Charging and discharging control method and apparatus for energy storage device, and user-side energy storage system

Abstract

The present disclosure provides a charging and discharging control method and apparatus for an energy storage device, and a user-side energy storage system. The method comprises: acquiring current electrical energy information of an energy storage device, current power generation information of a new energy power generation device, current electricity consumption information of an electric device, and electricity price information and power grid outage information in a target scheduling period; obtaining electrical energy prediction information in the target scheduling period on the basis of the current power generation information and the current electricity consumption information; obtaining charging and discharging control information of the energy storage device in the target scheduling period on the basis of the current electrical energy information, the electricity price information, the power grid outage information, and the electrical energy prediction information; and sending the charging and discharging control information to the energy storage device, so that the energy storage device performs charging and discharging on the basis of the charging and discharging control information. The method considers in advance the net electricity load demand during a future power outage, so that a scheduling plan is dynamically and flexibly made, and the charging and discharging of an energy storage device are controlled, thereby reducing the electricity costs of users, and satisfying the electricity demand of the users during power grid outages.

Description

Charging and discharging control method and device of energy storage equipment and user-side energy storage system TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of energy storage systems, and particularly to a charging and discharging control method and device of energy storage equipment and a user-side energy storage system. BACKGROUND

[0002] In order to ensure the power demand of users during power grid outage, the energy storage equipment is usually charged before the power outage according to the power outage information. In the related art, the energy storage equipment is usually manually fully charged before the power outage, or the stored power in the energy storage equipment is controlled at a specified power level at a specified time. This way of controlling the energy storage equipment has poor flexibility, resulting in high user power cost or difficulty in fully meeting the user power demand during the power grid outage. SUMMARY

[0003] The embodiments of the present disclosure provide a charging and discharging control method and device of energy storage equipment and a user-side energy storage system to consider the net power demand during future power outage in advance, so as to dynamically and flexibly make a scheduling plan, control the charging and discharging of the energy storage equipment, reduce the user power cost, and meet the user power demand during the power grid outage.

[0004] In a first aspect, the embodiments of the present disclosure provide a charging and discharging control method of energy storage equipment, which includes: obtaining current power information of the energy storage equipment, current power generation information of a new energy power generation device, current power consumption information of a power consumption device, and price information and power grid outage information in a target scheduling period; obtaining power prediction information in the target scheduling period based on the current power generation information and the current power consumption information; obtaining charging and discharging control information of the energy storage equipment in the target scheduling period based on the current power information, the price information, the power grid outage information, and the power prediction information; and sending the charging and discharging control information to the energy storage equipment, where the charging and discharging control information is used for the energy storage equipment to perform charging and discharging control.

[0005] The power prediction information includes power generation power prediction information and power consumption load prediction information; obtaining the power prediction information in the target scheduling period based on the current power generation information and the current power consumption information includes: inputting the current power generation information into a preset power generation power prediction model to obtain the power generation power prediction information in the target scheduling period; or inputting the current power consumption information into a preset load prediction model to obtain the power consumption load prediction information in the target scheduling period; or the power prediction information includes net power consumption load prediction information; obtaining the power prediction information in the target scheduling period based on the current power generation information and the current power consumption information includes: inputting the current power generation information and the current power consumption information into a preset net load prediction model to obtain the net power consumption load prediction information in the target scheduling period.

[0006] The target scheduling period includes a plurality of time points, and the power prediction information includes power prediction information corresponding to each of the plurality of time points in the target scheduling period; the obtaining of the charge-discharge control information of the energy storage device in the target scheduling period based on the current power information, the electricity price information, the power grid outage information, and the power prediction information includes: generating a plurality of sets of scenario prediction information based on the power prediction information and a plurality of sets of preset prediction error information; each set of prediction error information includes prediction error information corresponding to each of the plurality of time points in the target scheduling period, each set of scenario prediction information includes scenario prediction information corresponding to each of the plurality of time points in the target scheduling period, and each set of scenario prediction information corresponds to a scenario probability; inputting the plurality of sets of scenario prediction information, the scenario probability corresponding to each set of scenario prediction information, and the electricity price information into a preset energy scheduling model, and obtaining the charge-discharge control information of the energy storage device in the target scheduling period when a target function value of the energy scheduling model is lowest and a constraint condition is met; the constraint condition is established based on the power prediction information, the current power information, and the power grid outage information.

[0007] After the generating of the plurality of sets of scenario prediction information based on the power prediction information and the plurality of sets of preset prediction error information, the method further includes: obtaining an information distance between each two sets of scenario prediction information in the plurality of sets of scenario prediction information to obtain a distance matrix; taking each set of scenario prediction information as current prediction information one by one, determining a probability distance sum of the current prediction information and each set of scenario prediction information other than the current prediction information based on an initial scenario probability of each set of scenario prediction information and the distance matrix to obtain a probability distance matrix; merging at least part of the plurality of sets of scenario prediction information based on the distance matrix and the probability distance matrix to obtain new scenario prediction information and a scenario probability corresponding to the new scenario prediction information; and updating the plurality of sets of scenario prediction information based on the new scenario prediction information.

[0008] The power prediction information includes power generation prediction information and power consumption load prediction information, and the scenario prediction information includes power generation scenario prediction information and load scenario prediction information; the plurality of sets of scenario prediction information are generated based on the power prediction information and a plurality of sets of preset prediction error information, including: for the plurality of sets of prediction error information corresponding to the power generation prediction information, superimposing each set of prediction error information on the power generation prediction information at a time point to obtain a plurality of sets of power generation scenario prediction information; for the plurality of sets of prediction error information corresponding to the power consumption load prediction information, superimposing each set of prediction error information on the power consumption load prediction information at a time point to obtain a plurality of sets of load scenario prediction information; or, the power prediction information includes net power consumption load prediction information, and the scenario prediction information includes net load scenario prediction information; the plurality of sets of scenario prediction information are generated based on the power prediction information and a plurality of sets of preset prediction error information, including: for the plurality of sets of prediction error information corresponding to the net power consumption load prediction information, superimposing each set of prediction error information on the net power consumption load prediction information at a time point to obtain a plurality of sets of net load scenario prediction information.

[0009] The energy scheduling model includes: a power purchase cost and a power shortage cost, or a power purchase cost, a power shortage cost and a power selling income, or a power purchase cost, a power shortage cost and a storage cost, or a power purchase cost, a power shortage cost, a power selling income and a storage cost.

[0010] The constraint condition includes: a first condition; or, the first condition and a second condition; wherein the first condition includes: a power supply requirement condition of the energy storage device before the power failure time period arrives; the second condition includes one or more of the following conditions: during the power failure time period, the power purchase power and the power selling power are both zero; a maximum value of the charging power value and a maximum value of the discharging power value of the energy storage device; a minimum value and a maximum value of the energy storage capacity of the energy storage device; at the same time point within the target scheduling time period, the charge-discharge control information only includes a charging action or only includes a discharging action or neither includes a charging action nor includes a discharging action; a power balance constraint condition.

[0011] The power supply requirement condition includes: a sum of the storage capacity of the energy storage device before the power failure and the power generation amount of the new energy device during the power failure, being greater than or equal to a sum of the power consumption load of the power consumption device during the power failure and a first power supply amount; wherein the first power supply amount is a product of a preset first power supply coefficient and the power consumption load of the power consumption device during the power failure; or, a sum of the storage capacity of the energy storage device before the power failure and a preset second power supply amount, being greater than or equal to the net load during the power failure.

[0012] When the objective function value of the energy scheduling model is the lowest and the constraint condition is met, the charge-discharge control information of the energy storage device in the target scheduling period includes a plurality of groups of charge-discharge control information, and each group of charge-discharge control information includes charge control information or discharge control information at each time point in the target scheduling period. The method further includes: averaging the plurality of groups of charge-discharge control information according to the time points to obtain the charge control information or discharge control information at each time point in the target scheduling period as the final charge-discharge control information.

[0013] After the charge-discharge control information is sent to the energy storage device, the method further includes: collecting the energy storage state of the energy storage device, and the corresponding load demand value and power generation amount; in a case where the power prediction information includes the power generation prediction information and the power consumption load prediction information, updating the load prediction model and the power generation prediction model based on the energy storage state, the load demand value and the power generation amount; and in a case where the power prediction information includes the net power consumption load prediction information, updating the net load prediction model based on the energy storage state, the load demand value and the power generation amount.

[0014] In a second aspect, the embodiments of the present disclosure provide a user-side energy storage system, which includes: an energy storage device; a new energy power generation device, the energy storage device and a power consumption device being in communication connection with a server, and providing, to the server, power information of the energy storage device, power generation information of the new energy power generation device and power consumption information of the power consumption device respectively, the server further acquiring power grid outage information and electricity price information, so that the server executes the charge-discharge control method of the energy storage device according to any one of the first aspect.

[0015] In a third aspect, the embodiments of the present disclosure provide a server, which includes a processor and a memory, the memory storing computer executable instructions capable of being executed by the processor, and the processor executes the computer executable instructions to implement the charge-discharge control method of the energy storage device according to any one of the first aspect.

[0016] The embodiments of the present disclosure have the following beneficial effects:

[0017] The charge-discharge control method, device and user-side energy storage system described above first obtain power prediction information in a target scheduling period according to current power generation information and current power consumption information, then obtain charge-discharge control information in the target scheduling period according to current power information, electricity price information, power grid outage information and power prediction information in the target scheduling period, and then send the charge-discharge control information to the energy storage device. In the scenario of considering the power grid outage information, the net power consumption load demand during the future power outage is considered in advance based on the power prediction information, so that the scheduling plan is dynamically and flexibly made, the charge-discharge of the energy storage device is controlled, the user power consumption cost is reduced, and the user power consumption demand during the power grid outage is met.

[0018] Other features and advantages of the present disclosure will be set forth in the descriptions that follow and in part will be apparent from the description or can be learned by practice of the present disclosure. The purposes and other advantages of the present disclosure will be realized and attained by the structure particularly pointed out in the description.

[0019] In order to make the above objectives, features and advantages of the present disclosure more obvious and easy to understand, the following preferred embodiments are specifically described below, and the accompanying drawings are described in detail. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present disclosure, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0021] Fig. 1 is a flow chart of a charge and discharge control method of an energy storage device according to an embodiment of the present disclosure;

[0022] Fig. 2 is a flow chart of a management system for scheduling control of an energy storage device according to an embodiment of the present disclosure;

[0023] Fig. 3 is a flow chart of an optimization scheduling algorithm of a cloud scheduling center according to an embodiment of the present disclosure;

[0024] Fig. 4 is a structural schematic diagram of a charge and discharge control device of an energy storage device according to an embodiment of the present disclosure;

[0025] Fig. 5 is a structural schematic diagram of a user-side energy storage system according to an embodiment of the present disclosure;

[0026] Fig. 6 is a schematic diagram of a charge and discharge scheduling result of an energy storage device according to an embodiment of the present disclosure;

[0027] Fig. 7 is a schematic diagram of a server according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0028] In order to make the objectives, technical solutions and advantages of the embodiments of the present disclosure more clear, the technical solutions of the present disclosure will be described in detail below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present disclosure, but not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present disclosure.

[0029] With the popularization of new energy power generation, energy storage and other technologies in user electricity consumption, user-side energy storage systems (such as household energy storage systems, industrial and commercial energy storage systems, etc.) containing new energy power generation equipment and energy storage equipment and the energy management systems equipped with them will play an increasingly important role in reducing user electricity costs and improving user electricity comfort. However, it also faces challenges brought by various uncertainties, such as the uncertainty of new energy power generation power or generation capacity, and the uncertainty of user electricity load (i.e., the total power taken by user electricity equipment from the power system at a certain time). Moreover, with the advancement of global warming, extreme weather is becoming more frequent, which will also to some extent destroy the stability of power supply of the power grid. For example, extreme wind, freezing and other weather may cause faults of power transmission lines of the power grid. When the line fault occurs, planned power outage of the power grid is needed, which will make the user electricity face the situation of being unable to take electricity from the power grid.

[0030] On the other hand, with the advancement of industrialization and globalization, the contradiction between the increasing demand for user electricity and the insufficient local energy supply capacity is becoming increasingly prominent, so that the local has to make a corresponding power outage plan, so as to carry out planned power outage of the power grid according to the power outage plan to cope with this crisis.

[0031] Although the popularization of user-side energy storage systems can to some extent solve the problems caused by the above-mentioned planned power outage or unplanned power outage for user electricity. However, the uncertainty of taking electricity from the power grid caused by power outage, on the basis of the uncertainty of new energy power generation and the uncertainty of electricity load, adds a lot of challenges to the energy management system based on prediction. For example, power outage may cause sharp fluctuations in power supply and demand, reducing the accuracy of the prediction model and increasing the difficulty of system adjustment. How to further consider the impact of power outage information in the energy management system to maximize the reliability of user electricity, and on this basis, to reduce the electricity cost as much as possible, has become a problem that needs to be solved for the energy management system in the power outage scenario.

[0032] At present, in order to ensure the electricity demand of users when the power grid is powered off, the energy storage device is usually charged before the power outage according to the power outage information. In related technologies, the energy storage device is usually manually charged to full power before the power outage according to the power outage information, or the stored power in the energy storage device is controlled at a specified power level at a specified time. This fixed charging strategy does not consider real-time electricity demand, resulting in poor flexibility of the energy storage device controlled in this way, leading to high user electricity cost or difficulty in fully meeting the user electricity demand during power grid outage.

[0033] Based on this, the energy storage device charging and discharging control method provided by the embodiments of the present disclosure can be applied to an energy management system of a user-side energy storage system to flexibly control the charging and discharging of the energy storage device, reduce the user's electricity cost, and meet the user's electricity demand during power grid outage. It should be noted that the energy management system can be deployed in a server, which can be located in a cloud scheduling center, a local controller, etc. of the user-side energy storage system. Hereinafter, the server deployed in the cloud scheduling center is taken as an example for illustration.

[0034] To facilitate the understanding of the embodiments of the present disclosure, a specific process disclosed by the embodiments of the present disclosure is described in detail below. Referring to FIG. 1, one embodiment of the energy storage device charging and discharging control method in the embodiments of the present disclosure includes the following steps:

[0035] In step S101, the current energy information of the energy storage device, the current power generation information of the new energy power generation device, the current power consumption information of the power consumption device, and the electricity price information and power grid outage information in the target scheduling period are obtained.

[0036] The energy storage device can include energy storage batteries, super capacitors, etc. The new energy power generation device can include photovoltaic power generation devices, wind power generation devices, or tidal power generation devices, etc. without limitation.

[0037] The current energy information of the energy storage device can include static energy information such as maximum charging and discharging power and rated capacity. The current power generation information of the new energy power generation device can include the power generation amount at the current time and the power generation information in the past period (such as the past 7 days). The current power consumption information of the power consumption device can include the power consumption amount at the current time and the power consumption information in the past period (such as the past 7 days). The electricity price information can include the electricity purchase price for taking electricity from the power grid and the electricity selling price for selling electricity to the power grid in the target scheduling period. The power grid outage information can include the period when the power consumption device cannot obtain energy from the power grid, for example, at least one of planned outage information and unplanned outage information generated according to environmental data. In addition, in actual application, the power grid power supply area is usually divided according to multiple factors such as power consumption demand, power grid structure, load density, and management needs, and therefore, the power grid outage information can also include outage location information such as power supply area A location information. When the power grid outage information includes unplanned outage information generated according to environmental data, weather information can be input into a pre-trained outage prediction model for prediction to obtain outage prediction information.

[0038] In actual implementation, the current energy information of the energy storage device, the current power generation information of the new energy power generation device, the current power consumption information of the power consumption device, and the electricity price information and power grid outage information in the target scheduling period can be obtained through the energy management system.

[0039] At step S102, the power prediction information in the target scheduling period is obtained based on the current power generation information and the current power consumption information. For example, the power prediction information includes power generation prediction information and power consumption load prediction information, or the power prediction information includes net power consumption load prediction information.

[0040] That is, the power prediction information in the target scheduling period can be predicted according to the current power generation information and the current power consumption information. For example, the current power generation information and the current power consumption information can be input into a corresponding prediction model to obtain the power prediction information in the target scheduling period, or the power prediction information in the target scheduling period can be obtained by using other corresponding prediction algorithms.

[0041] In an embodiment, the power prediction information includes power generation prediction information and power consumption load prediction information. In this case, the obtained power generation information can be input into a pre-trained power generation prediction model to obtain the power generation prediction information in the target scheduling period. The obtained power consumption information can be input into a pre-trained power consumption load prediction model to obtain the power consumption load prediction information in the target scheduling period.

[0042] In another embodiment, the power prediction information includes net power consumption load prediction information. In this case, pre-set power generation information and power consumption information can be input into a pre-trained net power consumption load prediction model to obtain the net power consumption load prediction information in the target scheduling period. The net power consumption load is the difference between the power consumption load and the power generation.

[0043] At step S103, the charging and discharging control information of the energy storage device in the target scheduling period is obtained based on the current power information, the electricity price information, the power grid outage information, and the power prediction information.

[0044] It should be noted that the target scheduling period refers to a future period of time, and the charging and discharging control information of the energy storage device in the target scheduling period refers to the charging and discharging control information of the energy storage device in the future period of time, which can include charging and discharging actions and corresponding charging and discharging power values.

[0045] In actual implementation, the energy management system can obtain the charging and discharging control information of the energy storage device in the target scheduling period based on the current power information, the electricity price information, the power grid outage information, and the power prediction information.

[0046] In an embodiment, an energy scheduling model is constructed based on the current power information, the electricity price information, the power grid outage information, and the power prediction information, and the charging and discharging control information of the energy storage device in the target scheduling period is obtained by solving the energy scheduling model. The energy scheduling model can be constructed with the minimum total power consumption cost as the target and in combination with pre-set constraint conditions.

[0047] The total electricity cost can be obtained by the pre-established energy scheduling model, for example, the total electricity cost includes electricity purchase cost, electricity shortage cost, electricity sale revenue and / or energy storage cost, etc.

[0048] The constraint condition can be established according to the current electricity information and the electricity prediction information and the power grid outage information in the target scheduling period. The constraint condition can include a minimum reserved electricity quantity condition of the energy storage device; a standby power demand condition of the energy storage device before the power outage period; the electricity purchase power and the electricity sale power are both zero in the power outage period; etc.

[0049] The target scheduling period can include multiple time points (or time instants), and the charging and discharging control information of the energy storage device in the target scheduling period can include the charging and discharging power or the electricity quantity corresponding to each time point in the target scheduling period. Correspondingly, when predicting the power generation, the power generation instantaneous value corresponding to each time point can be predicted, or the average value in a short period of time can be predicted as the power generation of the corresponding time point; or the power generation quantity is predicted, and the product of the duration between adjacent two time points and the power generation can be taken as the power generation quantity of the corresponding time point.

[0050] In step S104, the charging and discharging control information is sent to the energy storage device.

[0051] The charging and discharging control information is used for the energy storage device to perform charging and discharging control, that is, the charging and discharging control information is sent to the energy storage device, so that the energy storage device performs charging and discharging based on the charging and discharging control information.

[0052] For example, the energy management system can send the charging and discharging control information to the energy storage device, so that the energy storage device performs charging and discharging according to the charging and discharging control information.

[0053] The charging and discharging control method of the energy storage device is to obtain the electricity prediction information in the target scheduling period according to the current power generation information and the current electricity consumption information, and then obtain the charging and discharging control information in the target scheduling period according to the current electricity information, the electricity price information in the target scheduling period, the power grid outage information and the electricity prediction information, and then send it to the energy storage device. In the scenario considering the power grid outage information, the net electricity load demand during the future power outage is considered in advance based on the electricity prediction information, so that the scheduling plan is dynamically and flexibly made, the charging and discharging of the energy storage device is controlled, the user electricity cost is reduced, and the user electricity demand during the power grid outage is met.

[0054] The power grid outage information includes: power outage plan information, and / or power outage prediction information generated based on environmental data.

[0055] The power grid outage information is outage information in a target scheduling period. Specifically, the power grid outage information can be outage plan information; can be outage prediction information generated according to environmental data; or can be outage information obtained by weighting the outage plan information and the outage prediction information generated according to the environmental data, for example, calculating the product of the outage probability of time a and time a in the outage plan information, calculating the product of the outage probability of time a and time a in the outage prediction information generated based on the environmental data, and then adding the two products to obtain the outage information of time a after weighting.

[0056] The outage plan information refers to planned outage information of the power grid to cope with insufficient power supply, and the outage prediction information refers to non-planned outage information that may cause power outage due to extreme weather and the like. When the power grid outage information is outage plan information, the outage plan information can be obtained in advance through the network, so as to obtain a time period in which power outage occurs in the future. When the power grid outage information includes outage prediction information generated based on environmental data, the outage prediction information can be obtained based on weather information, for example, the frequency of power outage of the power grid when historical extreme weather occurs can be counted, and an outage prediction model can be trained, and then the predicted extreme weather in the target scheduling period is input into the pre-trained outage prediction model for prediction, so as to obtain the outage prediction information such as the outage time period and the outage probability; for another example, when the amount of data used for model training is small, a threshold can be set to determine, for example, when the frequency of occurrence of historical extreme weather and the corresponding power grid outage event is small, some rules can be artificially set, for example, it is assumed that power outage may occur when the predicted wind speed exceeds a certain threshold or the predicted rainfall or snowfall exceeds a certain threshold, so as to determine the time period in which power outage is possible according to the corresponding predicted time.

[0057] In a specific implementation, when the power prediction information includes the power generation prediction information and the power consumption load prediction information, the power prediction information in the target scheduling period is obtained based on the current power generation information and the current power consumption information, including: inputting the current power generation information into a preset power generation prediction model to obtain the power generation prediction information in the target scheduling period; and inputting the current power consumption information into a preset load prediction model to obtain the power consumption load prediction information in the target scheduling period.

[0058] In an example, the power generation prediction model can be trained based on historical power generation data in advance to obtain a trained power generation prediction model, and the power consumption load prediction model can be trained based on historical power consumption load to obtain a trained power consumption load prediction model.

[0059] In actual application, the current power generation information is input into the trained power generation prediction model for prediction to obtain power generation prediction information in the target scheduling period, and the current power consumption information is input into the trained power consumption load prediction model for prediction to obtain power consumption load prediction information in the target scheduling period. At this time, the power generation prediction information and the power consumption load prediction information in the target scheduling period are the power prediction information in the target scheduling period.

[0060] In another specific implementation, when the power prediction information includes net power consumption load prediction information, the power prediction information in the target scheduling period is obtained based on the current power generation information and the current power consumption information, including: inputting the current power generation information and the current power consumption information into a preset net load prediction model to obtain net power consumption load prediction information in the target scheduling period.

[0061] In an example, the net load prediction model can be trained in advance based on the difference between historical power generation data and historical power consumption load to obtain a trained net load prediction model. In actual application, the current power generation information and the current power consumption information are input into the trained net power consumption load prediction model to obtain net power consumption load prediction information in the target scheduling period. At this time, the net power consumption load prediction information in the target scheduling period is the power prediction information in the target scheduling period.

[0062] In a specific implementation, the target scheduling period includes multiple time points, and the power prediction information includes power prediction information corresponding to each time point in the multiple time points in the target scheduling period. The charge and discharge control information of the energy storage device in the target scheduling period is obtained based on the current power information, the electricity price information, the power grid outage information, and the power prediction information, including:

[0063] First, based on the power prediction information and a plurality of preset groups of prediction error information, a plurality of groups of scenario prediction information are generated. Each group of prediction error information includes prediction error information corresponding to each time point in the multiple time points in the target scheduling period, and each group of scenario prediction information includes scenario prediction information corresponding to each time point in the multiple time points in the target scheduling period. The scenario prediction information corresponds to the power prediction information. Each group of scenario prediction information corresponds to a scenario probability.

[0064] It should be noted that the above scenario is a quantitative and qualitative description of all possible future development trends. The above generation of a plurality of groups of scenario prediction information based on the power prediction information and a plurality of preset groups of prediction error information is a scenario process, which refers to sampling the probability distribution model of a random variable to obtain a large number of scenario sets through sampling. The random variable corresponds to the prediction error information of the present embodiment.

[0065] When the power prediction information comprises the power generation prediction information, the prediction error information can be obtained by sampling the error distribution output when the pre-trained power generation prediction model is verified or tested; when the power prediction information comprises the power consumption load prediction information, the prediction error information can be obtained by sampling the error distribution output when the pre-trained load prediction model is verified or tested; when the power prediction information comprises the net power consumption load prediction information, the prediction error information can be obtained by sampling the error distribution output when the pre-trained net load prediction model is verified or tested.

[0066] Correspondingly, when the power prediction information comprises the power generation prediction information, the power generation prediction information and the corresponding multiple sets of prediction error information are respectively superimposed according to time points to obtain multiple sets of power generation scenario prediction information, for example, for each set of prediction error information, the power generation prediction information at time points t1, t2, etc. and the prediction error information at the corresponding time points are superimposed to obtain each set of power generation scenario prediction information; when the power prediction information comprises the power consumption load prediction information, the power consumption load prediction information and the corresponding multiple sets of prediction error information are respectively superimposed according to time points to obtain multiple sets of load scenario prediction information, for example, for each set of prediction error information, the power consumption load prediction information at time points t1, t2, etc. and the prediction error information at the corresponding time points are superimposed to obtain each set of load scenario prediction information; when the power prediction information comprises the net power consumption load prediction information, the net power consumption load prediction information and the corresponding multiple sets of prediction error information are respectively superimposed according to time points to obtain multiple sets of net load scenario prediction information, for example, for each set of prediction error information, the net power consumption load prediction information at time points t1, t2, etc. and the prediction error information at the corresponding time points are superimposed to obtain each set of net load scenario prediction information.

[0067] After generating the multiple sets of scenario prediction information, first, the initial scenario probability corresponding to each set of scenario prediction information is set, the information distance between each two sets of scenario prediction information is calculated to form a distance matrix. Then, according to the distance matrix, the sum of the probability distance of each set of scenario prediction information and each set of scenario prediction information except for the scenario prediction information is calculated to obtain a probability distance matrix. Next, according to the distance matrix and the probability distance matrix, at least part of the multiple sets of scenario prediction information is merged to obtain new scenario prediction information and new scenario probability.

[0068] Optionally, the target scene information corresponding to the minimum value in the probability distance matrix is obtained, and the minimum value of the information distance of the target scene information in the distance matrix is obtained. Then, the scene information corresponding to the minimum value of the information distance is determined from the multiple sets of scene prediction information, and the target scene information is merged to obtain a new set of scene prediction information; at the same time, the scene probability corresponding to the minimum value of the information distance is merged with the scene probability corresponding to the target scene information to obtain the scene probability corresponding to the new set of scene prediction information.

[0069] In addition, according to the new set of scene prediction information, the multiple sets of scene prediction information are updated, and the merged scene prediction information is deleted from the scene prediction information to obtain the multiple sets of scene prediction information after updating. In actual implementation, the above steps can be repeatedly executed until the multiple sets of scene prediction information after updating reach a pre-set group number threshold.

[0070] Then, the multiple sets of scene prediction information, the scene probability corresponding to each set of scene prediction information, and the electricity price information are input into a preset energy scheduling model, and the charge and discharge control information of the energy storage device in the target scheduling period is obtained when the objective function value of the energy scheduling model is the lowest and the constraint condition is satisfied; wherein the constraint condition is established based on the electricity prediction information, the current electricity information, and the power grid outage information.

[0071] Specifically, the scene prediction information, the scene probability, and the electricity price information can be input into the energy scheduling model for calculation to obtain a model that meets the demand. By setting corresponding constraint formulas for the electricity prediction information, the current electricity information, and the power grid outage information, the constraint condition is established.

[0072] Finally, the charge and discharge control information of the energy storage device in the target scheduling period is obtained when the objective function value of the energy scheduling model is the lowest and the constraint condition is satisfied; wherein the charge and discharge control information includes: a charge and discharge action and a charge and discharge power value corresponding to the charge and discharge action.

[0073] In a specific implementation, the electricity prediction information includes power generation power prediction information and electricity load prediction information; the multiple sets of prediction error information corresponding to the power generation power prediction information are obtained by sampling and processing the first error distribution output when the power generation power prediction model is model verified and / or model tested; and the multiple sets of prediction error information corresponding to the electricity load prediction information are obtained by sampling and processing the second error distribution output when the electricity load prediction model is model verified and / or model tested.

[0074] In one example, a first error distribution output by the power generation prediction model during model validation, or model testing, or both, is obtained, and the first error distribution is sampled to obtain a plurality of sets of prediction error information corresponding to the power generation prediction information. For example, the first error distribution is sampled to generate a plurality of sets of prediction error information corresponding to each time point in the target scheduling period.

[0075] In another example, a second error distribution output by the power consumption prediction model during model validation, or model testing, or both, is obtained, and the second error distribution is sampled to obtain a plurality of sets of prediction error information corresponding to the power consumption prediction information. For example, the second error distribution is sampled to generate a plurality of sets of prediction error information corresponding to each time point in the target scheduling period.

[0076] In another specific implementation, the power prediction information includes net power consumption prediction information, and the plurality of sets of prediction error information corresponding to the net power consumption prediction information is obtained by sampling a third error distribution output by the net load prediction model during model validation and / or model testing.

[0077] In one example, a third error distribution output by the net load prediction model during model validation, or model testing, or both, is obtained, and the third error distribution is sampled to obtain a plurality of sets of prediction error information corresponding to the net power consumption prediction information. For example, the third error distribution is sampled to generate a plurality of sets of prediction error information corresponding to each time point in the target scheduling period.

[0078] The power prediction information includes power generation prediction information and power consumption prediction information, and the scenario prediction information includes power generation scenario prediction information and load scenario prediction information. In one specific implementation, based on the power prediction information and a plurality of sets of preset prediction error information, a plurality of sets of scenario prediction information are generated, including: for the plurality of sets of prediction error information corresponding to the power generation prediction information, each set of prediction error information is superimposed on the power generation prediction information according to time points to obtain a plurality of sets of power generation scenario prediction information; and for the plurality of sets of prediction error information corresponding to the power consumption prediction information, each set of prediction error information is superimposed on the power consumption prediction information according to time points to obtain a plurality of sets of load scenario prediction information.

[0079] In one example, for the power generation prediction information, the prediction error information and the power generation prediction information are grouped according to a calculation unit, the prediction error information under the same calculation unit is added to the corresponding power generation prediction information to obtain a plurality of sets of power generation scenario prediction information.

[0080] In another example, for the electricity load prediction information, the prediction error information and the electricity load prediction information are grouped according to a calculation unit, the prediction error information under the same calculation unit is added to the corresponding electricity load prediction information, and a plurality of groups of load scene prediction information are obtained.

[0081] For example, in one embodiment, historical power generation data is collected and cleaned first, and a training set, a validation set and a test set are distinguished, and a prediction model is trained based on machine learning or deep learning. Specifically, the model can be trained based on historical power generation to obtain a power generation prediction model, and the error distribution of the power generation prediction model at each time in the validation set and the test set is counted. In actual application, the power generation prediction model is used to predict the power generation in the target scheduling period to obtain power generation prediction information at a plurality of time points in the target scheduling period, and then the error distribution is sampled to generate a plurality of groups of prediction error information of the power generation at each time point in the target scheduling period. Then, the prediction error information of the power generation can be added to the corresponding power generation prediction information according to a time unit, such as an hour, to obtain a plurality of groups of power generation scene prediction information in the target scheduling period. Similarly, the process of obtaining load scene prediction information can refer to the embodiment of power generation scene prediction information.

[0082] Alternatively, the above-mentioned power prediction data includes net load prediction data, and the scene prediction information includes net load scene prediction information. In one specific implementation, based on the power prediction information and a plurality of groups of preset prediction error information, a plurality of groups of scene prediction information are generated, including: for a plurality of groups of prediction error information corresponding to the net electricity load prediction information, each group of prediction error information is respectively superimposed on the net electricity load prediction information according to a time point, and a plurality of groups of net load scene prediction information are obtained.

[0083] In one example, for the net electricity load prediction information, the prediction error information and the net electricity load prediction information are grouped according to a calculation unit, the prediction error information under the same calculation unit is added to the corresponding net electricity load prediction information, and a plurality of groups of net load scene prediction information are obtained. Similarly, the process of obtaining net load scene prediction information can refer to the embodiment of power generation scene prediction information.

[0084] In one implementation, after generating the plurality of sets of scenario prediction information based on the power prediction information and the plurality of sets of preset prediction error information, the method further includes: setting an initial scenario probability for each set of scenario prediction information; obtaining an information distance between each two sets of scenario prediction information in the plurality of sets of scenario prediction information to obtain a distance matrix; taking each set of scenario prediction information as current prediction information one by one, determining a probability distance sum of the current prediction information and scenario prediction information other than the current prediction information based on the initial scenario probability of each set of scenario prediction information and the distance matrix to obtain a probability distance matrix; merging at least part of the sets of scenario prediction information in the plurality of sets of scenario prediction information based on the distance matrix and the probability distance matrix to obtain new scenario prediction information and a scenario probability corresponding to the new scenario prediction information; and updating the plurality of sets of scenario prediction information based on the new scenario prediction information.

[0085] To facilitate understanding, first, the theory involved in the embodiment is described. In the embodiment, scenario reduction is performed on a large number of scenario sets. The goal of the scenario reduction is to make the reduced scenario set close to the original scenario set from the perspective of probability measure, that is, to reduce the number of scenarios on the premise of retaining more information, thereby reducing the resource consumption of optimization solving. Common reduction methods include Fast Forward Selection and Simultaneous Backward Reduction. After scenario reduction, a number of representative scenarios are obtained, and each scenario has a corresponding probability value, and the sum of the probability values is 1.

[0086] Specifically, taking n sets of scenario prediction information as an example.

[0087] First, for each set of scenario prediction information, an initial scenario probability is set, for example, the initial scenario probabilities are equal, and each is 1 / n.

[0088] Then, for the n sets of scenario prediction information, an information distance between each two sets of scenario prediction information is calculated to obtain a distance matrix, where the distance can be understood as Manhattan distance, Euclidean distance, etc. And taking each set of scenario prediction information as current prediction information one by one, the probability distance sum of the current prediction information and the remaining scenario prediction information other than the current prediction information is calculated according to the initial scenario probability and the distance matrix to obtain a probability distance matrix according to the probability matrix formula.

[0089] Next, according to the calculated distance matrix and probability distance matrix, part or all of the sets of scenario prediction information in the n sets of scenario prediction information are merged to obtain new scenario prediction information and a scenario probability corresponding to the new scenario prediction information.

[0090] In one example, an average value of the scene prediction numbers of part or all of the n groups of scene prediction information is calculated, or a weighted average value is calculated according to the scene probabilities corresponding to the scene prediction information, so as to combine the scene prediction numbers of part or all of the n groups of scene prediction information.

[0091] Finally, the n groups of scene prediction information are updated according to the new scene prediction information.

[0092] In one specific implementation, a minimum value is obtained from the probability distance matrix, and a first scene information corresponding to the minimum value is determined; an information distance minimum value corresponding to the first scene information is obtained from the distance matrix, and a second scene information corresponding to the information distance minimum value is determined; the first scene information and the second scene information are combined to obtain new scene prediction information; and scene probabilities of the first scene information and the second scene information are combined to obtain a scene probability corresponding to the new scene prediction information.

[0093] Specifically, a first scene information i corresponding to a minimum value in the probability distance matrix is determined. Then, in the plurality of groups of scene prediction information, a second scene information j having a minimum information distance from the first scene information i is determined according to the distance matrix. Next, the second scene information j is combined into the first scene information i, such as calculating an average value of the first scene information i and the second scene information j, or performing a weighted calculation according to the probabilities of the first scene information i and the second scene information j to obtain new scene prediction information. At the same time, a scene probability p i of the first scene information i is added to a scene probability p j of the second scene information j to obtain a scene probability corresponding to the new scene prediction information.

[0094] Further, the new scene prediction information is updated to the plurality of groups of scene prediction information, and the combined scene prediction information is deleted from the plurality of groups of scene prediction information to obtain updated plurality of groups of scene prediction information; based on the updated plurality of groups of scene prediction information, the distance matrix and the probability distance matrix are updated, and the steps of combining at least part of the plurality of groups of scene prediction information based on the distance matrix and the probability distance matrix to obtain new scene prediction information and a scene probability corresponding to the new scene prediction information are continued to be executed until the number of information groups of the updated plurality of groups of scene prediction information meets a preset group number threshold.

[0095] That is, the new scene prediction information is updated to the n groups of scene prediction information, and the combined scene prediction information is deleted from the plurality of groups of scene prediction information, for example, the scene prediction information j is combined into the scene prediction information i, and then the scene prediction information i is deleted from the n groups of scene prediction information to obtain updated n groups of scene prediction information.

[0096] Specifically, the distance matrix and the probability distance matrix are updated according to the updated n groups of scene prediction information. Optionally, the step of merging at least part of the groups of scene prediction information in the plurality of groups of scene prediction information based on the distance matrix and the probability distance matrix to obtain new scene prediction information and a scene probability corresponding to the new scene prediction information is repeatedly performed until the number of information groups in the updated n groups of scene prediction information meets a preset group number threshold, and the step is stopped.

[0097] For example, the embodiment repeatedly performs the foregoing steps to continuously reduce the scenes until m scenes and probabilities p1, p2,..., pm of each scene are obtained. m , and

[0098] The energy scheduling model includes a power purchase cost and a power shortage cost.

[0099] The energy scheduling model further includes a power selling revenue and / or an energy storage cost.

[0100] That is, the energy scheduling model can specifically include the power purchase cost and the power shortage cost, or the power purchase cost, the power shortage cost, and the power selling revenue, or the power purchase cost, the power shortage cost, and the energy storage cost, or the power purchase cost, the power shortage cost, the power selling revenue, and the energy storage cost.

[0101] In one embodiment, when the energy scheduling model includes the power purchase cost, the power shortage cost, the power selling revenue, and the energy storage cost, the calculation formula of the energy scheduling model is as follows: obj all = obj buy - obj sell + obj bat + obj unmet

[0102] wherein, obj all is a target function value of the energy scheduling model, that is, a target function value of the total power consumption cost; obj buy is the power purchase cost, which can be determined based on scene prediction data, power purchase price information, and power purchase power; obj sell is the power selling revenue, which can be determined based on scene prediction data, power selling price information, and power selling power; obj bat is the energy storage cost, which can be determined based on scene prediction data, energy storage equivalent cost, and charge-discharge control information; and obj unmet is a cost when a constraint condition cannot be met, that is, a power shortage cost, which can be determined based on scene prediction data, an equivalent cost when power cannot be balanced, and a power shortage.

[0103] The calculation formula of each cost or revenue is as follows:

[0104] Wherein, obj is the cost or benefit; p s represents the scene probability of the s-th group of scene prediction information; c t，s represents various equivalent costs at the t-th time point in the s-th group of scene prediction information, for example, the electricity purchase price, the electricity selling unit price, the energy storage equivalent cost, the equivalent cost when the power cannot be balanced, etc.; u t，s represents various powers at the t-th time point in the s-th group of scene prediction information, for example, the electricity purchase power, the electricity selling power, the charge-discharge control information, the insufficient power, etc.; m is the total number of groups of scene prediction information; T is the number of multiple time points within the target scheduling period.

[0105] The above electricity purchase cost includes: the sub-electricity purchase cost corresponding to each group of scene prediction information; the sub-electricity purchase cost includes: the sum of the product of the electricity purchase price and the electricity purchase power at each time point within the target scheduling period.

[0106] The above electricity purchase cost further includes: the sum of the product of the scene probability corresponding to the scene prediction information and the sub-electricity purchase cost corresponding to the scene prediction information.

[0107] In the above manner, the sub-electricity purchase cost in the energy scheduling model is calculated, and then the electricity purchase cost is obtained.

[0108] The above electricity selling benefit includes: the sub-electricity selling benefit corresponding to each group of scene prediction information; the sub-electricity selling benefit includes: the sum of the product of the electricity selling unit price and the electricity selling power at each time point within the target scheduling period.

[0109] The above electricity selling benefit further includes: the sum of the product of the scene probability corresponding to the scene prediction information and the sub-electricity selling benefit corresponding to the scene prediction information.

[0110] In the above manner, the sub-electricity selling benefit in the energy scheduling model is calculated, and then the electricity selling benefit is obtained.

[0111] The above energy storage cost includes: the sub-energy storage cost corresponding to each group of scene prediction information; the sub-energy storage cost includes: the sum of the product of the energy storage equivalent cost and the charge-discharge control parameter at each time point within the target scheduling period, wherein the charge-discharge control parameter can usually be the discharge power value.

[0112] The above energy storage cost further includes: the sum of the product of the scene probability corresponding to the scene prediction information and the sub-energy storage cost corresponding to the scene prediction information.

[0113] In the above manner, the sub-energy storage cost in the energy scheduling model is calculated, and then the energy storage cost is obtained.

[0114] The above power shortage cost includes: the sub-power shortage cost corresponding to each group of scene prediction information; the sub-power shortage cost includes: the sum of the product of the equivalent cost when the power cannot be balanced and the insufficient power that cannot be met at each time point within the target scheduling period.

[0115] The power shortage cost further includes a sum of products of scenario probabilities corresponding to the scenario prediction information and sub-power shortage costs corresponding to the scenario prediction information.

[0116] In this way, the sub-power shortage cost in the energy scheduling model is calculated, and then the power shortage cost is obtained.

[0117] The constraint condition includes the first condition or the first condition and the second condition, and the first condition includes a power supply requirement condition of the energy storage device before the power failure time period arrives.

[0118] The second condition includes one or more of the following conditions:

[0119] First, the power purchase power and the power selling power are both zero in the power failure time period.

[0120] That is, the second condition can further include the power supply requirement condition of the energy storage device before the power failure time period arrives.

[0121] The second condition includes a basic condition for safe and stable operation of the entire system, and specifically can include the power supply requirement condition of the energy storage device before the power failure time period arrives, the power purchase power and the power selling power are both zero in the power failure time period, and the like.

[0122] Second, the maximum value of the charging power value and the maximum value of the discharging power value of the energy storage device. That is, the condition is: That is, the charging and discharging control information at the t-th moment in the s-th set of scenario prediction information cannot exceed the two maximum values, where is charging control information in the charging and discharging control information at the t-th moment in the s-th set of scenario prediction information, is discharging control information in the charging and discharging control information at the t-th moment in the s-th set of scenario prediction information, is the maximum value of the charging power value, indicates the maximum value of the discharging power value.

[0123] Third, the minimum value and the maximum value of the energy storage capacity of the energy storage device;

[0124] Fourth, the charging and discharging control information at the same moment within the target scheduling time period includes only charging actions or only discharging actions or neither charging actions nor discharging actions;

[0125] Fifth, a power balance constraint condition;

[0126] In one example, the calculation formula of the power balance constraint condition is as follows:

[0127] wherein, a charging control parameter in the charging and discharging control information corresponding to the t-th time in the s-th group of scenario prediction information; denotes the selling power at the t-th time in the s-th group of scenario prediction information; denotes the load power at the t-th time in the s-th group of scenario prediction information, denotes the generation power at the t-th time in the s-th group of scenario prediction information, a discharging control parameter in the charging and discharging control information corresponding to the t-th time in the s-th group of scenario prediction information; denotes the shortage power at the t-th time in the s-th group of scenario prediction information, which cannot be balanced in real time due to the occurrence of power failure.

[0128] It should be noted that the above-mentioned power supply demand condition can be converted into the form of net power load . .

[0129] The above-mentioned power supply demand condition includes: the sum of the storage capacity of the energy storage device before the power failure and the power generation amount of the new energy device during the power failure is greater than or equal to the sum of the power load of the power consumption device during the power failure and the first power supply amount; wherein the first power supply amount is the product of the first power supply coefficient and the power load of the power consumption device during the power failure. In one embodiment, the calculation formula of the power supply demand condition is as follows:

[0130] wherein, denotes the time before the occurrence of power failure corresponding to the s-th group of scenario prediction information, denotes the energy storage state of the energy storage device at the time before the occurrence of power failure corresponding to the s-th group of scenario prediction information; denotes the power that can be provided by the new energy power generation device during the power failure; denotes the start time of the power failure; denotes the end time of the power failure; denotes the generation power corresponding to the t-th time in the s-th group of scenario prediction information, cap bat denotes the capacity of the energy storage device, unit: kWh; ts denotes the resolution of time, unit: min; the first power supply coefficient a≥0, which represents the energy storage capacity reserved to meet the stable and reliable power demand; denotes the load power corresponding to the t-th time in the s-th group of scenario prediction information.

[0131] The physical meaning of the above-mentioned formula can be understood as follows: under each group of scenarios, the energy storage power before the power failure and the new energy power generation amount during the power failure can meet the load amount during the power failure, and a certain amount of excess is left.

[0132] In actual applications, when a power outage occurs, the energy storage device cannot obtain power from the power grid, and the output power of new energy power generation equipment, such as photovoltaic energy, is affected by the weather and has uncertainty, so it is necessary to charge the energy storage device in advance before the arrival of the power outage period to ensure the reserved specified energy storage capacity, so that the energy storage device can provide power for the power outage period. Through the above-mentioned manner, a first mode is provided to set a constraint condition for the power supply demand condition of the energy storage device before the arrival of the power outage period.

[0133] The value of the first power supply coefficient a can be determined based on prediction information of the power prediction data, or the position of the power outage time from the current time, or the prediction information of the power prediction data and the position of the power outage time from the current time.

[0134] In another mode, the sum of the storage capacity of the energy storage device before the power outage and the second power supply amount β is greater than or equal to the net load during the power outage.

[0135] wherein, represents the time before the occurrence of the power outage corresponding to the s-th set of scenario prediction information, is the energy storage state of the energy storage device at the time before the occurrence of the power outage corresponding to the s-th set of scenario prediction information; the second power supply amount β≥0 represents the energy storage capacity reserved to meet the stable and reliable power demand; represents the start time of the power outage; represents the end time of the power outage; represents the net load corresponding to the t-th time in the s-th set of scenario prediction information, cap bat represents the capacity of the energy storage, and the unit is kWh; ts represents the resolution of time, and the unit is min.

[0136] The physical meaning of the above formula can be understood as that the energy storage capacity before the power outage plus the power supply amount should be greater than or equal to the net load demand under each set of scenarios.

[0137] Through the above-mentioned manner, a second mode is provided to set a constraint condition for the power supply demand condition of the energy storage device before the arrival of the power outage period.

[0138] The value of the above-mentioned β is determined based on the prediction error of the power prediction data, or the position of the power outage time from the current time, or the prediction error of the power prediction data and the position of the power outage time from the current time.

[0139] It should be noted that the value of a or β can be pre-set as a constant value, such as 10%. However, this mode is not flexible enough and may not fully utilize the potential of energy storage.

[0140] In addition, the values of a or β can be adaptively adjusted according to the prediction error statistics of load prediction and new energy power generation prediction, such as photovoltaic prediction. For example, the distribution of prediction errors of a certain time period by the photovoltaic prediction and load prediction system in the previous 14 days is statistically analyzed. If the deviation between the historical prediction and the actual value of the time period is within 25%, it is considered that the prediction robustness is good, and the value of a or β can be appropriately reduced. If the overall gap is greater than 25%, it indicates that the prediction stability of the time period is low. In order to meet the requirement of reliable power utilization during power outage, the conventional value of a or β needs to be kept unchanged or appropriately increased.

[0141] In an implementation manner, when the target function value of the energy scheduling model is the lowest and the constraint condition is met, the charging and discharging control information of the energy storage device in the target scheduling period includes a plurality of groups of charging and discharging control information, each group of charging and discharging control information includes charging control information or discharging control information at each time point in the target scheduling period, and the method further includes: averaging the plurality of groups of charging and discharging control information according to the time points to obtain the charging control information or discharging control information at each time point in the target scheduling period as the final charging and discharging control information.

[0142] After obtaining a plurality of groups of scenario prediction information through the foregoing manner, the plurality of groups of scenario prediction information, the scenario probability corresponding to each group of scenario prediction information, and the price information at each time point in the target scheduling period can be input into the energy scheduling model, and based on the foregoing constraint condition, a plurality of groups of charging and discharging control information can be calculated, which makes the target function value of the energy scheduling model the lowest, wherein each group of charging and discharging control information includes charging control information or discharging control information at each time point in the target scheduling period. Then, for each time point in the target scheduling period, the charging control information and the discharging control information in the plurality of groups of charging and discharging control information are averaged to obtain the final charging control information or discharging control information at each time point. For example, assuming that there are m groups of scenarios in a certain time length, the charging control parameter discharging control parameter may have m different values, and the average value of the m values is calculated to obtain the final charging and discharging control information.

[0143] In an implementation manner, after the charging and discharging control information is sent to the energy storage device, the method further includes: collecting the energy storage state of the energy storage device, and the corresponding load demand value and power generation amount; in the case that the electric energy prediction information includes power generation prediction information and power consumption load prediction information, updating the load prediction model and the power generation prediction model based on the energy storage state, the load demand value and the power generation amount; in the case that the electric energy prediction information includes net power consumption load prediction information, updating the net load prediction model based on the energy storage state, the load demand value and the power generation amount.​

[0144] Specifically, after the charging and discharging control of the energy storage device according to the charging and discharging control information, the energy management system can collect the energy storage state of the energy storage device, and information such as the load demand value and the power generation capacity of the target scheduling period corresponding to the energy storage state, and then update the prediction model based on the collected information, thereby improving the accuracy of the prediction of the prediction model.

[0145] In one embodiment, referring to the flow chart of the energy management system for scheduling control of the energy storage device shown in FIG. 2, the following will be described in detail.

[0146] First, the cloud scheduling center (including the energy management system) obtains input information, which includes current power information, power price information in the target scheduling period, power grid outage information and power prediction information, then calculates output information, i.e. charging and discharging control information, based on the input information, and inputs the charging and discharging control information to the local controller (which can be a local energy management system), which controls the charging and discharging of the energy storage device based on the charging and discharging control information; at the same time, the cloud scheduling center updates the power prediction information.

[0147] Specifically, the data acquisition module in the energy management system of the cloud scheduling center acquires the information of the energy storage device, such as the current SOC value of the energy storage, the maximum charging and discharging power of the energy storage device, the rated capacity of the energy storage, etc. The data prediction module acquires the power generation information, load demand information and power grid outage information in the target scheduling period. Other information, such as real-time power price information, can also be obtained from the data acquisition module or the data prediction module.

[0148] Based on the above obtained information, according to specific targets such as minimum grid power cost and stability of energy storage power supply during power outage, a corresponding energy scheduling model is constructed, wherein the constraint conditions include power balance, energy storage charging and discharging power limit, minimum reserved power, etc., and the model is solved to obtain the charging and discharging control parameters of the energy storage device in the future period of time, and the charging and discharging control parameters are downloaded to the local energy management system for execution.

[0149] After waiting for execution, the data acquisition module of the local controller collects the information required by the cloud scheduling center at the next time, such as the SOC value of the energy storage after executing the control instruction of the cloud scheduling center, and uploads it to the cloud scheduling center. At the next time, the cloud scheduling center further updates the corresponding power generation power prediction model and load prediction model according to the latest acquired energy storage device information and the power generation value and load demand value at the last time, so as to obtain the power generation power information and load demand information at multiple time points in the target scheduling period, and repeatedly cycle the prediction, optimization and control process.

[0150] It should be noted that the target of the embodiment scheduling is to reduce the power grid power as much as possible under the premise of meeting the power supply during power failure. The required power generation prediction information in the scheduling under the power failure scenario is the expected power generation prediction, that is, the influence of light abandonment due to power failure or zero power grid is excluded when training the power generation prediction model.

[0151] In addition, the time resolution of the energy management system of the cloud scheduling center and the local controller is inconsistent. For example, the time resolution of the cloud may be 5 minutes, 15 minutes, 30 minutes, etc., and the time resolution of the local may reach seconds, milliseconds, etc. Therefore, when the local receives the instruction from the cloud scheduling center, the instruction needs to be corrected and executed in combination with the real-time data of the local to ensure the safe and stable operation of the system.

[0152] In one embodiment, referring to the cloud scheduling center optimization scheduling algorithm flowchart shown in FIG. 3, the following will be described in detail.

[0153] After the cloud scheduling center optimization scheduling algorithm starts,

[0154] In step S301, the current power information, the current power generation information of the new energy power generation equipment, the current power consumption information of the power consumption equipment, and the price information and power grid power failure information of the target scheduling period are obtained. The power prediction information of the target scheduling period is determined according to the power generation information and the power consumption information;

[0155] In step S302, the power prediction information and the preset multiple sets of prediction error information are used for scene generation to generate a large number of scene prediction information;

[0156] In step S303, the scene reduction is performed on the large number of scene sets, and finally multiple sets of scene prediction information are obtained;

[0157] In step S304, the mathematical model of the optimization target scheduling period is established, and the random optimization is solved;

[0158] In step S305, the charge and discharge control information of the energy storage equipment in the target scheduling period is obtained;

[0159] In step S306, the charge and discharge control information in the target scheduling period is applied to the system;

[0160] In step S307, the energy storage state and other information of the energy storage equipment are updated;

[0161] In step S308, the prediction model is updated and optimized, and the time in the target scheduling period is updated to the next time;

[0162] In step S309, it is determined whether the optimization scheduling algorithm needs to be continued. If yes, return to step S301; if no, execute step S310;

[0163] Step S310: The cloud-based scheduling center completes the optimization of the scheduling algorithm.

[0164] During the process of updating the energy storage device's electrical information, the energy storage device's electrical information after executing cloud commands, along with the corresponding load demand and power generation for that time period, is fed back to the cloud dispatch center. This updates the prediction model online, optimizing and predicting the target scheduling time for the next time period, repeating the prediction and optimization control process. After executing specific commands locally, the energy storage status and other information of the energy storage device for that time period will be updated, and this information needs to be promptly updated and fed back to the cloud dispatch center.

[0165] Simultaneously, the load demand and power generation during this period can be uploaded to the cloud dispatch center in a timely manner, thereby utilizing the concept of data assimilation to update the load forecasting model and power generation forecasting model of the cloud dispatch center online. Assume the prior prediction at time k is... The measured value is y k Given that the error statistics for time k based on historical prediction data are σ1 and the observation error of the measured value are σ2, the posterior estimate at time k, i.e., the optimal estimate, is... Then The input to the prediction model ultimately allows the cloud scheduling center algorithm to update the optimization and prediction of the target scheduling time to the next time, repeating the above prediction, optimization, and control process.

[0166] This embodiment processes the data in a cloud-based dispatch center, which has the following advantages:

[0167] Elastic computing resources: Cloud-based scheduling centers typically have larger computing resources, enabling them to easily handle large-scale data and complex computing tasks. This allows intelligent scheduling algorithms running in cloud-based scheduling centers to more flexibly respond to tasks of different scales and needs.

[0168] Centralized Management: The cloud-based scheduling center offers the advantage of centralized management, allowing administrators to more easily monitor and manage the algorithm's operational status, and perform maintenance and updates. This reduces management costs and improves system maintainability.

[0169] Global optimization: Intelligent scheduling algorithms running in a cloud-based scheduling center have access to more global information, such as the overall system status and resource utilization. This allows the algorithms to perform more comprehensive optimizations to achieve optimal overall system performance.

[0170] Suitable for large-scale deployments: If an application needs to be deployed across multiple geographic locations or on a large number of devices, using a cloud-based dispatch center makes it easier to achieve unified management and coordination, avoiding the complexity of deploying and maintaining algorithms on each edge device.

[0171] Easy to expand: Cloud scheduling center usually has good scalability, can flexibly expand computing resources according to demand, to meet the growth needs of the system. This makes the intelligent scheduling algorithm running in the cloud scheduling center able to adapt to changing needs and scale.

[0172] For the above method embodiment, refer to the structure diagram of the charge and discharge control device of the energy storage device shown in FIG. 4. The device comprises:

[0173] The information acquisition module 401 is configured to acquire current energy information of the energy storage device, current power generation information of the new energy power generation device, current power consumption information of the power consumption device, and price information and power grid outage information in a target scheduling period.

[0174] The prediction module 402 is configured to obtain energy prediction information in the target scheduling period based on the current power generation information and the current power consumption information.

[0175] The control information generation module 403 is configured to obtain charge and discharge control information of the energy storage device in the target scheduling period based on the current energy information, the price information, the power grid outage information, and the energy prediction information.

[0176] The information sending module 404 is configured to send the charge and discharge control information to the energy storage device, so that the energy storage device performs charge and discharge based on the charge and discharge control information.

[0177] In the scenario considering the power grid outage information, the method considers the net power load demand during future power outage based on the energy prediction information in advance, so as to dynamically and flexibly make a scheduling plan, control the charge and discharge of the energy storage device, reduce the user power consumption cost, and meet the user power consumption demand during power grid outage.

[0178] The energy prediction information includes power generation prediction information and power consumption load prediction information; the prediction module is further configured to input the current power generation information into a preset power generation prediction model to obtain power generation prediction information in the target scheduling period; input the current power consumption information into a preset load prediction model to obtain power consumption load prediction information in the target scheduling period; or the energy prediction information includes net power load prediction information; the prediction module is further configured to input the current power generation information and the current power consumption information into a preset net load prediction model to obtain net power load prediction information in the target scheduling period.

[0179] The target scheduling period includes a plurality of time points, and the power prediction information includes power prediction information corresponding to each of the plurality of time points in the target scheduling period; the control information generation module is further configured to generate a plurality of sets of scenario prediction information based on the power prediction information and a plurality of sets of preset prediction error information; each set of prediction error information includes prediction error information corresponding to each of the plurality of time points in the target scheduling period, each set of scenario prediction information includes scenario prediction information corresponding to each of the plurality of time points in the target scheduling period, and each set of scenario prediction information corresponds to a scenario probability; the plurality of sets of scenario prediction information, the scenario probability corresponding to each set of scenario prediction information, and the electricity price information are input into a preset energy scheduling model, and when a target function value of the energy scheduling model is the lowest and a constraint condition is satisfied, the charging and discharging control information of the energy storage device in the target scheduling period is obtained; the constraint condition is established based on the power prediction information, current power information, and power grid outage information.

[0180] The device further includes an error distribution acquisition module configured to acquire an information distance between each two sets of scenario prediction information in the plurality of sets of scenario prediction information to obtain a distance matrix; each set of scenario prediction information is sequentially taken as current prediction information, a probability distance sum of the current prediction information and each set of scenario prediction information other than the current prediction information is determined based on an initial scenario probability of each set of scenario prediction information and the distance matrix to obtain a probability distance matrix; at least part of the plurality of sets of scenario prediction information is merged based on the distance matrix and the probability distance matrix to obtain new scenario prediction information and a scenario probability corresponding to the new scenario prediction information; and the plurality of sets of scenario prediction information is updated based on the new scenario prediction information.

[0181] The power prediction information includes power generation prediction information and power consumption load prediction information, and the scenario prediction information includes power generation scenario prediction information and load scenario prediction information; the control information generation module is further configured to, for a plurality of sets of prediction error information corresponding to the power generation prediction information, superimpose each set of prediction error information on the power generation prediction information according to time points to obtain a plurality of sets of power generation scenario prediction information; for a plurality of sets of prediction error information corresponding to the power consumption load prediction information, superimpose each set of prediction error information on the power consumption load prediction information according to time points to obtain a plurality of sets of load scenario prediction information; or the power prediction information includes net power consumption load prediction information, and the scenario prediction information includes net load scenario prediction information; the control information generation module is further configured to, for a plurality of sets of prediction error information corresponding to the net power consumption load prediction information, superimpose each set of prediction error information on the net power consumption load prediction information according to time points to obtain a plurality of sets of net load scenario prediction information.

[0182] The energy scheduling model includes: a power purchase cost and a power shortage cost, or a power purchase cost, a power shortage cost and a power selling income, or a power purchase cost, a power shortage cost and a storage cost, or a power purchase cost, a power shortage cost, a power selling income and a storage cost.

[0183] The constraint condition includes: a first condition, or the first condition and a second condition; the first condition includes: a power supply requirement condition of the storage device before the power failure time period arrives; the second condition includes one or more of the following conditions: the power purchase power and the power selling power are both zero during the power failure time period; a maximum value of the charging power value and a maximum value of the discharging power value of the storage device; a minimum value and a maximum value of the storage capacity of the storage device; the charging and discharging control information includes only a charging action or only a discharging action or neither a charging action nor a discharging action at the same time within the target scheduling time period; a power balance constraint condition.

[0184] The power supply requirement condition includes: a sum of the storage capacity of the storage device before the power failure and the power generation of the new energy device during the power failure, is greater than or equal to a sum of the power load of the power consumption device during the power failure and a first power supply amount; the first power supply amount is a product of a preset first power supply coefficient and the power load of the power consumption device during the power failure; or a sum of the storage capacity of the storage device before the power failure and a preset second power supply amount, is greater than or equal to the net load during the power failure.

[0185] When the objective function value of the energy scheduling model is the lowest and the constraint condition is satisfied, the charging and discharging control information of the storage device within the target scheduling time period includes a plurality of groups of charging and discharging control information, each group of charging and discharging control information includes charging control information or discharging control information at each time point in the target scheduling time period, and the device further includes: a final control information generation module configured to average the plurality of groups of charging and discharging control information according to the time points to obtain the charging control information or the discharging control information at each time point in the target scheduling time period as the final charging and discharging control information.

[0186] The device further includes: an electric energy prediction information updating module configured to collect the storage state of the storage device and corresponding load demand value and power generation; in a case where the electric energy prediction information includes power generation power prediction information and power load prediction information, update the load prediction model and the power generation power prediction model based on the storage state, the load demand value and the power generation; in a case where the electric energy prediction information includes net power load prediction information, update the net load prediction model based on the storage state, the load demand value and the power generation.

[0187] The power grid power failure information includes: power failure plan information, and / or power failure prediction information generated based on environmental data.

[0188] The embodiment also provides a user-side energy storage system, as shown in FIG. 5, which comprises: an energy storage device 52; a new energy power generation device 51, the energy storage device 52 and a power consumption device 53 are in communication connection with a server 54, and the server 54 is provided with the power information of the energy storage device, the power generation information of the new energy power generation device and the power consumption information of the power consumption device respectively, and the server further acquires the power grid outage information and the electricity price information, so that the server executes the charge-discharge control method of the energy storage device in any one of the above embodiments.

[0189] The server is also connected with a power grid 55 for acquiring power grid information such as power grid outage information and electricity price information.

[0190] It should be noted that the application scenario of the user-side energy storage system can be household, i.e. civil or commercial, and the power consumption device can be a household load, and the new energy power generation device can be a photovoltaic string.

[0191] In one example, as shown in FIG. 6, a four-day, i.e. 0-96 hours (unit: h) energy storage charge-discharge scheduling result is given. In the traditional energy storage charge-discharge scheduling scheme, i.e. the pre-outage full charging scheme in FIG. 6, when the system receives or learns that there will be a power outage event in the future, the energy storage device will be fully charged in advance, i.e. the SOC of the energy storage device reaches about 100%, such as at about 2 hours, about 10 hours, about 18 hours, etc. In addition, according to the charging behavior of the energy storage device, the energy storage device is manually fully charged at 10h in the morning of the first day, i.e. 10h in FIG. 6, at 8h in the morning of the second day, i.e. 32h in FIG. 6, at 8h in the morning of the third day, i.e. 56h in FIG. 6, etc. time, also causes the subsequent time period to occur the abandoned light / abandoned electricity behavior of the new energy device. However, in the case of using the scheme of the present application, i.e. the optimized scheduling scheme in FIG. 6, the system will determine the SOC of the energy storage device to be at a reasonable position according to the power consumption demand during the power outage, i.e. the load power in FIG. 6, and the power generation of the new energy device, such as the photovoltaic device, such as in the 2-4 hours before the power outage, the 10-12 hours before the power outage, the 32-34 hours before the power outage, the 48-50 hours before the power outage, etc. Time period, the energy storage device does not blindly charge first. And using the scheme of the present application in the simulation environment, through four days of energy storage charge-discharge optimization scheduling, about 39.50% of the power grid power can be saved, and the self-generation and self-use rate of the new energy device power generation is further improved, and the abandoned light amount is reduced.

[0192] The above embodiments of the present disclosure model the power generation and load demand by using the scene generation and scene reduction method, fully consider the random uncertainty of multiple scenarios by constructing multiple scenarios, so as to intelligently optimize and schedule the household energy storage system with power outage uncertainty.

[0193] The above embodiments of the present disclosure further reduce costs and increase benefits by dynamically and adaptively adjusting the standby power based on error statistics of historical power generation prediction and load prediction.

[0194] The above embodiments of the present disclosure further reduce the impact of uncertainty on scheduling by using the data assimilation idea to make online updates to the load prediction and power generation prediction models of the cloud scheduling center.

[0195] The embodiment also provides a server, including a processor and a memory, the memory storing computer executable instructions capable of being executed by the processor, and the processor executes the computer executable instructions to implement the above energy storage device charging and discharging control method.

[0196] Referring to FIG. 7, the server includes a processor 100 and a memory 101, the memory 101 storing computer executable instructions capable of being executed by the processor 100, and the processor 100 executes the computer executable instructions to implement the above energy storage device charging and discharging control method.

[0197] The embodiment also provides a computer readable storage medium storing computer executable instructions, and the computer executable instructions, when invoked and executed by a processor, cause the processor to implement the above energy storage device charging and discharging control method.

[0198] The energy storage device charging and discharging control method and device and the computer program product of the user side energy storage system provided by the embodiments of the present disclosure include a computer readable storage medium storing program codes, the instructions included in the program codes can be used to execute the method described in the foregoing method embodiments, and specific implementation can be referred to the method embodiments, which will not be described here.

[0199] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above described system and device can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0200] The above is only a specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited to this, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A charge and discharge control method of an energy storage device, characterized by, The method comprises: obtaining current power information of a power storage device, current power generation information of a new energy power generation device, current power consumption information of a power consumption device, and power price information and power grid outage information in a target scheduling period; obtaining power prediction information in the target scheduling period based on the current power generation information and the current power consumption information; obtaining charge-discharge control information of the power storage device in the target scheduling period based on the current power information, the power price information, the power grid outage information and the power prediction information; sending the charge-discharge control information to the power storage device, and the charge-discharge control information is used for the power storage device to perform charge-discharge control.

2. The method of claim 1, wherein, The power prediction information comprises power generation power prediction information and power consumption load prediction information; and the obtaining of the power prediction information in the target scheduling period based on the current power generation information and the current power consumption information comprises: inputting the current power generation information into a preset power generation power prediction model to obtain power generation power prediction information in the target scheduling period; and inputting the current power consumption information into a preset load prediction model to obtain power consumption load prediction information in the target scheduling period. Alternatively, the power prediction information comprises net power consumption load prediction information; and the obtaining of the power prediction information in the target scheduling period based on the current power generation information and the current power consumption information comprises: inputting the current power generation information and the current power consumption information into a preset net load prediction model to obtain net power consumption load prediction information in the target scheduling period.

3. The method of claim 1, wherein, The target scheduling period comprises a plurality of time points, the power prediction information comprises power prediction information corresponding to each time point in the plurality of time points in the target scheduling period; and the obtaining of the charge-discharge control information of the power storage device in the target scheduling period based on the current power information, the power price information, the power grid outage information and the power prediction information comprises: generating a plurality of groups of scene prediction information based on the power prediction information and a plurality of groups of preset prediction error information; wherein each group of the prediction error information comprises prediction error information corresponding to each time point in the plurality of time points in the target scheduling period, each group of the scene prediction information comprises scene prediction information corresponding to each time point in the plurality of time points in the target scheduling period, and each group of the scene prediction information corresponds to a scene probability; inputting the plurality of groups of the scene prediction information, the scene probability corresponding to each group of the scene prediction information and the power price information into a preset energy scheduling model, and obtaining the charge-discharge control information of the power storage device in the target scheduling period when a target function value of the energy scheduling model is lowest and a constraint condition is satisfied; wherein the constraint condition is established based on the power prediction information, the current power information and the power grid outage information.

4. The method of claim 3, wherein, After the generating of the plurality of groups of scene prediction information based on the power prediction information and the plurality of groups of preset prediction error information, the method further comprises: obtaining an information distance between each two groups of the scene prediction information in the plurality of groups of the scene prediction information to obtain a distance matrix. determining, based on the initial scene probability of each group of the scene prediction information and the distance matrix, a probability distance sum of the current prediction information and each group of the scene prediction information other than the current prediction information, to obtain a probability distance matrix; based on the distance matrix and the probability distance matrix, merging at least part of the groups of the scene prediction information to obtain new scene prediction information and a scene probability corresponding to the new scene prediction information; updating the groups of the scene prediction information based on the new scene prediction information.

5. The method of claim 3, wherein, The power prediction information includes power generation prediction information and power consumption load prediction information, and the scene prediction information includes power generation scene prediction information and load scene prediction information; the generating the groups of the scene prediction information based on the power prediction information and the preset groups of prediction error information comprises: for the groups of the prediction error information corresponding to the power generation prediction information, superimposing each group of the prediction error information on the power generation prediction information at a time point to obtain the groups of the power generation scene prediction information; for the groups of the prediction error information corresponding to the power consumption load prediction information, superimposing each group of the prediction error information on the power consumption load prediction information at a time point to obtain the groups of the load scene prediction information; alternatively, the power prediction information includes net power consumption load prediction information, and the scene prediction information includes net load scene prediction information; the generating the groups of the scene prediction information based on the power prediction information and the preset groups of prediction error information comprises: for the groups of the prediction error information corresponding to the net power consumption load prediction information, superimposing each group of the prediction error information on the net power consumption load prediction information at a time point to obtain the groups of the net load scene prediction information.

6. The method of claim 3, wherein, The energy scheduling model comprises: a power purchase cost and a power shortage cost, or a power purchase cost, a power shortage cost and a power selling income, or a power purchase cost, a power shortage cost and a storage cost, or a power purchase cost, a power shortage cost, a power selling income and a storage cost.

7. The method of claim 3, wherein, The constraint condition comprises: a first condition, or a first condition and a second condition; wherein the first condition comprises: a power supply demand condition of the energy storage device before a power outage time period arrives; and the second condition comprises one or more of the following conditions: in the power outage time period, the power purchase power and the power selling power are both zero; a maximum value of the charging power value and a maximum value of the discharging power value of the energy storage device; a minimum value and a maximum value of the energy storage capacity of the energy storage device; at the same time in the target scheduling time period, the charge-discharge control information only includes a charging action or only includes a discharging action or neither includes a charging action nor includes a discharging action; a power balance constraint condition.

8. The method of claim 7, wherein, The power supply demand condition comprises: The sum of the storage capacity of the energy storage device before power failure and the power generation of the new energy device during power failure is greater than or equal to the sum of the power load of the power consumption device during power failure and the first standby power amount; wherein the first standby power amount is the product of a preset first standby power coefficient and the power load of the power consumption device during power failure; Or, the sum of the storage capacity of the energy storage device before power failure and the preset second standby power amount is greater than or equal to the net load during power failure.

9. The method of claim 3, wherein, When the objective function value of the energy scheduling model is the lowest and the constraint condition is met, the charge-discharge control information of the energy storage device in the target scheduling period includes a plurality of groups of charge-discharge control information, and each group of the charge-discharge control information includes charge control information or discharge control information at each time point in the target scheduling period, and the method further comprises: Taking the average value of a plurality of groups of the charge-discharge control information according to time points to obtain charge control information or discharge control information at each time point in the target scheduling period as the final charge-discharge control information.

10. The method of claim 2, wherein, After the charge-discharge control information is sent to the energy storage device, the method further comprises: Collecting the energy storage state of the energy storage device, and the corresponding load demand value and power generation; In the case that the power prediction information includes power generation power prediction information and power consumption load prediction information, updating the load prediction model and the power generation power prediction model based on the energy storage state, the load demand value and the power generation; In the case that the power prediction information includes net power consumption load prediction information, updating the net load prediction model based on the energy storage state, the load demand value and the power generation.

11. A user-side energy storage system, characterized by The system comprises an energy storage device; The energy storage device, the new energy power generation device and the power consumption device are in communication connection with a server, and the server is provided with power information of the energy storage device, power generation information of the new energy power generation device and power consumption information of the power consumption device, respectively, and the server further acquires power grid power failure information and electricity price information, so that the server executes the charge-discharge control method of the energy storage device according to any one of claims 1-10.

12. A server, characterized by A processor and a memory are included, the memory stores computer executable instructions which can be executed by the processor, and the processor executes the computer executable instructions to realize the charge-discharge control method of the energy storage device according to any one of claims 1-10.

Images