Power management system and method

Abstract

A power management system is electrically coupled to a solar photovoltaic array, a power transmission network and a battery. The power management system includes a power converter and a processing module. The power converter is configured to operate in one of a plurality of operating modes, including a self-generating and self-using mode, a time electricity price mode, and a standby mode, to regulate power flowing between the solar photovoltaic array, the power transmission grid, and the battery. The processing module is configured to output prediction data representing the power demand in the first time period, the output power of the solar photovoltaic array and the electricity price of the power transmission network through a prediction model. One of a plurality of operation modes is set as a target mode of the power converter and a corresponding switching time according to a time sequence of the prediction data. And controlling the power converter to operate in the target mode at the switching time.

Description

Technical Field

[0001] This invention relates to power management, and more particularly to power management using artificial intelligence (AI) technology. Background Technology

[0002] Electricity management refers to the efficient production, distribution, storage, and consumption of electricity to ensure a stable and reliable power supply. It includes monitoring energy flux, optimizing efficiency, and balancing supply and demand to prevent waste and save costs. The three key components of modern electricity management are solar photovoltaic (PV) arrays, the transmission grid, and batteries. PV arrays generate electricity from sunlight, the transmission grid distributes and regulates power across large networks, and batteries store excess power for later use. The combination of these three elements enhances energy sustainability, transmission grid stability, and resilience to diverse application scenarios.

[0003] Conventional power management systems typically offer three common modes: self-consumption, time-of-use (TOU) pricing, and reserve. These modes utilize PV arrays, transmission networks, and batteries in different ways to help optimize energy efficiency, reduce costs, and improve stability.

[0004] In self-consumption mode, the system prioritizes using the power generated by the PV array before drawing power from the transmission grid and batteries. In TOU mode, the system considers the scenario where electricity prices are higher during peak hours and lower during off-peak hours. Therefore, during peak hours, the system prioritizes using the power generated by the PV array and batteries to supply load demand. During off-peak hours, the system prioritizes using the power generated by the PV array to supply load demand and charge the batteries. In standby mode, the system prioritizes charging the batteries.

[0005] However, without sufficient knowledge, ordinary users find it difficult to choose an appropriate mode to optimize electricity use. Users will face difficulties when analyzing factors such as electricity price trends, solar energy availability, and real-time battery state of charge (SOC). Furthermore, manually switching between various modes requires continuous monitoring and decision-making, which is both inconvenient and inefficient. Therefore, choosing an inappropriate mode can lead to relatively poor energy utilization, increased costs, and reduced system efficiency.

[0006] Therefore, there is an urgent need for a power management system and method that can solve the above problems. Summary of the Invention

[0007] An embodiment of the present invention provides a power management system for a building, electrically coupled to a solar photovoltaic (PV) array, a power grid, and a battery. The power management system includes: a power converter configured to operate in one of multiple operating modes to regulate the power flow between the PV array, the power grid, and the battery, wherein the operating modes include a self-consumption mode, a time-of-use (TOU) mode, and a standby mode; and a processing module configured to: generate forecast data through a forecast model, the forecast data representing the building's power demand, the PV array's power output, and the power grid's electricity price within a first time period; set one of the operating modes as a target mode for the power converter and a corresponding switching time based on a time series of the forecast data; and control the power converter to switch to the target mode operation at the switching time.

[0008] In one embodiment, the power converter is configured to: in self-consumption mode, prioritize the use of power generated by the PV array and power from the battery to supply the building's power needs; in TOU mode, prioritize the use of power generated by the PV array and power from the battery during peak hours to supply the building's power needs; and in off-peak hours, prioritize the use of power from the power grid to supply the building's power needs; and in standby mode, prioritize the use of power generated by the PV array to charge the battery.

[0009] In one embodiment, the processing module is further configured to: when the power converter operates in self-consumption mode, if the forecast data representing the electricity price of the transmission network shows that the electricity price of the transmission network will fall, and the forecast data representing the electricity demand of the building shows that the electricity demand of the building will decrease, then the target mode is set to standby mode; when the power converter operates in standby mode, if the forecast data representing the electricity price of the transmission network shows that the electricity price of the transmission network will rise, and the forecast data representing the output power of the PV array shows that the output power of the PV array will decrease, and the forecast data representing the electricity demand of the building shows that the electricity demand of the building will increase, then the target mode is set to self-consumption mode; and when the power converter operates in TOU mode, if the forecast data representing the electricity price of the transmission network shows that the electricity price of the transmission network will rise, and the forecast data representing the output power of the PV array shows that the output power of the PV array will increase, and the forecast data representing the electricity demand of the building shows that the electricity demand of the building will decrease, then the target mode is set to self-consumption mode.

[0010] In one embodiment, the processing module is further configured to: record data representing the building's power demand, the PV array's output power, and the power price of the transmission network during a second time period, wherein the second time period occurs within the first time period, after the target mode is determined, and before the switching time; compare the predicted data and the recorded data during the second time period; and, in response to the matching of the predicted data and the recorded data during the second time period, control the power converter to switch to operating in the target mode at the switching time.

[0011] In one embodiment, the processing module is further configured to: switch the power converter from self-consumption mode to TOU mode if the building's power demand does not decrease, in response to a discrepancy between the predicted data and the recorded data during the second time period; switch the power converter from standby mode to TOU mode if the PV array's output power decreases; and switch the power converter from TOU mode to self-consumption mode if the battery's state of charge (SOC) exceeds a charging neighbor value.

[0012] In one embodiment, the processing module is further configured to: maintain the current operating mode of the power converter in response to the discrepancy between the predicted data and the recorded data during the second time period; and repeatedly record the recorded data representing the building's power demand, the PV array's output power, and the electricity price of the transmission network during the second time period, and compare the predicted data and the recorded data during the second time period.

[0013] In one embodiment, the processing module is further configured to: generate, through a prediction model, a prediction data with a time resolution less than or equal to that of a second time period, based on historical data representing the building's power demand, the PV array's power output, and the electricity price of the transmission network during a past time period, wherein the past time period is longer than the first time period.

[0014] In one embodiment, the prediction model is trained on at least one of a weather database, an appliance database, a PV generation dataset, and an electricity demand database.

[0015] In one embodiment, the processing module is further configured to: detect, before setting the target mode, whether the electricity price of the transmission network, the state of charge (SOC) of the battery, and the output power of the PV array meet a criterion corresponding to the current operating mode of the power converter; and maintain the current operating mode of the power converter in accordance with the criterion.

[0016] In one embodiment, the criteria include: corresponding to the self-consumption mode, the grid electricity price and the battery's SOC need to exceed a maximum critical price and a maximum charging critical value, respectively; corresponding to the TOU mode, the grid electricity price and the power generated by the PV array need to be lower than a minimum critical price and a minimum PV threshold, respectively; and corresponding to the standby mode, the grid electricity price needs to be lower than a minimum critical price and the battery's SOC needs to exceed a maximum charging critical value.

[0017] An embodiment of the present invention provides a power management method implemented on a system applied to a building, wherein the system includes a power converter and a processing module, and is electrically coupled to a PV array, a transmission network, and a battery. The power management method includes: the power converter operating in one of multiple operating modes to regulate the power flowing between the PV array, the transmission network, and the battery, wherein the operating modes include a self-consumption mode, a time-of-use (TOU) mode, and a standby mode; and the processing module: generating forecast data through a forecast model, the forecast data representing the building's power demand, the PV array's power output, and the transmission network's electricity price within a first time period; setting one of the multiple operating modes as a target mode for the power converter and a corresponding switching time based on a time series of the forecast data; and controlling the power converter to switch to operating in the target mode at the switching time.

[0018] The power management system provided by this invention automatically sets an operating mode for the power converter to regulate the power flow between PV arrays, the transmission network, and batteries. Specifically, by applying AI technology, the power management system can flexibly and accurately switch the power converter between different operating modes according to past, present, and future data, adapting to various scenarios. This maximizes energy efficiency, reduces costs, and improves reliability. Attached Figure Description

[0019] A more complete understanding of the present invention can be obtained by referring to the following detailed description and embodiments, in which:

[0020] Figure 1 This is a structural diagram of a power management system according to an embodiment of the present disclosure;

[0021] Figure 2 This is a flowchart illustrating how a processing module executes a power management method according to an embodiment of this disclosure;

[0022] Figure 3 This is a schematic diagram illustrating the operation of determining a target mode and a switching time according to an embodiment of the present disclosure;

[0023] Figure 4This is a flowchart illustrating how to control a power converter to operate in a target mode during a switching time, according to an embodiment of this disclosure.

[0024] Figure 5 This is a schematic diagram illustrating a discrepancy between predicted data and recorded data after comparing predicted data with recorded data, according to an embodiment of this disclosure.

[0025] Figure 6 This is a schematic diagram illustrating the process before determining the target mode and switching time, according to an embodiment of this disclosure.

[0026] The reference numerals in the attached figures are explained as follows:

[0027] 10: Power Management System

[0028] 101: Power Converter

[0029] 102: Processing Module

[0030] 11: PV array

[0031] 12: Power Transmission Network

[0032] 13: Battery

[0033] 20: Power Management Methods

[0034] 200, 201, 202, 203, 2031, 2032, 2033, 2034: Operations

[0035] 202_1, 202_2, 202_3, 2034_1, 2034_2, 2034_3, 200_1, 200_2, 200_3: Criteria Detailed Implementation

[0036] The following description is intended to illustrate the general principles of this disclosure and should not be construed as limiting. The scope of the invention is best determined by referring to the appended claims.

[0037] In the following embodiments, similar reference numerals are used to denote similar or equal elements.

[0038] The sequential numbering in the specification and claims, such as "first" and "second", is for illustrative purposes only and there is no order between them.

[0039] The descriptions of apparatus or systems in the embodiments of this disclosure also apply to the methods in the embodiments, and vice versa.

[0040] Figure 1 This is a structural diagram of a power management system 10 according to an embodiment of the present disclosure. The power management system 10 can be installed in a building. Figure 1As shown, the power management system 10 includes a power converter 101 and a processing module 102, and is electrically coupled to a solar photovoltaic (PV) array 11, a power transmission network 12, and a battery 13.

[0041] The power management system 10 can be implemented on any computing system with computing capabilities, such as a microcontroller, a personal computer (e.g., a desktop computer or a laptop), a server computer, or a mobile device (e.g., a tablet or a smartphone), or on a computer cluster consisting of multiple computers working together, but this disclosure is not limited to the above.

[0042] The power converter 101 can be implemented on any computing system with computing capabilities, such as a microcontroller, a personal computer (desktop or laptop), a server computer, or a mobile device (tablet or smartphone). Alternatively, the power converter 101 can also be implemented using integrated circuits, such as application-specific integrated circuits (ASICs), systems on a chip (SOCs), or field-programmable gate arrays (FPGAs), but this disclosure is not limited to the above.

[0043] Processing module 102 may include one or more general-purpose or special-purpose processors and combinations thereof to execute instructions, such as a central processing unit (CPU) and / or a graphics processing unit (GPU). Processing module 102 may also include volatile memory, such as dynamic random access memory (DRAM) and / or static random access memory (SRAM), but this disclosure is not limited to the above.

[0044] In one embodiment, the power converter 101 can operate in one of several operating modes to regulate the power flowing between the PV array 11, the power grid 12, and the battery 13. For example, the power converter 101 may designate the PV array 11 and the battery 13 as power sources, or the power converter 101 may designate the power grid 12 as a power source alone.

[0045] In one embodiment, the operating modes include a self-consumption mode, a time-of-use (TOU) mode, and a standby mode.

[0046] In self-consumption mode, the power converter 101 prioritizes the use of power generated by the PV array 11 and the battery 13 to supply the building's power needs. In one embodiment, when the power generated by the PV array 11 exceeds the power demand, the power converter 101 can further utilize the power generated by the PV array 11 to charge the battery 13. In another embodiment, when the power generated by the PV array 11 is insufficient to meet the power demand, the power converter 101 can further utilize the power from the power grid 12 or the battery 13.

[0047] In TOU mode, during peak hours, power converter 101 can prioritize the use of power generated by PV array 11 and power from battery 13 to supply the building's power needs. During off-peak hours, power converter 101 prioritizes the use of power from transmission grid 12 to supply the building's power needs. In one embodiment, power converter 101 can further utilize power from transmission grid 12 to charge battery 13 during off-peak hours.

[0048] In another embodiment of the TOU mode, during off-peak hours, the power converter 101 may preferentially use the power generated by the PV array 11 to supply the building's power needs and charge the battery 13. If the power generated by the PV array 11 is insufficient to supply the power needs, the power converter 101 may further utilize the power from the transmission network 12.

[0049] In standby mode, the power converter 101 can prioritize charging the battery 13 with the power generated by the PV array 11. In addition, if the power generated by the PV array 11 is insufficient, the power converter 101 can draw power from the power grid 12 to ensure that the battery 13 reaches a preset charging state, thereby maintaining the availability of backup power.

[0050] Figure 2 This is a flowchart illustrating the execution of power management method 20 by processing module 102 according to an embodiment of this disclosure. Figure 2 As shown, power management method 20 includes operations 201 to 203.

[0051] In operation 201, processing module 102 can generate predicted data (e.g., for the next 3 hours, 6 hours, or 12 hours) representing the building's power demand, the PV array's power output, and the electricity price of the transmission network within a first time period through a prediction model.

[0052] In one embodiment, the processing module 102 generates predictive data using a predictive model based on historical data representing the building's electricity demand, the PV array's power output, and the transmission network's electricity price over a past period. The past period is longer than the first period. The predictive model is trained on historical data and learns the patterns and correlations of electricity demand, PV array power output, and transmission network electricity prices, enabling it to produce accurate predictions for future periods, such as the first period mentioned above.

[0053] In one embodiment, the processing module 102 generates prediction data through a prediction model, which is based not only on historical data but also on external variables, including but not limited to rainfall rate, temperature, humidity, and / or the electricity consumed by household appliances.

[0054] In one embodiment, the predictive model can be implemented as a time series model, which uses the temporal dependencies and inherent patterns of historical data to predict future values. For example, the predictive model may apply an Autoregressive Integrated Moving Average (ARIMA), a Seasonal ARIMA (SARIMA), a Long Short-Term Memory (LSTM) model, or a Holt-Winters model, and this disclosure is not limited to the above.

[0055] In one embodiment, the prediction model may be implemented as a regression model, such as decision tree regression, random forest regression, extreme gradient boosting (XGboost), or neural networks (NN), but this disclosure is not limited to the above.

[0056] In one embodiment, the prediction model is trained on a weather database, a household appliance database, a PV generation dataset, or an electricity demand database. The weather database represents environmental data, such as rainfall rate, temperature, humidity, and / or air pressure. The household appliance database represents appliance data, such as the compressor rate and fan rate of an air conditioner. Depending on the algorithm used by the prediction model, the model can learn from these training databases in a supervised or unsupervised manner.

[0057] In operation 202, the processing module 102 can set a target mode for the power converter 101 based on predicted data and set a corresponding switching time. This setting can be based on factors such as predicted electricity prices, expected solar energy output, and battery charging status to optimize energy use. In operation 203, the processing module 102 can control the power converter 101 to switch to the target mode at the switching time to ensure efficient power management and save energy costs.

[0058] Figure 3 According to one embodiment of this disclosure, it is shown that Figure 2 A schematic diagram of operation 202. (See diagram below.) Figure 3 As shown, operation 202 may further include the application of guidelines 202_1 to 202_3.

[0059] In one embodiment, criterion 202_1 may correspond to a self-consumption mode, wherein the predicted data indicating the electricity price of the transmission network must show that the electricity price of the transmission network will decrease, and the predicted data indicating the electricity demand of the building must show that the electricity demand of the building will decrease. If criterion 202_1 is met, the processing module 102 can set the target mode of the power converter 101, which was originally operating in self-consumption mode, to a standby mode. If criterion 202_1 is not met, the processing module 102 can set the target mode of the power converter 101, which was originally operating in self-consumption mode, to self-consumption mode.

[0060] In one embodiment, criterion 202_2 may correspond to a standby mode, wherein the predicted data for the transmission network's electricity price must indicate that the transmission network's electricity price will increase, the predicted data for the PV array's output power must indicate that the PV array's output power will decrease, and the predicted data for the building's electricity demand must indicate that the building's electricity demand will increase. If criterion 202_2 is met, the processing module 102 can set the target mode of the power converter 101, which was originally operating in standby mode, to a self-consumption mode. If criterion 202_2 is not met, the processing module 102 can set the target mode of the power converter 101, which was originally operating in standby mode, to standby mode.

[0061] In one embodiment, criterion 202_3 may correspond to the TOU mode, wherein the predicted data for the transmission network's electricity price must indicate that the transmission network's electricity price will increase, the predicted data for the PV array's output power must indicate that the PV array's output power will increase, and the predicted data for the building's electricity demand must indicate that the building's electricity demand will decrease. If criterion 202_3 is met, the processing module 102 can set the target mode of the power converter 101, which was originally operating in TOU mode, to a self-consumption mode. If criterion 202_3 is not met, the processing module 102 can set the target mode of the power converter 101, which was originally operating in TOU mode, to TOU mode.

[0062] In one embodiment, the trend of the predicted data can be observed using various mathematical and algorithmic methods, such as: Rate of Charge (ROC) calculation, Moving Average (MA), Regression Analysis, or Time Series Analysis (e.g., ARIMA sequence model), but this disclosure is not limited to the above.

[0063] Figure 4 According to one embodiment of this disclosure, it is shown that Figure 2 The flowchart for operation 203. (Example) Figure 4 As shown, operation 203 further includes operations 2031 to 2033.

[0064] In operation 2031, processing module 102 can record data representing the building's power demand, PV array output power, and grid electricity price during a second time period. The second time period occurs within the first time period, after the target mode has been set and before the switching time arrives.

[0065] In operation 2032, the processing module 102 can compare the predicted data and the recorded data in the second time period to verify whether the prediction is accurate.

[0066] In operation 2033, if the predicted data in the second time period matches the recorded data, the processing module 102 can control the power converter 101 to operate in the target mode at the switching time. If the predicted data matches the recorded data, it means that the prediction of the processing module 102 is accurate; therefore, the processing module 102 does not need to reset the target mode. The processing module 102 can control the power converter 101 to operate in the previously set target mode at the previously set switching time.

[0067] Conversely, if the predicted data does not match the recorded data, it means that the prediction of the processing module 102 is inaccurate. Therefore, the processing module 102 needs to reset the target mode.

[0068] In one embodiment, if the predicted data and recorded data do not match in the second time period, the processing module 102 can maintain the current operating mode of the power converter 101 and repeat operations 2031 and 2032 until the predicted data and recorded data match in the second time period. When the predicted data and recorded data match, the processing module 102 executes operation 2033 to control the power converter 101 to switch to target mode operation.

[0069] Figure 5This is a schematic diagram illustrating operation 2034 according to an embodiment of this disclosure. If the predicted data does not match the recorded data, operation 2034 is executed after operation 2032. Figure 5 As shown, operation 2034 may further include the application of guidelines 2034_1 to 2034_3.

[0070] In one embodiment, criterion 2034_1 may correspond to a self-consumption mode, where the building's electricity demand needs to be reduced. If criterion 2034_1 is not met, processing module 102 can switch power converter 101 from self-consumption mode to TOU mode. If criterion 2034_1 is met, processing module 102 can keep power converter 101 operating in self-consumption mode. Specifically, processing module 102 can reset the target mode of power converter 101, which was previously operating in self-consumption mode, to TOU mode. Then, processing module 102 can control power converter 101 to immediately operate in TOU mode, rather than waiting for a switching time.

[0071] In one embodiment, criterion 2034_2 may correspond to a standby mode, where the power output of the PV array needs to be reduced. If criterion 2034_2 is met, processing module 102 can switch power converter 101 from standby mode to TOU mode. If criterion 2034_2 is not met, processing module 102 can keep power converter 101 operating in standby mode. Specifically, processing module 102 can reset the target mode of power converter 101, which was previously operating in standby mode, to TOU mode. Then, processing module 102 can control power converter 101 to immediately operate in TOU mode.

[0072] In one embodiment, criterion 2034_3 may correspond to the TOU mode, where a state of charge (SOC) of the battery must be higher than a charging neighbor value. If criterion 2034_3 is met, the processing module 102 can switch the power converter 101 from the TOU mode to the self-consumption mode. If criterion 2034_3 is not met, the processing module 102 can keep the power converter 101 operating in the TOU mode. Specifically, the processing module 102 can reset the target mode of the power converter 101, which was originally operating in the TOU mode, to the self-consumption mode. Then, the processing module 102 can control the power converter 101 to immediately operate in the self-consumption mode.

[0073] The aforementioned charging thresholds can be set based on factors including battery type, system requirements, and power management strategies. Typically, manufacturers provide recommended SOC ranges to extend battery life; for example, the charging threshold for lithium-ion batteries is usually 80-90% to prevent overcharging and extend battery life. Lead-acid batteries, on the other hand, can be charged to 100% for optimal performance. This disclosure is not limited to the above.

[0074] In particular, verifying whether the predicted data matches the recorded data helps assess the accuracy of the predictive model. This prevents the system from relying on an inaccurate model to set the target mode. Therefore, a more appropriate target mode can be selected, improving power utilization efficiency.

[0075] In one embodiment, the processing module 102 may... Figure 2 Before operation 202, operation 200 is further performed. In operation 200, processing module 102 can confirm whether a criterion corresponding to the current operating mode of power converter 101 is met. The criterion includes at least two of the following: the electricity price of the transmission network, the state of charge (SOC) of the battery, and the power output of the PV array. Then, if the corresponding criterion is met, processing module 102 can maintain the current operating mode of power converter 101.

[0076] Figure 6 This is a schematic diagram illustrating operation 200 according to an embodiment of the present disclosure. Figure 6 As shown, operation 200 may further include the application of guidelines 200_1 to 200_3.

[0077] In one embodiment, criterion 200_1 may correspond to a self-consumption mode, where the grid electricity price and the battery's SOC must be high. Criterion 200_2 may correspond to a TOU mode, where both the grid electricity price and the PV array's output power must be low. Criterion 200_3 may correspond to a standby mode, where the grid electricity price must be low and the battery's SOC must be high.

[0078] Specifically, Criterion 200_1 may include that the electricity price of the transmission network must exceed a maximum critical price, and the state of charge (SOC) of the battery must exceed a maximum charging critical value. Criterion 200_2 may include that the electricity price of the transmission network must be lower than a minimum critical price, and the power output of the PV array must be lower than a minimum PV critical value. Criterion 200_3 may include that the electricity price of the transmission network must be lower than a minimum critical price, and the state of charge (SOC) of the battery must exceed a maximum charging critical value.

[0079] The aforementioned maximum critical electricity price, maximum charging critical value, minimum critical electricity price, and minimum PV threshold can be set using machine learning models or statistical methods based on relevant historical data. For example, the maximum and minimum critical electricity prices can be set based on past electricity price fluctuations or historical energy utilization patterns. Alternatively, the minimum critical electricity price can be set based on historical solar energy production patterns.

[0080] In one embodiment, if the corresponding criteria are not met, the processing module 102 may randomly switch the power converter 101 to another mode. For example, the processing module 102 may switch the power converter 101, which was originally operating in TOU mode, to standby mode. Alternatively, the processing module 102 may switch the power converter 101, which was originally operating in self-consumption mode, to TOU mode. This disclosure is not limited to the above.

[0081] In one embodiment, the predicted data and recorded data are represented by a chart. In another embodiment, the processing module 102 updates the chart when the predicted data changes and waits 30 minutes when the predicted data remains unchanged.

[0082] In one embodiment, when the chart is updated, the processing module 102 confirms the above criteria and continues to perform operation 201. In particular, by setting the above criteria, the switching time, and reducing chart updates, the system can avoid switching between different operating modes too frequently.

[0083] The power management system disclosed herein can automatically set an operating mode for power converters to regulate the power flowing between PV arrays, transmission networks, and batteries. Specifically, by applying AI technology, the power management system can flexibly and accurately switch power converters between different operating modes according to past, present, and future data, adapting to various scenarios. This maximizes energy efficiency, reduces costs, and improves reliability.

[0084] The preceding paragraphs describe the subject in various ways. Clearly, the teachings of this document can be implemented in multiple ways, and any particular architecture or functionality disclosed in the examples is merely representative. Based on the teachings of this document, those skilled in the art should understand that the various aspects disclosed herein can be implemented independently, or two or more aspects can be combined and implemented.

[0085] Although this disclosure has been described with reference to the above embodiments, they are not intended to limit the scope of this disclosure. Those skilled in the art can make modifications without departing from the spirit and scope of this disclosure. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A power management system applied to a building, the power management system electrically coupled to a solar array, a power grid, and a battery, the power management system comprising: a power converter configured to operate in one of a plurality of operating modes to regulate power flow among the solar array, the power grid, and the battery, wherein the operating modes include a self-consumption mode, a time-of-use mode, and a backup mode; and a processing module configured to: output, by a prediction model, a prediction data representing a power demand of the building, an output power of the solar array, and a price of the power grid for a first time period; set, according to a time series of the prediction data, one of the operating modes as a target mode of the power converter and a corresponding switching time; and control the power converter to switch to operate in the target mode at the switching time.

2. The power management system of claim 1, wherein the power converter is configured to: in the self-consumption mode, prioritize using the output power of the solar array and a power of the battery to supply the power demand of the building; in the time-of-use mode, in an on-peak period, prioritize using the output power of the solar array and the power of the battery to supply the power demand of the building; and in an off-peak period, prioritize using a power of the power grid to supply the power demand of the building; and in the backup mode, prioritize using the output power of the solar array to charge the battery.

3. The power management system of claim 1, wherein the processing module is further configured to: in response to the power converter operating in the self-consumption mode, set the target mode to the backup mode if the prediction data representing the price of the power grid indicates that the price of the power grid will decrease and the prediction data representing the power demand of the building indicates that the power demand of the building will decrease; in response to the power converter operating in the backup mode, set the target mode to the self-consumption mode if the prediction data representing the price of the power grid indicates that the price of the power grid will increase and the prediction data representing the output power of the solar array indicates that the output power of the solar array will decrease and the prediction data representing the power demand of the building indicates that the power demand of the building will increase; and in response to the power converter operating in the time-of-use mode, set the target mode to the self-consumption mode if the prediction data representing the price of the power grid indicates that the price of the power grid will increase and the prediction data representing the output power of the solar array indicates that the output power of the solar array will increase and the prediction data representing the power demand of the building indicates that the power demand of the building will decrease.

4. The power management system of claim 1, wherein the processing module is further configured to: ​ ​ ​ record data representing the power demand of the building, the output power of the solar array, and the electricity price of the power grid during a second time period, wherein the second time period is within the first time period, and after the target mode is determined, and before the switching time is reached; compare the predicted data and the recorded data during the second time period; and in response to the predicted data and the recorded data during the second time period matching, control the power converter to switch to operate in the target mode at the switching time.

5. The power management system of claim 4, wherein in response to the predicted data and the recorded data during the second time period not matching, the processing module is further configured to: if the power demand of the building does not decrease, switch the power converter from the self-generate self-consume mode to the time-of-use mode; if the output power of the solar array decreases, switch the power converter from the backup mode to the time-of-use mode; and if a state of charge of the battery exceeds a maximum charge threshold, switch the power converter from the time-of-use mode to the self-generate self-consume mode.

6. The power management system of claim 4, wherein the processing module is further configured to: in response to the predicted data and the recorded data during the second time period not matching, maintain the current operating mode of the power converter; and repeat the operations of recording the recorded data representing the power demand of the building, the output power of the solar array, and the electricity price of the power grid during the second time period, and comparing the predicted data and the recorded data during the second time period, until the predicted data and the recorded data during the second time period match.

7. The power management system of claim 4, wherein the processing module is further configured to generate, by the prediction model, the prediction data with a time resolution less than or equal to the second time period based on historical data representing the power demand of the building, the power output of the solar array, and the electricity price of the power grid over a past time period, wherein, the past time period is longer than the first time period.

8. The power management system of claim 4, wherein before setting the target mode, the processing module is further configured to: detect whether the electricity price of the power grid, a state of charge of the battery, and the output power of the solar array satisfy a criterion corresponding to the current operating mode of the power converter; and in response to satisfying the corresponding criterion, maintain the current operating mode of the power converter.

9. The power management system of claim 8, wherein the criterion comprises: for the self-generate self-consume mode, the electricity price of the power grid needs to exceed a maximum threshold price, and the state of charge of the battery needs to exceed a maximum charge threshold; for the time-of-use mode, the electricity price of the power grid needs to be below a minimum threshold price, and the output power of the solar array needs to be below a minimum solar threshold; and for the backup mode, the electricity price of the power grid needs to be below a minimum threshold price, and the state of charge of the battery needs to exceed a maximum charge threshold.

10. The power management system of claim 1, wherein the prediction model is trained with at least one of a weather database, an appliance database, a solar output database, and a power demand database. ​ 11. A power management method performed on a system, wherein the system is applied to a building, and the system comprises a power converter and a processing module, and is electrically coupled to a solar array, a power grid, and a battery, the power management method comprising: operating, by the power converter, in one of a plurality of operating modes to regulate power flow among the solar array, the power grid, and the battery, wherein the operating modes comprise a self-consumption mode, a time-of-use (TOU) mode, and a backup mode; and setting, by the processing module: a target mode and a corresponding switching time among the operating modes for the power converter according to a time series of prediction data representing a power demand of the building, a power output of the solar array, and a price of the power grid for a first time period; and controlling the power converter to switch to operate in the target mode at the switching time.

12. The power management method of claim 11, further comprising: setting, by the power converter: in the self-consumption mode, a priority order of using the power output of the solar array and a power of the battery to supply the power demand of the building; in the TOU mode, a priority order of using the power output of the solar array and the power of the battery to supply the power demand of the building during a peak period, and a priority order of using a power of the power grid to supply the power demand of the building during an off-peak period; and in the backup mode, a priority order of using the power output of the solar array to charge the battery.

13. The power management method of claim 11, further comprising: setting, by the processing module: in response to the power converter operating in the self-consumption mode, the target mode to the backup mode if the prediction data representing the price of the power grid indicates that the price of the power grid will decrease and the prediction data representing the power demand of the building indicates that the power demand of the building will decrease; in response to the power converter operating in the backup mode, the target mode to the self-consumption mode if the prediction data representing the price of the power grid indicates that the price of the power grid will increase and the prediction data representing the power output of the solar array indicates that the power output of the solar array will decrease and the prediction data representing the power demand of the building indicates that the power demand of the building will increase; and in response to the power converter operating in the TOU mode, the target mode to the self-consumption mode if the prediction data representing the price of the power grid indicates that the price of the power grid will increase and the prediction data representing the power output of the solar array indicates that the power output of the solar array will increase and the prediction data representing the power demand of the building indicates that the power demand of the building will decrease.

14. The power management method of claim 11, further comprising: setting, by the processing module: ​ ​ ​ ​ recording recorded data representing the power demand of the building, the output power of the solar array, and the electricity price of the power grid for a second time period, wherein the second time period is within the first time period, and after the target mode is determined, and before the switching time is reached; comparing the predicted data and the recorded data for the second time period; and in response to the predicted data and the recorded data for the second time period matching, controlling the power converter to operate in the target mode at the switching time.

15. The power management method of claim 14, further comprising: in response to the predicted data and the recorded data for the second time period not matching, the processing module: switching the power converter from the self-generate self-consume mode to the time-of-use mode if the power demand of the building does not decrease; switching the power converter from the backup mode to the time-of-use mode if the output power of the solar array decreases; and switching the power converter from the time-of-use mode to the self-generate self-consume mode if a state of charge of the battery exceeds a maximum charge threshold.

16. The power management method of claim 14, further comprising: the processing module: in response to the predicted data and the recorded data for the second time period not matching, maintaining the current operating mode of the power converter; and repeating the operations of recording the recorded data representing the power demand of the building, the output power of the solar array, and the electricity price of the power grid for the second time period, and comparing the predicted data and the recorded data for the second time period, until the predicted data and the recorded data for the second time period match.

17. The power management method of claim 14, further comprising: the processing module: using the prediction model to generate the predicted data having a time resolution less than or equal to the second time period based on historical data representing the power demand of the building, the output power of the solar array, and the electricity price of the power grid for a past time period, wherein the past time period is longer than the first time period.

18. The power management method of claim 14, further comprising: before setting the target mode, the processing module: detecting whether the electricity price of the power grid, a state of charge of the battery, and the output power of the solar array satisfy a criterion corresponding to the current operating mode of the power converter; and in response to satisfying the corresponding criterion, maintaining the current operating mode of the power converter.

19. The power management method of claim 18, wherein the criterion comprises: for the self-generate self-consume mode, the electricity price of the power grid exceeding a maximum threshold price, and the state of charge of the battery exceeding a maximum charge threshold; for the time-of-use mode, the electricity price of the power grid being lower than a minimum threshold price, and the output power of the solar array being lower than a minimum solar threshold; and for the backup mode, the electricity price of the power grid being lower than a minimum threshold price, and the state of charge of the battery exceeding a maximum charge threshold. ​ ​ 20. The power management method of claim 11, wherein the prediction model is trained with at least one of a weather database, an appliance database, a solar power production database, and a power demand database.

Images