, executing battery charging management, discharging management and anti-reflux strategy according to instructions and predicted data issued by the cloud end module, and improving the utilization efficiency of energy, so that the whole energy management system has high intelligent and automatic levels, and can make a rapid and accurate decision according to real-time data.

In the above embodiment, the present application further includes S5, where the cloud module calculates benefits based on the real-time data, and optimizes the energy management policy according to the benefits.

And optimizing the energy management strategy according to the profit situation by acquiring the profit situation in real time, carrying out scheme design from the profit maximization angle, and optimizing the energy management strategy.

In the embodiment, a linear regression mode is adopted for the photovoltaic power generation prediction model, a state space model is used for the load prediction model, a K-means clustering algorithm is used for the electricity price management model, control logic of energy management comprises battery management, photovoltaic power generation management, charging management, discharging management and anti-reflux strategies, and an EMS controller dynamically adjusts the photovoltaic power generation and battery charging and discharging strategies according to cloud module prediction data and real-time data so as to achieve energy optimization management.

The application is suitable for energy management of a photovoltaic energy storage system, and is also suitable for energy management of an expansion system based on light storage, such as a system with a combined fan, a diesel engine and the like.

The application relates to photovoltaic power generation prediction, load energy consumption prediction, peak-valley brix based on predicted energy consumption and other aspects, and relates to independent subsystems in the parts above.

The foregoing describes specific embodiments of the present application. It is to be understood that the application is not limited to the particular embodiments described above, and that various changes and modifications may be made by one skilled in the art within the scope of the claims without affecting the spirit of the application. The above-described preferred features may be used in any combination without collision.

Claims

1. The energy management system based on the cloud-edge integrated optical storage system is characterized by comprising a cloud end module and an edge end module;

the cloud module predicts future photovoltaic power generation power and power generation capacity through the photovoltaic power generation prediction model by receiving data acquired by the EMS controller, predicts future load through the load prediction model, analyzes electricity price peak-valley period through the electricity price management model and sends predicted data to the edge end module;

The edge end module performs energy management according to the battery SOC, chargeable quantity, future photovoltaic power generation power and generated energy, superposition peak Gu Tao and safety anti-backflow logic through the EMS controller;

The EMS controller monitors the photovoltaic power generation amount in real time and reports the photovoltaic power generation amount to the cloud module to help the cloud module to correct the predicted value of the photovoltaic power generation amount in real time;

the EMS controller dynamically adjusts photovoltaic power generation and battery charging and discharging strategies according to the cloud module prediction data and the real-time data so as to realize energy optimization management;

the photovoltaic power generation prediction model adopts a linear regression mode, the load prediction model uses a state space model, and the electricity price management model uses a K-means clustering algorithm.

2. The cloud-edge integrated optical storage system-based energy management system of claim 1, wherein the photovoltaic power generation prediction model comprises:

3. The energy management system based on a cloud-edge integrated optical storage system of claim 2, wherein the linear regression equation is:

,β、β₁、β₂、β₃、β₄、β₅、β₆、β₇、β₈、β₉ is a coefficient to be fitted, which is estimated through training data, epsilon is correction deviation, p is photovoltaic power generation prediction data, S _pv is a photovoltaic panel state, S _inv is an inverter state, L _lat is latitude of a geographic position, L _lon is longitude of the geographic position, C is cloud cover, S is irradiation intensity, T is temperature, H is humidity, and A is altitude.

4. The cloud-edge integrated optical storage system-based energy management system of claim 1, wherein the load prediction model comprises:

5. The energy management system based on the cloud-edge integrated optical storage system as claimed in claim 4, wherein,

The state equation is X _t＝Ax_t-1+Bu_t+w_t;

the observation equation is Y _t＝Cx_t+Du_t+v_t;

6. The energy management system based on the cloud-edge integrated optical storage system as claimed in claim 4, wherein,

The method specifically comprises the steps of estimating parameters of a state space model by using a Bayesian method, and carrying out state estimation and prediction by using a Kalman filter.

7. The energy management system based on a cloud-edge integrated optical storage system of claim 1, wherein the electricity price management model comprises:

8. The energy management system based on a cloud-edge integrated optical storage system of claim 1, wherein the EMS controller includes:

9. An energy management method based on a cloud-edge integrated optical storage system, which is characterized by comprising the following steps:

The cloud module calculates benefits based on real-time data and optimizes an energy management strategy according to the benefits;

10. The energy management method based on the cloud-edge integrated optical storage system as claimed in claim 9, wherein,

The control logic of the energy management comprises battery management, photovoltaic power generation management, charge management, discharge management and anti-reflux strategy.