An energy storage energy optimization decision method, device and medium of a source network load storage system

Abstract

The present application relates to the technical field of energy optimization, and specifically provides a kind of energy storage energy optimization decision method, device and medium of source network load storage system, the method is the predicted power generation of renewable energy and the predicted power consumption of load prediction in future T time period obtained by prediction model, according to the relationship between electricity consumption and income, the income before and after using energy optimization algorithm is calculated by the predicted power generation and predicted power consumption obtained;According to the predicted data accuracy calculated by the predicted power generation and predicted power consumption according to historical data;Whether to use energy optimization algorithm is judged according to income and predicted data accuracy;According to the charge-discharge power of each energy storage device in source network load storage system is adjusted according to the optimized dispatch power.In power generation and power consumption prediction, through the dual mechanism of accuracy evaluation and yield improvement estimation, avoid the error optimization instruction issuing caused by prediction deviation, solve the dispatch mismatch problem caused by data inaccuracy in traditional method.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of energy optimization, and particularly relates to an energy storage energy optimization decision method, device and medium for a source-grid-load-storage system. BACKGROUND

[0002] In traditional power systems, the randomness of loads and the uncontrollability of power grid power changes with load size lead to low energy utilization efficiency and environmental pollution problems. In recent years, with the development of renewable energy technology, a new type of power system based on source-grid-load-storage integration has gradually become an effective way to solve the above problems. This new system integrates photovoltaic power generation, wind power generation, tidal power generation and other types of renewable clean energy, and combines the application of lithium ion batteries, flow batteries, sodium ion batteries and other energy storage units, aiming to realize the rational allocation of power grid power and non-renewable energy, to reduce industrial and commercial electricity costs and reduce environmental pollution. However, due to the influence of climate and weather changes on renewable energy, its power generation is unstable, and the energy and power of the energy storage unit are limited. If the power distribution of these resources cannot be effectively managed, it may lead to increased light / wind curtailment, unbalanced peak and valley electricity prices, and ultimately affect the economy and stability of the system.

[0003] The existing source-grid-load-storage energy optimization decision method mainly uses the direct correlation logic of "prediction-optimization". This process first predicts the power generation and load power consumption in the future unit time through a single or combined model, which is used as the input data of the optimization algorithm. Then, without considering the accuracy of the prediction results and the revenue potential assessment, the optimization algorithm is directly started based on the predicted data, the power generation and power consumption are substituted into the preset revenue function, and the "revenue maximization" is taken as the goal to solve the optimal energy distribution strategy under the equipment operation constraints. Finally, the optimized energy distribution scheme is sent to the local controller through the communication interface to drive the source-grid-load-storage equipment to adjust the operating state. Although this method improves the energy utilization efficiency to some extent, since the accuracy of the power generation and load power consumption prediction results is not assessed, once the prediction deviation is large, the optimization algorithm will generate a strategy based on incorrect data, which may lead to excessive discharge of energy storage, frequent start and stop of power generation equipment, and thus increase the system loss and operation risk. In addition, even if the prediction data is accurate, if the actual revenue improvement space is small, the execution of the optimization algorithm with high algorithm power consumption will also cause waste of computing resources, thereby reducing the overall operation efficiency of the system. Therefore, how to effectively assess the prediction quality and revenue improvement potential has become the key to improving the performance of the source-grid-load-storage system. SUMMARY

[0004] In order to solve the problems existing in the prior art, the present application provides an energy storage energy optimization decision method for a source-grid-load-storage system, comprising the following steps:

[0005] Step S1, obtaining the renewable energy power generation and load power consumption in the T time period in the actual data of the source network load storage system, and establishing the relationship between the power generation, power consumption and income;

[0006] Step S2, obtaining the predicted power generation and predicted power consumption of the renewable energy in the future T time period through the prediction model to obtain the predicted power generation and predicted power consumption;

[0007] Step S3, according to the relationship between the power generation, power consumption and income, calculating the income before and after using the energy optimization algorithm through the predicted power generation and predicted power consumption;

[0008] Step S4, calculating the predicted power generation accuracy θ 1 and the predicted power consumption accuracy θ 2 according to the historical data;

[0009] Step S5, obtaining the prediction confidence θ 1 and the predicted power consumption accuracy θ 2; θ

[0010] Step S6, if the income after using the energy optimization algorithm ≥ threshold value and the reliable income ≥ threshold value, using the energy optimization algorithm to obtain the scheduling power, wherein the reliable income is the product of the income after using the energy optimization algorithm and the prediction confidence;

[0011] Step S7, adjusting the charging and discharging power of each energy storage device in the source network load storage system according to the optimized scheduling power.

[0012] On the basis of the above scheme, the method for calculating the income before and after using the energy optimization algorithm in step S3 is:

[0013] Step S31: obtaining the renewable energy power generation in each t unit time in the future T time period , and the load power consumption in each t unit time in the future T time period ;

[0014] Step S32, calculating the total electricity cost of the future T time period without using the optimization algorithm :

[0015] ,

[0016] ,

[0017] ,

[0018] wherein t2-t1=T,​ P(t) represents the power exchanged between the source-grid-load-storage system and the large power grid at time t, C(t) represents the electricity price at time t, , , Pd(t) represents the power consumed by the load at time t, Prg(t) represents the predicted power generated by the renewable energy at time t, and Pst(t) represents the sum of the charging and discharging powers of the energy storage unit at time t when no optimization algorithm is used;

[0019] Step S33, calculating the total electricity cost in the future T time period when the optimization algorithm is used :

[0020] ,

[0021] ,

[0022] ,

[0023] wherein t2-t1=T, P(t) represents the power exchanged between the source-grid-load-storage system and the large power grid at time t, C(t) represents the electricity price at time t, , , Pd(t) represents the power consumed by the load at time t, Prg(t) represents the predicted power generated by the renewable energy at time t, and Pst(t) represents the sum of the charging and discharging powers of the energy storage unit at time t when the optimization algorithm is used;

[0024] Step S34, calculating the total electricity cost difference in the future T time period :

[0025] .

[0026] Optionally, the step S31 is to obtain the sum of the renewable energy power generated every t unit time in the future T time period and the sum of the load power consumed every t unit time in the future T time period or to obtain the difference D between the renewable energy power generated every t unit time and the load power consumed every t unit time in the future T time period.

[0027] In a preferred embodiment, the step S1 is:

[0028] Step S11, obtaining the sum of the power generated and the sum of the power consumed by the load in the T time period in the actual data;

[0029] Step S12, taking the sum of the power generated and the sum of the power consumed by the load obtained in S11 as the target to perform cluster analysis;

[0030] Step S13: Input the total power generation and total power consumption of the cluster centers into the local scheduling algorithm and the energy optimization algorithm respectively to obtain the difference in total electricity cost before and after the energy optimization algorithm.

[0031] Step S14: The total power generation and the total load power consumption in the actual data time period T are used as input variables, and the total electricity cost difference is used as the dependent variable for linear fitting to obtain the relationship between power generation, electricity consumption and revenue.

[0032] In another preferred embodiment, step S1 is:

[0033] Using the total renewable energy power generation per unit time (t) and the total load power consumption per unit time (t) in the actual time period T as input variables, and the difference D between renewable energy power generation per unit time (t) and load power consumption per unit time (t) in the actual time period T, or the difference D between renewable energy power generation per unit time (t) and load power consumption per unit time (t) in the actual time period T as input variables; using the difference in total electricity cost before and after the energy optimization algorithm calculated by the local scheduling algorithm and the energy optimization algorithm respectively as the prediction target, deep learning or neural network model is used to learn and obtain the mathematical correspondence between power generation and consumption and revenue.

[0034] Based on the above scheme, step S4 includes:

[0035] Based on historical data and the input parameters of power generation and power consumption, similar days for predicting power generation and power consumption are identified. The average prediction accuracy of these similar days is then used as the accuracy of future prediction days. The input parameters for power generation include at least the forecast weather conditions, cloud cover rate, and average temperature of the prediction day obtained from public meteorological data. The input parameters for power consumption include at least the daily production plan, work and leave information, or the average temperature of the prediction day obtained from public meteorological data.

[0036] Specifically, in step S5, the confidence level is predicted. θ The calculation is as follows:

[0037] Determined based on historical data θ 1 and θ 2. The weights of the impact on returns are multiplied by the impact coefficient when calculating the prediction confidence: .

[0038] Based on the above scheme, the specific steps for obtaining the scheduling power using the energy optimization algorithm are as follows:

[0039] Step S61: Establish the optimization objective function, using a single-objective economic optimization function:

[0040] ,

[0041] Wherein, t2-t1=T, is the total electricity cost minimum value after using the optimization algorithm;

[0042] Step S62, the algorithm constraint condition is established, including the power upper and lower limit constraint A1 of the energy storage unit, the SOC upper and lower limit constraint A2 of the energy storage unit and the power equal constraint A3 of power supply and power consumption:

[0043] A1: min(Pi)≤Pi(t)≤max(Pi),

[0044] Wherein, min(Pi) represents the power lower limit of the i th energy storage unit, max(Pi) represents the power upper limit of the i th energy storage unit, and Pi(t) represents the power of the i th energy storage unit at t time;

[0045] A2: min(SOCi)≤SOCi(t)≤max(SOCi),

[0046] Wherein, min(SOCi) represents the SOC lower limit of the i th energy storage unit, max(SOCi) represents the SOC upper limit of the i th energy storage unit, and SOCi(t) represents the SOC of the i th energy storage unit;

[0047] A3: ;

[0048] Step S63, according to the optimization objective function and the constraint condition, the pareto optimization is carried out, the solution meeting the condition is found, and the algorithm is terminated, and the charging and discharging power of each energy storage unit in each unit time in the future T time period is output.

[0049] Secondly, the application discloses an energy storage energy optimization decision device of a source network load storage system, wherein the system is configured to execute the energy storage energy optimization decision method as described above, and the device comprises:

[0050] A first calculation module is configured to acquire the power generation of renewable energy and the load power consumption in a T time period in actual data of the source network load storage system, and establish the relationship among the power generation, the power consumption and the income;

[0051] A prediction module is configured to acquire the predicted power generation of renewable energy and the load predicted power consumption in a T time period in the future through a prediction model, and obtain the predicted power generation and the predicted power consumption;

[0052] A second calculation module is configured to calculate the income before and after using the energy optimization algorithm according to the predicted power generation and the predicted power consumption acquired by the income calculation module;

[0053] An accuracy calculation module is configured to calculate the predicted power generation accuracy and the predicted power consumption accuracy according to the historical data of the predicted power generation and the predicted power consumption θ 1θ 2; and according to the predicted power generation accuracy θ 1 and the predicted power consumption accuracy θ 2 obtain a predicted reliability θ

[0054] a judging module, if the income after using the energy optimization algorithm is greater than or equal to a threshold value and the reliable income is greater than or equal to a threshold value, the scheduling power is obtained by using the energy optimization algorithm;

[0055] an adjusting module, configured to adjust the charging and discharging power of each energy storage device in the source network load storage system according to the optimized scheduling power.

[0056] In a third aspect, the present application further provides an energy storage energy optimization decision device, the device comprising:

[0057] a memory storing executable program codes;

[0058] a processor coupled with the memory;

[0059] the processor invokes the executable program codes stored in the memory to execute the energy storage energy optimization decision method as described above.

[0060] In a fourth aspect, the present application discloses a computer readable storage medium storing a computer program, the computer program, the computer storage medium storing computer instructions, the computer instructions being invoked to execute the steps of the energy storage energy optimization decision method as described above.

[0061] Compared with the prior art, the present application has the following beneficial effects:

[0062] 1. The optimization algorithm is enabled through the double evaluation of the income and the prediction accuracy, which effectively avoids the resource waste and decision errors caused by blind optimization in the prediction deviation or low income scenarios.

[0063] 2. The prediction accuracy evaluation method based on historical similar day data is introduced, and the comprehensive prediction reliability is obtained by combining the influence of prediction error on income, which can dynamically reflect the influence degree of current prediction quality on optimization results, and provides a more scientific and reliable quantitative basis for whether to execute the optimization algorithm.

[0064] 3. The "prediction-evaluation-decision" closed-loop energy optimization decision system is established, the prediction quality and the income value are considered, and the traditional decision mode of single dependence on prediction data or simple income calculation is changed. Through comprehensive evaluation, it is judged whether to execute the optimization algorithm, the source network load storage system energy optimization decision is changed from "experience driven" "data driven" to "value driven", and the precision and economy of the decision are significantly improved. BRIEF DESCRIPTION OF DRAWINGS​

[0065] Figure 1 Flow chart of the overall method of energy optimization decision of the present application;

[0066] Figure 2 Flow chart for judging whether the present application can be energy optimized;

[0067] Figure 3 Flow chart for the specific embodiment of the present application. DETAILED DESCRIPTION

[0068] The application will be further described below in conjunction with specific embodiments.

[0069] Embodiment 1

[0070] As shown in Figure 1 and Figure 2 , the present application provides an energy storage energy optimization decision method of a source network load storage system, comprising the following steps:

[0071] Step S1, obtaining the power generation amount of renewable energy and the load power consumption amount in T time period in the actual data of the source network load storage system, and establishing the relationship between the power generation amount, the power consumption amount and the income;

[0072] In this embodiment, the step S1 specifically comprises:

[0073] Step S11, obtaining the total power generation amount and the total load power consumption amount in T time period in the actual data;

[0074] Step S12, taking the total power generation amount and the total load power consumption amount obtained in S11 as the target for cluster analysis;

[0075] Step S13, inputting the total power generation amount and the total load power consumption amount of the cluster center into the local scheduling algorithm and the energy optimization algorithm respectively, obtaining the total power consumption electricity fee difference before and after the energy optimization algorithm, and the total power consumption electricity fee difference is the income;

[0076] Step S14, taking the total power generation amount and the total load power consumption amount in T time period in the actual data as the input variable, and taking the total power consumption electricity fee difference as the dependent variable for linear fitting, to obtain the relationship between the power generation amount, the power consumption amount and the income.

[0077] Step S2, obtaining the predicted power generation power of renewable energy and the load predicted power consumption power in the future T time period through the prediction model, and further calculating the predicted power generation amount and the predicted power consumption amount;

[0078] The prediction model is established according to actual historical data. First, historical data is acquired, and a prediction model is established and trained according to the corresponding power generation of parameters such as weather conditions, cloud coverage, and average temperature in the historical data, and the corresponding power consumption of parameters such as daily production plans and work and leave conditions and average temperature.

[0079] Step S3, according to the relationship between the power generation, power consumption, and income obtained in step S1, the predicted power generation and predicted power consumption are taken as inputs to calculate the income before and after using the energy optimization algorithm, as a first evaluation index;

[0080] Specifically, the step S3 includes the following steps:

[0081] Step S31: acquiring renewable energy power generation per t unit time in a future T time period , and load power consumption per t unit time in the future T time period ;

[0082] Step S32, calculating the total electricity fee of the future T time period without using the optimization algorithm :

[0083] ,

[0084] ,

[0085] ,

[0086] wherein t2-t1=T, represents the power exchanged between the source-grid-load-storage system and the large power grid at t moment, represents the electricity price at t moment, , , respectively represent the power consumption of the load, the predicted power generation of the renewable energy, and the total power of the charging and discharging of the energy storage unit at t moment;

[0087] Step S33, calculating the total electricity fee of the future T time period using the optimization algorithm :

[0088] ,

[0089] ,

[0090] ,

[0091] wherein t2-t1=T, represents the power exchanged between the source-grid-load-storage system and the large power grid at t moment, Let t represent the electricity price at time t, defined as positive when purchasing electricity from the grid and negative when selling electricity to the grid. , , These represent the power consumption of the load at time t, the predicted power generation of renewable energy, and the total charging and discharging power of all energy storage units using optimization algorithms, respectively.

[0092] Step S34: Calculate the total electricity cost difference within the future time period T. :

[0093] .

[0094] Step S4: Calculate the accuracy of the predicted power generation based on historical data for both predicted power generation and predicted power consumption. θ 1. Accuracy of predicted electricity consumption θ 2;

[0095] Step S4 includes: finding similar days for predicted power generation and predicted power consumption from historical data based on the input parameters of power generation and power consumption, and using the average prediction accuracy of similar days for predicted power generation as the prediction accuracy of future prediction days. θ 1. The average prediction accuracy of similar electricity consumption days is used as the prediction accuracy of future prediction days. θ 2. The input parameters for power generation may include forecast weather conditions, cloud cover, average temperature, average air pressure, average humidity, and air pollution status obtained from public meteorological data for the predicted day; the input parameters for electricity consumption may include daily production plans, work and leave arrangements, or the average temperature of the predicted day obtained from public meteorological data.

[0096] It should be noted that the prediction accuracy for similar days is calculated based on the predicted power generation and power consumption data for similar days and the actual power generation and power consumption data.

[0097] Step S5, based on the accuracy of the predicted power generation θ 1 and the accuracy of the predicted electricity consumption θ 2. Obtain the prediction confidence level θ , θ As a second evaluation indicator .

[0098] Preferred options are determined based on historical data. θ 1 and θ 2. The weights of the impact on returns are multiplied by the impact coefficient when calculating the prediction confidence: ,in x 1, x 2 are respectively θ 1. Impact on returnsθ 2The impact on the income. In actual calculation, x 1, x 2may be greater than 1, so the obtained θ need to be further processed to make 0 θ ≤1, when θ <1, θ unchanged ， If θ ≥1, let θ 1.

[0099] Step S6, determining whether to use the energy optimization algorithm according to the first evaluation index and the second evaluation index;

[0100] If the income after using the energy optimization algorithm is less than the threshold value, the local control algorithm is used without using the optimization algorithm; if the income after using the energy optimization algorithm is greater than or equal to the threshold value, the reliable income is calculated, that is, the income after using the energy optimization algorithm * the predicted reliability θ , if the reliable income is greater than or equal to the threshold value, the energy optimization algorithm is used to obtain the scheduling power.

[0101] Generally, the threshold value can be determined according to the cost of using the energy optimization algorithm, for example, the cost of using the power generation prediction data (such as the cost of calling the public weather data API) X yuan, the cost of using the load prediction data Y yuan, the server running cost used by the algorithm and the communication cost between the local and the cloud Z yuan, so the comprehensive cost is X+Y+Z yuan, wherein the power generation prediction and the load prediction are steps that must be performed in the evaluation process, that is, X+Y is the cost that must be used, if the optimization algorithm is used, Z yuan needs to be spent on the basis of X+Y, if the income is greater than Z yuan, the optimization algorithm is used after the local control algorithm, so the threshold value can be determined as Z yuan.

[0102] Preferably, the threshold value can also be determined according to the actual application of people's will or scene.

[0103] In this embodiment, by finding the date similar to the predicted day as the similar day from the historical data, and calculating the prediction accuracy of the similar day as the prediction accuracy of the future prediction day, the prediction accuracy is calculated by using the accuracy of the historical similar day, and the prediction reliability is adjusted according to the different influences of the prediction accuracy on the income, whether to use the energy optimization algorithm is further judged, and unnecessary calculation resource consumption is reduced.

[0104] According to this embodiment, the optional energy optimization algorithm includes genetic algorithm, particle swarm optimization algorithm, sparrow search algorithm, ant colony algorithm or grey wolf optimization algorithm and the like. Further, the specific steps of obtaining the scheduling power by using the energy optimization algorithm are as follows:

[0105] Step S61, an optimization objective function is established, and a single-target economic optimization function is adopted:

[0106] ,

[0107] wherein t2-t1=T, is a total electricity fee minimum value after using an optimization algorithm, the smaller the value is, the less the total electricity fee in the T time period is, and the optimal fitness is; represents power exchanged with a large power grid at t time, and an expression thereof is:

[0108] ,

[0109] respectively represent predicted power of a load at t time, predicted power generation of a renewable energy source, and total charging and discharging power of all energy storage units when the optimization algorithm is used; represents an electricity price at t time, is positive when electricity is purchased from the power grid, and is negative when electricity is sold to the power grid, and is different according to a current system buying and selling state, and an expression thereof is:

[0110] .

[0111] Step S62, algorithm constraint conditions are established, including power upper and lower limit constraints A1 of the energy storage units, SOC upper and lower limit constraints A2 of the energy storage units, and power consumption and power supply power equalization constraints A3:

[0112] A1: min(Pi)≤Pi(t)≤max(Pi),

[0113] wherein min(Pi) represents a power lower limit of the i th energy storage unit, max(Pi) represents a power upper limit of the i th energy storage unit, and Pi(t) represents power of the i th energy storage unit at t time, which should be changed between the power upper and lower limits;

[0114] A2: min(SOCi)≤SOCi(t)≤max(SOCi),

[0115] wherein min(SOCi) represents an SOC use lower limit of the i th energy storage unit, max(SOCi) represents an SOC use upper limit of the i th energy storage unit, and SOCi(t) represents SOC use of the i th energy storage unit, and the SOC of the energy storage unit at t time should not exceed the use upper and lower limits;

[0116] A3: ;

[0117] Step S63, according to the optimization objective function and constraint conditions, find the solution that meets the conditions and terminate the algorithm, get the charging and discharging power of each energy storage unit in each t unit time in the future T period.

[0118] Step S7, according to the optimized scheduling power, adjust the charging and discharging power of each energy storage device in the source network load storage system.

[0119] The generated power vector of each energy storage unit in T period is sent to the local controller through the communication network. After the local controller receives the information, if the judgment result is to perform the optimization algorithm, adjust the charging and discharging power of each energy storage device in the source network load storage system according to the optimization result.

[0120] Embodiment 2

[0121] The difference between this embodiment and the implementation of embodiment 1 is that the step S1 of establishing the relationship between the power generation, power consumption and income is:

[0122] The input variables in the actual data are the renewable energy power generation and load power consumption in T period, which can be the total renewable energy power generation in each t unit time in T period The total load power consumption in each t unit time in T period It can also be the difference between the renewable energy power generation and the load power consumption in each t unit time in T period Wherein It can also be the renewable energy power generation in each t unit time in T period ( And the load power consumption It should be noted that the renewable energy power generation in each t unit time is a vector, and the total renewable energy power generation in each t unit time is the result of vector addition. The two are different, and different input variable forms are suitable for different prediction tasks.

[0123] Further, the total electricity bill difference before and after the energy optimization algorithm is calculated by the local scheduling algorithm and the energy optimization algorithm as the prediction target, and the deep learning or neural network model is used for learning to obtain the mathematical corresponding relationship between the power generation, power consumption and income.

[0124] The embodiment introduces a deep learning or neural network model to model and learn the electricity bill difference under local scheduling and energy optimization strategy, and establishes a nonlinear mapping relationship between power generation, power consumption and optimization income. The prediction accuracy and generalization ability of the system for optimization gain are effectively improved, the efficient judgment of the optimization strategy is realized, the calculation cost is greatly reduced while ensuring the economy, and the system response speed and intelligent level are improved.

[0125] Similarly, when using the above model, the input of step S31 can also be the total renewable energy power generation per t unit time in the future T time period and the total load power consumption per t unit time in the future T time period , the difference D between the renewable energy power generation per t unit time in the future T time period and the load power consumption per t unit time in the future T time period, or the renewable energy power generation per t unit time in the future T time period and the load power consumption per t unit time in the future T time period . The predicted power generation of renewable energy and the predicted power consumption of load per t unit time in the source network load storage system in the future T time period are added respectively, and the total power generation and the total power consumption in the future T time period can be obtained.

[0126] Embodiment 3

[0127] As shown in Figure 3 , the specific judgment of energy optimization is taken as daily income as follows:

[0128] According to the prediction model, the daily power generation and the daily power consumption are predicted, and the daily income when using the energy optimization algorithm is estimated according to the model formed between the power generation, the power consumption and the income;

[0129] The accuracy of the predicted power generation θ 1 and the accuracy of the predicted power consumption θ 2 are calculated according to the parameters of similar days in the historical data respectively;

[0130] If the estimated daily income is greater than the threshold value, the prediction credibility θ is further calculated, and if the estimated daily income is less than the threshold value, the related message of not using the energy optimization algorithm is generated, and the energy distribution and strategy are issued;

[0131] If the prediction credibility is greater than the threshold value, the optimization algorithm is used to calculate the energy storage scheduling power, the energy distribution and the strategy are issued, and if the prediction credibility is less than the threshold value, the related message of not using the energy optimization algorithm is generated, and the energy distribution and the strategy are issued.

[0132] Specifically, the generated related message not using the optimization algorithm is sent to the local controller through a communication network (such as 5G, industrial Ethernet, etc.), and after the local controller receives the information, if the judgment result is that the optimization algorithm needs to be performed, the charging and discharging power of each energy storage device in the source network load storage system is adjusted according to the result of solving the optimization target; if the judgment result is that the optimization algorithm does not need to be performed, the local controller maintains the local preset control strategy.

[0133] In a second aspect, the present application discloses an energy storage energy optimization decision device of a source network load storage system, which is configured to perform the energy storage energy optimization decision method as described above, comprising:

[0134] A first calculation module is configured to obtain the power generation of renewable energy and the power consumption of load in a T time period in the actual data of the source network load storage system, and establish a relationship between the power generation, the power consumption and the income;

[0135] A prediction module is configured to obtain the predicted power generation of renewable energy and the predicted power consumption of load in a future T time period through a prediction model, and obtain the predicted power generation and the predicted power consumption;

[0136] A second calculation module is configured to calculate the income before and after using the energy optimization algorithm according to the predicted power generation and the predicted power consumption obtained by the income calculation module;

[0137] An accuracy calculation module is configured to calculate the predicted power generation accuracy θ 1 and the predicted power consumption accuracy θ 2 according to the historical data of the predicted power generation and the predicted power consumption, and obtain the prediction credibility θ 1 according to the predicted power generation accuracy θ 2 and the predicted power consumption accuracy θ .

[0138] A judgment module is configured to use the energy optimization algorithm to obtain the dispatching power if the income after using the energy optimization algorithm is greater than or equal to a threshold value and the credible income is greater than or equal to a threshold value;

[0139] An adjustment module is configured to adjust the charging and discharging power of each energy storage device in the source network load storage system according to the optimized dispatching power.

[0140] The energy optimization decision device described above estimates the income according to the mathematical relationship between the power generation, the power consumption and the income, and determines whether to use the energy optimization algorithm according to the double judgment mechanism of the income and the accuracy of the prediction data, which effectively avoids the waste of resources and the decision errors caused by blind optimization due to prediction deviation or low income scenarios.

[0141] In a third aspect, the present disclosure provides a device for energy storage energy optimization decision, which can include one or more memories and one or more processors. The memory is configured to store computer-executable instructions non-transitorily; and the processor is configured to execute the computer-executable instructions, which can cause the processor to perform one or more steps of the method for energy storage energy optimization decision according to any of the embodiments of the present disclosure.

[0142] The specific implementation of each step of the method for energy storage energy optimization decision and the related explanations can refer to the related content in the above-mentioned embodiments of the method for energy storage energy optimization decision, which will not be repeated here. In one embodiment, the processor and the memory can communicate with each other directly or indirectly. For example, the processor and the memory can communicate through a network connection. The network can include a wireless network, a wired network, and / or any combination of a wireless network and a wired network, and the type and function of the network are not limited in the present disclosure. For another example, the processor and the memory can also communicate through a bus connection. The bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. For example, the processor and the memory can be arranged at a remote data server end (cloud end) or a distributed energy system end (local end), and can also be arranged at a client end (for example, a mobile device such as a mobile phone). For example, the processor can be a central processing unit (CPU), a tensor processing unit (TPU), or a graphics processing unit (GPU), etc. which has data processing capability and / or instruction execution capability, and can control other components in the device to perform the desired functions. The central processing unit (CPU) can be X86 or ARM architecture, etc.

[0143] In one embodiment, the memory can include one or more computer program products in any combination, which can include various forms of computer readable storage media, such as volatile memory and / or non-volatile memory. For example, the volatile memory can include random access memory (RAM), cache memory, etc. The non-volatile memory can include read-only memory (ROM), hard disk, erasable programmable read-only memory (EPROM), compact disc read-only memory (CD-ROM), USB memory, flash memory, etc. One or more computer-executable instructions can be stored on the computer readable storage medium, and the processor can execute the computer-executable instructions to implement various functions of the device. Various application programs and various data, as well as various data used and / or generated by the application programs, etc. can also be stored in the memory.

[0144] In a fourth aspect, the present application discloses a computer readable storage medium, which stores a computer program, the computer program, the computer storage medium stores computer instructions, the computer instructions are called to execute one or more steps of the energy storage energy optimization decision method.

[0145] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functionality, and operations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowcharts or block diagrams can represent a module, a segment, or a portion of code, which comprises at least one executable instruction for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or in the reverse order, depending on the functionality involved. It will also be noted that each block in the block diagrams and / or flowchart illustrations, and combinations of blocks in the block diagrams and / or flowchart illustrations, can be implemented by special purpose hardware-based systems that perform the specified functions or operations, or combinations of special purpose hardware and computer instructions.

[0146] In general, various example embodiments of the present application can be implemented in hardware or special-purpose circuits, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while other aspects can be implemented in firmware or software which can be executed by a controller, microprocessor or other computing device, Although various example embodiments of the present application can have been described in connection with one embodiment, it is not intended to be limited to that embodiment or only that embodiment. It is intended that changes and modifications can be made by those skilled in the art to which this application pertains that do not depart from the spirit and scope of the claims.

[0147] The specific implementation described above is only a preferred embodiment of the present application, and is not intended to limit the present application. For those skilled in the art, various modifications and changes can be made to the present application without departing from the spirit and principles of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0148] The above describes the specific embodiments of the present application, but is not intended to limit the protection scope of the present application. Those skilled in the art should understand that various modifications or changes made to the technical solutions of the present application without inventive labor are still within the protection scope of the present application.

Claims

1. An energy storage energy optimization decision method of a source network load storage system, characterized in that, Comprising: Step S1, obtaining the renewable energy power generation and load power consumption in T time period in the source network load storage system actual data, and establishing the relationship between power generation, power consumption and income; Step S2, obtaining the predicted power generation and load predicted power consumption of the renewable energy in the future T time period through the prediction model to obtain the predicted power generation and predicted power consumption; Step S3, based on the relationship between power generation, power consumption and income, calculating the total electricity fee before and after using the energy optimization algorithm according to the obtained predicted power generation and predicted power consumption, and calculating the income according to the difference between the total electricity fee before using the energy optimization algorithm and the total electricity fee after using the energy optimization algorithm; Step S4, calculating the predicted power generation accuracy according to historical data θ 1 and the predicted power consumption accuracy θ 2; Step S5, obtaining a prediction accuracy of the predicted power generation amount based on the predicted power consumption amount θ 1 and the predicted power consumption amount θ 2 to obtain a prediction reliability θ ; Step S6, if the income after using the energy optimization algorithm is greater than or equal to the threshold value and the credible income is greater than or equal to the threshold value, the scheduling power is obtained by using the energy optimization algorithm, and the credible income is the product of the income after using the energy optimization algorithm and the prediction credibility; Wherein, the step of obtaining the scheduling power by using the energy optimization algorithm is: Step S61, establishing an optimization objective function, and adopting a single objective economic optimization function: Where t2-t1=T, minF D To minimize the total electricity cost after using the optimization algorithm, P pwr ( t () represents the electricity price at time t. P' grid ( t () represents the power exchanged between the source-grid-load-storage system and the main power grid at time t after applying the energy optimization algorithm; P load ( t ), P pv ( t ), P' bat ( t ) represent the power consumption of the load at time t, the predicted power generation of renewable energy, and the total charging and discharging power of all energy storage units using the optimization algorithm, respectively; P use ( t Let t be the electricity purchase price at time t. P sell ( t Let t be the electricity price at time t; Step S62, establishing algorithm constraint conditions, including power upper and lower limit constraints A1 of the energy storage unit, SOC upper and lower limit constraints A2 of the energy storage unit, and power consumption and power supply power equalization constraints A3; Step S63, performing Pareto optimization according to the optimization objective function and the constraint conditions, finding a solution that meets the conditions, and then terminating the algorithm, and outputting the charging and discharging power of each energy storage unit at t time in the future T time period; Step S7, adjusting the charging and discharging power of each energy storage device in the source network load storage system according to the optimized scheduling power. 2.The energy storage energy optimization decision method of a source-network-load-storage system according to claim 1, characterized in that, The method for calculating the income before and after using the energy optimization algorithm in step S3 is: Step S31: acquire renewable energy power generation power at time t in future T time period P pv ( t ), load power consumption at time t in future T time period P load ( t ); Step S32, calculate the total electricity cost in the future T period of time when the optimization algorithm is not used : where t2-t1=T, P grid ( t ) represents the power exchanged between the source-grid-storage system and the large grid at time t, P pwr ( t ) represents the electricity price at time t, P load ( t ), P pv ( t ), P bat ( t ) respectively represent the power consumed by the load, the predicted power generated by the renewable energy source and the total power charged and discharged by the energy storage unit at time t. Step S33, calculate total electricity cost in future T time period when using optimization algorithm : where t2-t1=T, P' grid ( t ) represents the power exchanged between the source-grid-storage system and the large grid at time t, P pwr ( t ) represents the electricity price at time t, P load ( t ), P pv ( t ), P' bat ( t ) represents the power consumed by the load, the predicted power generated by the renewable energy source and the total power charged and discharged by the energy storage unit using the optimization algorithm at time t, respectively. Step S34, calculate the total electricity bill difference in the future T time period D price : 。 3. The energy storage energy optimization decision method of a source-network-load-storage system according to claim 1, characterized in that, The step S1 is: Step S11, obtaining the total power generation and total load power consumption in T time period in the actual data; Step S12, taking the total power generation and total load power consumption obtained in S11 as the target for cluster analysis; Step S13, inputting the total power generation and total load power consumption of the cluster center into the local scheduling algorithm and the energy optimization algorithm respectively to obtain the total power consumption difference before and after the energy optimization algorithm; Step S14, taking the total power generation and total load power consumption in T time period in the actual data as the input variable, and taking the total power consumption difference as the dependent variable for linear fitting to obtain the relationship between power generation, power consumption and income.

4. The energy storage energy optimization decision method of a source-network-load-storage system according to claim 1, characterized in that, The step S1 is: Taking the total renewable energy power generation in T time period and the total load power consumption in T time period, the difference D between the renewable energy power generation and the load power consumption in T time period, or the renewable energy power generation and the load power consumption at t time in T time period as the input variable; taking the total power consumption difference before and after the energy optimization algorithm calculated by the local scheduling algorithm and the energy optimization algorithm as the prediction target, using the deep learning or neural network model for learning to obtain the mathematical corresponding relationship between power generation and power consumption and income.

5. The energy storage energy optimization decision method of a source-network-load-storage system according to claim 1, wherein, The step S4 comprises: The similar day is found from historical data according to the input parameters of power generation and the input parameters of power consumption, the average of the prediction accuracy of the similar day is obtained as the accuracy of the future prediction day; the input parameters of power generation at least include the predicted weather condition of the prediction day obtained from public weather data, the cloud cover rate and the average temperature; the input parameters of power consumption at least include the daily production plan, the work and holiday situation or the average temperature of the prediction day obtained from public weather data.

6. The energy storage energy optimization decision method of a source-network-load-storage system according to claim 5, wherein, The step S5 of predicting the reliability θ The calculation is: determined from historical data θ 1 and θ 2 impact on the yield, the impact factor is multiplied when calculating the prediction reliability: .

7. The energy storage energy optimization decision method of a source-network-load-storage system according to claim 2, characterized in that, The power upper and lower limit constraints A1 of the energy storage unit are: min ( Pi ) ≤Pi ( t ) ≤max ( Pi ), wherein, min ( Pi ) represents the power lower limit of the i-th energy storage unit, max ( Pi ) represents the power upper limit of the i-th energy storage unit, Pi ( t ) represents the power of the i-th energy storage unit at time t. The upper and lower limits of the SOC of the energy storage unit A2 are: min ( SOCi ) ≤SOCi ( t ) ≤max ( SOCi ), wherein, min SOCi represents the SOC lower limit of use of the i-th energy storage unit, i max SOCi represents the SOC upper limit of use of the i-th energy storage unit, SOCi t represents the SOC of use of the i-th energy storage unit,​​​​ The power consumption and power supply power equal constraint A3 is: .

8. An energy storage energy optimization decision device of a source network load storage system, characterized in that, The system is configured to perform the energy storage energy optimization decision method of any one of claims 1-7, comprising: A first calculation module is configured to obtain the power generation of renewable energy and the power consumption of load in a T time period in the actual data of the source network load storage system, and establish a relationship between the power generation, power consumption and income; A prediction module is configured to obtain the predicted power generation of renewable energy and the predicted power consumption of load in a future T time period through a prediction model, and obtain the predicted power generation and the predicted power consumption; A second calculation module is configured to calculate the income before and after using the energy optimization algorithm according to the predicted power generation and the predicted power consumption obtained by the income calculation module; an accuracy calculation module configured to calculate a predicted power generation accuracy according to historical data of the predicted power generation and the predicted power consumption θ 1and the predicted power consumption accuracy θ 2; and obtain a predicted credibility according to the predicted power generation accuracy θ 1and the predicted power consumption accuracy θ 2 θ ; A judgment module is configured to use the energy optimization algorithm to obtain the dispatching power if the income after using the energy optimization algorithm is greater than or equal to a threshold value and the reliable income is greater than or equal to a threshold value; An adjustment module is configured to adjust the charging and discharging power of each energy storage device in the source network load storage system according to the optimized dispatching power.

9. An energy storage energy optimization decision device, characterized by, The device comprises: a memory storing executable program codes; a processor coupled with the memory; the processor invokes the executable program codes stored in the memory to execute the energy storage energy optimization decision method of any one of claims 1-7.

10. A computer readable storage medium storing a computer program, characterized in that, The computer program, the computer storage medium stores computer instructions, the computer instructions are called to execute the steps of the energy storage energy optimization decision method of any one of claims 1 to 7.

Images