Device adaptive control method and apparatus, energy management system, and storage medium

Abstract

The application relates to the technical field of control, and provides a device adaptive control method and device, an energy management system and a storage medium. The method comprises the following steps: in response to a target input operation of a user, determining at least two energy management targets corresponding to the energy management system; constructing a multi-target function based on the energy management targets; determining dynamic weights corresponding to the energy management targets based on running state data of controlled devices in the energy management system; performing multi-target optimization based on the multi-target function and the dynamic weights to generate a control strategy; and controlling the controlled devices to run according to the control strategy. The above method can improve the multi-target comprehensive achievement rate and the system adaptive capability.

Description

Technical Field

[0001] This invention relates to the field of control technology, and in particular to an adaptive control method, device, energy management system, and storage medium for equipment. Background Technology

[0002] Among related technologies, Home Energy Management Systems (HEMS) and Park Energy Management Systems (PEMS) generally suffer from two core problems. First, current energy management systems have a single optimization objective, making it difficult to simultaneously consider multiple dimensions such as electricity costs, green electricity ratio, energy storage lifespan, and backup power security, resulting in suboptimal energy management performance. Second, control strategies are complex, requiring users to manually configure multiple parameters for different devices, leading to cumbersome operation and negatively impacting user experience. Therefore, existing solutions suffer from poor user experience and insufficient intelligence. Summary of the Invention

[0003] In view of the above, it is necessary to provide a device for adaptive control, an energy management system and a storage medium to solve the technical problems of poor user experience and insufficient intelligence level.

[0004] This application provides an adaptive control method for devices applied to control devices in an energy management system. The method includes: responding to a user's target input operation, determining at least two energy management objectives corresponding to the energy management system; constructing a multi-objective function based on the energy management objectives; determining dynamic weights corresponding to each energy management objective based on the operating status data of the controlled devices in the energy management system; performing multi-objective optimization based on the multi-objective function and the dynamic weights to generate a control strategy; and controlling the operation of the controlled devices according to the control strategy.

[0005] The device adaptive control method provided in this application determines at least two energy management objectives in response to user input operations and constructs a multi-objective function based on these objectives. Compared to traditional energy management control schemes with a single optimization objective, this application adapts to diverse user needs through multiple energy management objectives and objective functions, improving the scenario adaptability of the energy management system. It performs multi-objective optimization and generates control strategies based on the dynamic weights of the controlled device's operating status data. Specifically, by adjusting the weights of each energy management objective, the optimization direction of the energy management objectives is precisely matched with the actual operating status and scenario requirements of the controlled device, significantly improving the rationality and accuracy of the control strategy. This results in a more automated and intelligent control strategy, significantly improving energy utilization efficiency, the overall achievement rate of multiple objectives, and user experience. Through this device adaptive control method, the user's operational threshold can be effectively reduced, and the overall achievement rate of multiple objectives and the system's adaptive capability can be improved.

[0006] This application embodiment also provides a device adaptive control apparatus, the apparatus comprising: a determining module, configured to determine at least two energy management objectives corresponding to the energy management system in response to a user's target input operation; a constructing module, configured to construct a multi-objective function based on the energy management objectives; the determining module is further configured to determine dynamic weights corresponding to each of the energy management objectives based on the operating status data of the controlled devices in the energy management system; a generating module, configured to perform multi-objective optimization based on the multi-objective function and the dynamic weights to generate a control strategy; and a control module, configured to control the operation of the controlled devices according to the control strategy.

[0007] This application embodiment also provides an energy management system, the energy management system including a control device, the control device including: at least one processor; and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executed by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the device adaptive control method as described above.

[0008] This application also provides a computer-readable storage medium storing at least one instruction that, when executed by a processor, implements the device adaptive control method described above. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the application environment of the device adaptive control method provided in an embodiment of this application.

[0010] Figure 2 This is a schematic diagram of the structure of a control device provided in an embodiment of this application.

[0011] Figure 3 This is a flowchart of a device adaptive control method provided in an embodiment of this application.

[0012] Figure 4 This is a schematic diagram of an energy management template provided in an embodiment of this application.

[0013] Figure 5 This is a flowchart illustrating the sub-step of step S303 in the device adaptive control method provided in an embodiment of this application.

[0014] Figure 6 This is a flowchart of a device adaptive control method provided in another embodiment of this application.

[0015] Figure 7 This is a schematic diagram of the structure of an adaptive control device for equipment provided in an embodiment of this application. Detailed Implementation

[0016] For ease of understanding, some concepts related to the embodiments of this application are illustrated and explained by way of example for reference.

[0017] It should be noted that the terms "first", "second", "third", "fourth", "fifth", etc. (if present) in the specification, claims and drawings of this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0018] To better understand the device adaptive control method, apparatus, energy management system, and storage medium provided in the embodiments of this application, the application environment of the device adaptive control method provided in this application is described below.

[0019] Figure 1 This is a schematic diagram illustrating the application environment of the device adaptive control method applicable to this application, provided as an embodiment of the present application. The device adaptive control method is applied in an energy management system 1. The energy management system 1 includes a control device 10 and a controlled device 20, and the control device 10 and the controlled device 20 are communicatively connected.

[0020] In this embodiment, the control device 10 may include an Energy Management System (EMS) function. Based on the EMS function, the control device 10 can receive user commands, collect operating status data of each controlled device 20 within the energy management system 1, execute the device adaptive control method described in this application to perform multi-objective optimization decisions, and distribute the generated control strategy to the corresponding controlled device 20 for execution, thereby achieving centralized monitoring, intelligent scheduling, and optimized management of energy flow.

[0021] In some embodiments of this application, the control device 10 in the energy management system 1 can be a standalone device, such as a separately configured industrial control computer. In other embodiments of this application, the control device 10 in the energy management system 1 can also be integrated with other devices. Alternatively, the control device 10 can be a controller for a certain device. For example, the control device 10 can be a controller in a power conversion device.

[0022] In some embodiments of this application, the controlled device 20 may be a photovoltaic device, an energy storage device, an air conditioner, a heat pump, a charging pile, a smart socket, a lighting device, a storage water heater, or other similar devices.

[0023] To better understand the adaptive control method for equipment provided in this application, the specific structure of the control equipment is described below. (See reference...) Figure 2 The diagram shown is a structural schematic of a control device 10 provided in an embodiment of this application. Figure 2 As shown, when the control device 10 is an independent device, the control device 10 includes, but is not limited to, a communication module 101, a memory 102, a processor 103, an input / output (I / O) interface 104, and a bus 105. The processor 103 is coupled to the communication module 101, the memory 102, and the input / output interface 104 via the bus 105.

[0024] In some embodiments of this application, the control device 10 can establish a communication connection with the controlled device 20 through the communication module 101.

[0025] The memory 102 may include one or more random access memory (RAM) and one or more non-volatile memory (NVM). The RAM can be directly read and written by the processor 103, and can be used to store executable programs (such as machine instructions) of the operating system or other running programs, as well as user and application data.

[0026] Random access memory can include static random-access memory (SRAM), dynamic random-access memory (DRAM), synchronous dynamic random-access memory (SDRAM), double data rate synchronous dynamic random-access memory (DDR SDRAM), etc.

[0027] Non-volatile memory can also store executable programs and user and application data, and can be pre-loaded into random access memory for direct reading and writing by the processor 103. Non-volatile memory can include disk storage devices and flash memory.

[0028] Memory 102 is used to store one or more computer programs. The one or more computer programs are configured to be executed by processor 103. The one or more computer programs include multiple instructions.

[0029] In other embodiments, the control device 10 also includes an external memory interface for connecting to an external memory to expand the storage capacity of the control device 10.

[0030] Processor 103 may include one or more processing units, such as an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural network processing unit (NPU). These different processing units may be independent devices or integrated into one or more processors.

[0031] Processor 103 provides computing and control capabilities.

[0032] I / O interface 104 is used to provide a channel for user input or output. For example, I / O interface 104 can be used to connect various input and output devices, such as mouse, keyboard, touch screen, etc., so that users can enter information or visualize information.

[0033] Bus 105 is used at least to provide a channel for communication between the communication module 101, memory 102, processor 103, and input / output interface 104 in the control device 10.

[0034] It is understood that the structure illustrated in the embodiments of this application does not constitute a specific limitation on the control device 10. In other embodiments of this application, the control device 10 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0035] In other embodiments, when the control device 10 is integrated with a device in the energy management system 1, the control device 10 can be a controller in that related device.

[0036] Figure 3 This is a flowchart of a device adaptive control method provided in an embodiment of this application, as shown below. Figure 3 As shown, the adaptive control method is applied to the control equipment of an energy storage system. Depending on different requirements, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0037] Step S301: In response to the user's target input operation, at least two energy management targets corresponding to the energy management system are determined.

[0038] In this embodiment of the application, the user can interact with the energy management system through target input operations and convey energy management intentions to the energy management system.

[0039] In some implementations, the target input operation includes: a user selecting at least one target template from a plurality of preset energy management templates via a graphical user interface, and filling in the parameter items in the target template to form a structured energy management target. The target template can be a text-based energy management template or a voice-based energy management template.

[0040] For example, the energy management template includes "Electricity Cost Control", "Green and Low-Carbon", "Safe Backup Power", and "Equipment Maintenance". When a user selects the "Electricity Cost Control" template and enters the value "300" in the associated parameter field, the energy management target "Electricity cost this month shall not exceed 300 yuan" is generated.

[0041] In some implementations, the target input operation includes: a user inputting a statement containing at least two energy management dimensions in natural language via a voice input device or text input box; the energy management system then parses the statement using a natural language processing module to extract structured energy management targets. For example, refer to... Figure 4 The diagram shown illustrates an energy management template. When the target template is a text-based energy management template, users can fill in parameters based on the options in the user interface. For example, in the option interface for electricity cost energy goals, users can enter the target cost for this month's electricity bill; or in the voice-based energy management template, users can obtain the corresponding energy management goals through voice. This is not limited here.

[0042] For example, if a user inputs the voice command "Don't let the electricity bill exceed 400 yuan this month, and use more of your own solar energy," the natural language processing module will parse it and determine the energy management goals as "This month's electricity bill is less than or equal to 400 yuan" and "Maximize the proportion of self-consumption of photovoltaic green electricity."

[0043] In some implementations, a mapping relationship exists between the preset template and natural language input. The user's fuzzy expression via natural language input can be matched to the corresponding template and parameter completion performed. For example, if the user inputs "save electricity and ensure power supply at night," the control device matches it to a combined template of "electricity cost control" and "safe backup power," and uses default parameters or initiates secondary confirmation to complete the specific numerical constraints, ultimately generating a clear set of energy management objectives. These energy management objectives refer to quantifiable or assessable indicators expected to be achieved in the operation of the energy management system; there are at least two of these objectives, covering different energy management dimensions.

[0044] In some implementations, the energy management objective includes specific numerical constraints, which take the form of setting an upper limit, lower limit, or range for a certain energy indicator. For example, "Total electricity expenditure this month is less than or equal to 500 yuan", "The state of charge (SOC) of the energy storage system is always greater than or equal to 25%", and "Carbon emissions are reduced by 15% compared to last month".

[0045] In some implementations, the energy management objective includes a proportion-optimization objective, which takes the form of maximizing or minimizing a certain proportion of energy. For example, "maximizing the proportion of photovoltaic power generation for self-consumption" or "minimizing the proportion of electricity purchased from the grid".

[0046] In some implementations, the energy management objective includes a fuzzy balance objective, which takes the form of a comprehensive trade-off requirement for at least two potentially conflicting dimensions, without requiring specific numerical values ​​from the user. Examples include "achieving a good balance between electricity costs and energy storage battery life" or "balancing electricity economy and environmental friendliness." For such objectives, the control device will assign initial quantitative weights and constraints based on historical data or default rules.

[0047] In some implementations, the control device can determine at least two energy management objectives corresponding to the energy management system in response to the same target input operation from the user. For example, the user can simultaneously determine multiple energy management objectives covering multiple dimensions such as electricity cost, green electricity ratio, and backup power security through a single template selection (e.g., checking "green and low carbon" and setting "30% backup power") or a natural language command (e.g., "Electricity bill not exceeding 300, use more green electricity, and reserve backup power").

[0048] In this embodiment, users can set multi-dimensional energy management goals that may be interconnected or conflicting, through template selection or natural language description that aligns with everyday thinking. The control device automatically translates the user's high-level intent into structured target parameters (including target type, constraints, priority cues, etc.) that can be used for subsequent optimization calculations using template parsing or Natural Language Processing (NLP) technology. This encapsulates the complex multi-objective setting process into a simple "one-click" interaction, greatly reducing the system's usage threshold and laying a foundation for clear user intent for subsequent automated and intelligent optimization control. For example, an elderly user only needs to input "Housekeeper, help me save on electricity bills, don't affect evening electricity usage" through the control device. The control device can then automatically parse and determine two core energy management goals: "minimize electricity costs" and "maintain the reliability of power supply for necessary evening loads (implying backup power goals)," initiating adaptive collaborative optimization of all home appliances.

[0049] Step S302: Construct a multi-objective function based on the energy management objective.

[0050] In some embodiments of this application, the multi-objective function is a model used to mathematically describe multiple objectives that need to be optimized simultaneously and their possible conflicting relationships. The multi-objective function can be a vector containing multiple sub-objective functions. For example, the multi-objective function is F(x), where F(x) = [f_cost(x), f_green(x), f_B...]. [x], where f_cost(x) represents the electricity cost function, f_green(x) represents the green electricity usage ratio function, and f_B (x) represents the battery life function.

[0051] In this embodiment, constructing a multi-objective function based on the energy management objective requires transforming the user's verbal description into mathematical constraints. For example, the user objective "electricity cost not exceeding 300 yuan" is transformed into the inequality constraint f_cost(x) ≤ 300. The user objective "maximizing green electricity" is transformed into finding the maximum value of f_green(x). Physical constraints can also be introduced during the construction of this multi-objective function. For example, the upper limit of the charging and discharging power of energy storage devices, the maximum output power of photovoltaic inverters, and the total household load demand, etc., together constitute the feasible solution space of the optimization problem. For example, for the dual objectives of "electricity cost ≤ 500 yuan and green electricity ≥ 60%" input by the user, the multi-objective function constructed by the control device will simultaneously minimize the electricity cost and maximize the green electricity ratio, and under the condition of satisfying the physical constraints of energy storage charging and discharging, find a series of possible optimal compromise solutions (e.g., Pareto optimal solutions).

[0052] Step S303: Determine the dynamic weights corresponding to each energy management objective based on the operating status data of the controlled equipment in the energy management system.

[0053] In the embodiments of this application, in an energy management system, there are often inherent conflicts and trade-offs among multiple energy management objectives (such as reducing electricity costs, increasing the proportion of green electricity, and ensuring backup power security) determined in response to user input operations. For example, to pursue the lowest electricity cost, the system may tend to use energy storage batteries extensively during peak electricity price periods, but this will accelerate battery aging and undermine the energy objective of "extending energy storage lifespan." Similarly, to maximize the use of photovoltaic green electricity, the system may want to absorb as much photovoltaic output as possible, but if there are consecutive cloudy or rainy days, over-reliance on photovoltaics will lead to insufficient backup power, jeopardizing the objective of "ensuring backup power security." Traditional multi-objective optimization schemes typically assign fixed weights to each energy management objective, thereby performing static trade-offs between at least two energy management objectives. However, the operating environment of the energy system (such as sunlight, electricity price, and load demand) and equipment status (such as energy storage battery capacity and health) are dynamically changing. A reasonable weight allocation at one time may completely fail at another time, or even lead to system operational risks. For example, when the energy storage battery has sufficient power, the weight of the "reducing electricity costs" target can be appropriately increased, allowing the battery to discharge more to save on electricity costs; however, when the battery power drops to a safe boundary, the weight of the "ensuring backup power security" target can be dynamically increased immediately, prioritizing the reliability of the system power supply, even if this means temporarily sacrificing some economic efficiency.

[0054] The dynamic weights corresponding to each energy management objective are determined based on the operational status data of the controlled equipment in the energy management system. This is not simply assigning a fixed importance coefficient to each energy management objective, but rather determining the weights based on real-time collected operational status data of the controlled equipment. This allows multi-objective optimization to move beyond static preset rules and intelligently and adaptively adjust the priority of each energy management objective based on real-time fluctuations in photovoltaic power generation, the current state of charge (SOC) of energy storage batteries, instantaneous changes in grid electricity prices, and the actual load of each electrical device.

[0055] In this embodiment of the application, the operating status data refers to a set of parameters collected in real time from the controlled equipment of the energy management system, reflecting its current operating conditions and performance.

[0056] In some embodiments, the operating status data includes the energy storage device's state of charge (SOC), state of health (SOH), charge / discharge efficiency, and temperature. For example, the current energy storage SOC is 45% and the battery temperature is 32°C, obtained in real time through the Battery Management System (BMS).

[0057] In some embodiments, the operating status data includes the real-time output power of the photovoltaic power generation system, the short-term predicted output power, and the prediction error. For example, the photovoltaic inverter reports a current power generation of 2.5kW, while the weather forecast interface provides a prediction of the solar irradiance for the next hour, and the calculated prediction error is 20%.

[0058] In some embodiments, the operational status data includes real-time power drawn from the grid, the time-of-use electricity price for the current period, and future electricity price plans. For example, a smart meter monitors that the current household power drawn from the grid is 1.2 kW, and the electricity price table indicates that it is currently during peak electricity price hours (1.2 yuan / kWh).

[0059] In some embodiments, the operational status data includes the current operating power, setting mode, and working status of each controlled device (such as an air conditioner, an electric vehicle charging station, and a water heater). For example, the smart socket reports that the current cooling power of the living room air conditioner is 1.5kW, and the electric vehicle charging station is in standby mode.

[0060] In this embodiment, the dynamic weights calculated in real-time based on the operating status data of the controlled equipment can represent the relative importance of each energy management objective in the current system state. These coefficients adaptively adjust with time and state changes. Specifically, the method for determining the dynamic weights corresponding to each energy management objective based on the operating status data of the energy equipment is described below. Figure 5 As shown.

[0061] Step S304: Perform multi-objective optimization based on multi-objective functions and dynamic weights to generate a control strategy.

[0062] In this embodiment, multi-objective optimization based on a multi-objective function and dynamic weights is the process of finding a set of non-dominated solutions (Pareto optimal solutions) in a feasible solution space with conflicting objectives. This application can employ an improved Dynamic Weight-Adaptive NSGA-II (NSGA-II) as the optimization engine. This algorithm iteratively generates an optimization strategy population by simulating selection, crossover, and mutation in biological evolution, and utilizes non-dominated sorting and crowding calculation to ensure the quality and diversity of solutions.

[0063] In this embodiment, NSGA-II is a classic algorithm for multi-objective scheme optimization, specifically including: population initialization, evolutionary iteration, termination judgment, and policy output steps. Population initialization includes: based on the current system state and user objectives, initializing a population P_t consisting of N individuals (i.e., candidate control policies). Each individual, through chromosome encoding, represents the control sequence for each controlled device (such as energy storage charging and discharging power, air conditioning set temperature adjustment, charging pile start and stop time, etc.) within a future scheduling cycle.

[0064] The evolutionary iteration steps include: a. Offspring generation: Using selection, crossover, and mutation operators, generate an offspring population Q_t of the same size from the current parent population P_t; b. Merge populations: Merge the parent population P_t with the offspring population Q_t to form a merged population R_t of size 2N; c. Dynamic weight loading: Load the dynamic weight vectors W(t) = [w_1(t), w_2(t), ..., w_m(t)] of each objective calculated at the current time from the policy generation domain; d. Weighted Fitness Assessment: For each individual in R_t, based on its encoded control sequence, simulate the system operation under future scenarios and calculate its performance on various original energy targets (electricity cost f_cost, green electricity ratio f_green, energy storage lifetime loss f_B). The function value F(x) on the objective function is given by [f_1(x), f_2(x), ..., f_m(x)]. By using dynamic weights W(t) to weight the original objective function value, a weighted objective value F'(x) for ranking is obtained: F'(x) = [w_1(t)*f_1(x), w_2(t)*f_2(x), ..., w_m(t)*f_m(x)]. For maximizing objectives (such as green electricity ratio), it is usually first converted to a minimization form. e. Non-dominated sorting: Based on the weighted objective value F'(x), perform a fast non-dominated sorting on all individuals in the merged population R_t, dividing them into different non-dominated layers F1, F2, ... . Layer F1 contains all Pareto optimal solutions that are not dominated by any other individual. f. Crowding Distance Calculation: For individuals within the same non-dominated layer, calculate their crowding distance in the weighted target space. The calculation of crowding distance takes into account dynamic weights to ensure that the solution set has better distribution in the target dimension that the user is currently more concerned with; g. Environmental selection: Following the order of superiority to inferiority in the non-dominant layer (F1 takes precedence over F2) and the order of crowding distance within the layer from largest to smallest, N individuals are selected sequentially from R_t to form the parent population P_{t+1} of the new generation. This process ensures the preservation of elites and population diversity.

[0065] The termination judgment steps include: if the preset maximum number of generations Gen_max is reached, or the population convergence index meets the requirements, then terminate the iteration; otherwise, let t = t + 1 and return to step a.

[0066] Policy output: From the first non-dominated layer (F1) of the final population P_{final}, select the optimal individual based on the weighted comprehensive score Score(x) = Σ [w_i(t) * f_i'(x)] (f_i'(x) is the normalized value), and decode it into a specific, executable control policy.

[0067] In this embodiment, a dynamic weight W(t) is directly applied to the evaluation phase of each individual in each generation of the population. This makes the criterion for determining "non-dominant" relationships time-varying. For example, when the dynamic weight w_security(t) increases due to a decrease in SOC, a strategy that performs slightly better in maintaining SOC but is less economical may have a better weighted evaluation result than a strategy that is highly economical but depletes SOC. Consequently, the former will have a higher rank in the non-dominant ranking and a greater probability of being retained and evolving. This drives the algorithm to automatically search for the optimal compromise solution that meets the current priority.

[0068] In some implementations, a resource penalty mechanism may also be introduced. This mechanism is based on real-time device utilization calculations. When a device (such as a charging station) is overloaded, a penalty coefficient is applied, increasing its scheduling cost. This automatically reduces its priority in multi-objective optimization, achieving load balancing. For a detailed description, please refer to [link to relevant documentation]. Figure 6 The flowchart shown.

[0069] Step S305: Control the operation of the controlled equipment according to the control strategy.

[0070] In this embodiment, the operation of controlled devices can be controlled according to a control strategy. For example, the controlled device may include an energy storage converter, which executes an energy storage scheduling strategy by adjusting its operating mode (charging / discharging / standby) and power level. The controlled device may include a smart switch or smart socket for remotely switching on or off or adjusting the power supply of loads such as air conditioners, water heaters, and charging piles. The controlled device may include a photovoltaic inverter, which affects the consumption method of photovoltaic power generation (all for self-consumption, part for grid connection) by adjusting its operating point. The controlled device may also include a thermostat in a smart home system, which indirectly controls the load by adjusting the set temperature of air conditioners and underfloor heating.

[0071] In summary, the device adaptive control method provided in this application determines at least two energy management objectives based on the user's target input operation; constructs a multi-objective function based on the energy management objectives; determines the dynamic weights corresponding to each energy management objective based on the operating status data of the controlled devices in the energy management system; performs multi-objective optimization based on the multi-objective function and the dynamic weights to generate a control strategy; and controls the operation of the controlled devices according to the control strategy. Through a closed-loop control strategy based on one-click user input of multiple objectives, reverse derivation, adaptive adjustment of dynamic weights, and multi-device collaboration, it achieves true "one-click" multi-objective energy management, effectively reducing the user's operational threshold and improving the overall achievement rate of multiple objectives and the system's adaptive capability.

[0072] Understandably, after executing step S305, the device adaptive control method further includes: continuously monitoring the operating status data of the controlled device; updating the dynamic weight based on the current operating status data when the change in the operating status data meets a preset triggering condition; and updating the control strategy based on the updated dynamic weight.

[0073] In this embodiment, the operating status data of the controlled devices can be monitored at preset intervals. For example, every 5 minutes, the operating status data is collected, aggregated, and updated in real time through sensors, smart meters, and communication modules deployed on each controlled device and key node. When changes in the operating status data meet preset trigger conditions, the dynamic weights are updated based on the current operating status data, and the control strategy is updated based on the updated dynamic weights.

[0074] In this embodiment, the triggering conditions are a set of pre-defined thresholds or event rules used to determine significant changes in the energy management operating environment or internal state. Once a triggering condition is detected and met, it is determined that the currently executed control strategy may no longer be optimal or face risks, thereby immediately initiating a strategy recalculation process. The triggering conditions include: real-time grid price fluctuations exceeding a first threshold (e.g., 20%); photovoltaic prediction errors exceeding a second threshold (e.g., 30%); energy storage system state of charge falling below a third threshold (e.g., 20%); and the user temporarily modifying or adding energy management objectives via an app or voice.

[0075] In this embodiment, the price fluctuations are caused by drastic changes in the spot electricity market price or by receiving an emergency demand response signal from the power grid. These sudden price changes directly impact the optimization basis for the "reducing electricity costs" objective. Therefore, when the real-time electricity price fluctuation exceeds a first threshold, the dynamic weights are updated based on the current operating status data, and the control strategy is updated based on the updated dynamic weights.

[0076] In this embodiment, rapid weather changes (such as sudden heavy rain or rapid cloud movement) can affect the actual power generation of the photovoltaic system, causing it to deviate significantly from the power generation forecast based on historical data and short-term weather forecasts. Increased photovoltaic forecasting error means a decrease in the reliability of strategies relying on existing forecasts. Therefore, when the photovoltaic forecasting error exceeds a second threshold, the dynamic weights are updated based on the current operating status data, and the control strategy is updated based on the updated dynamic weights.

[0077] In this embodiment of the application, when the state of charge of the energy storage system is lower than the third threshold, the system's backup energy reserves are insufficient, and it is necessary to increase the priority of the "ensuring backup power security" objective, limit further discharge of the energy storage, and possibly adjust other loads.

[0078] In this embodiment of the application, when a user temporarily modifies or adds energy management objectives via an APP or voice, a multi-objective function can be constructed based on the updated energy management objectives, thereby updating the dynamic weights based on the current operating status data, and updating the control strategy based on the updated dynamic weights.

[0079] Please see Figure 5In some embodiments, step S303 includes the following sub-steps: Step S501: Based on the operating status data, determine the device status matrix associated with each of the at least two energy management objectives.

[0080] In this embodiment of the application, the device state matrix S i (t) is a vector or set of state parameters of key real-time controlled equipment closely related to energy management objectives. Operating parameters of various devices such as photovoltaic power generation equipment, energy storage equipment, and loads are acquired periodically or event-driven through a real-time data acquisition network, and the device state matrix is ​​constructed based on predefined mapping relationships.

[0081] In one embodiment, the device state matrix can be described by the following formula (1): (1), in, This represents the state of charge of the stored energy at time t; This represents the photovoltaic prediction error; This indicates the energy storage charging and discharging efficiency.

[0082] Step S502: Obtain the dynamic weight of each energy management objective based on the equipment status matrix.

[0083] In this embodiment, after determining the device state matrix, it is necessary to determine the dynamic weight of each corresponding energy management objective. Determining this dynamic weight is a process of quantifying the real-time state of the controlled device into an objective priority adjustment signal. This allows the establishment of a mathematical mapping model that can respond to changes in device state and has clear physical meaning and configurability. When the state parameters of the controlled device deviate from their ideal or safe range, the system can automatically and reasonably adjust the optimization weights of the relevant objectives, thereby achieving adaptive adjustment of the optimization strategy. Specifically, obtaining the dynamic weight of each energy management objective based on the device state matrix includes: obtaining the critical lower limit and critical upper limit corresponding to the device state matrix; obtaining a normalized state value based on the device state matrix, the critical lower limit, and the critical upper limit; obtaining a state coefficient based on the normalized state value and the adjustment coefficient; and obtaining the dynamic weight based on the state coefficient and the initial weight corresponding to each energy management objective.

[0084] In this embodiment of the application, for the state matrix S i For each key state parameter (such as SOC(t)) in (t), the system presets or determines its critical lower limit value through learning. and critical upper limit value These critical values ​​define the reasonable or safe operating range for this parameter. For example, for energy storage SOC, It can be set to guarantee 20% of the minimum backup power requirement. It can be set to 95% to prevent overcharging. For the photovoltaic prediction error ΔPV, Possibly 0% (ideal prediction) This may be the maximum tolerance of the system, such as 40%.

[0085] In this embodiment, a normalized state value is obtained based on the device state matrix, the lower critical limit, and the upper critical limit. The normalized state value is calculated using the following formula (2). , (2).

[0086] The above formulas can map original state parameters with different units and dimensions (such as SOC expressed as a percentage or power expressed in kilowatts) to the interval [0, 1]. Normalization allows the influence of different state parameters on the weights to be compared on the same scale.

[0087] In this embodiment, an adjustment coefficient is introduced to precisely control the response characteristics of the weights to changes in device state. This coefficient is a configurable real number, and its sign determines the direction in which the weight changes with the state (a positive sign indicates that the weight changes with the state). An increase in the sign indicates an increase in the weight, while a negative sign indicates an increase in the weight (the opposite is true for negative signs). The absolute value of this coefficient determines the sensitivity of the weight to state changes. In the embodiments of this application, the state coefficient... A specific response curve can be simulated using a nonlinear function (such as an exponential function). The state coefficient can be calculated using the following formula (3): (3).

[0088] In this embodiment of the application, each energy management objective corresponds to a benchmark weight. Based on the state coefficients and the initial weights corresponding to each energy management objective, dynamic weights can be obtained, which can be calculated using the following formula (4): (4).

[0089] Among them, the benchmark weight It can be set by the user or by the system default, representing the importance of the target under typical conditions.

[0090] Through the above steps S501 and S502, this embodiment of the application implements a data-driven, state-aware, and parameter-configurable dynamic weight determination mechanism. This mechanism deeply couples the physical reality of equipment operation (state matrix) with the decision preferences of multi-objective optimization (dynamic weights), enabling the optimization behavior of the energy management system to closely integrate with real-time operating conditions and make adaptive intelligent adjustments.

[0091] Figure 6 This is a flowchart of a device adaptive control method provided in another embodiment of this application, such as... Figure 6 As shown, the adaptive control method is applied to the control equipment of an energy storage system. Depending on different requirements, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0092] Step S601: In response to the user's target input operation, determine at least two energy management targets corresponding to the energy management system.

[0093] Step S602: Construct a multi-objective function based on energy management objectives.

[0094] Step S603: Determine the dynamic weights corresponding to each energy management objective based on the operating status data of the controlled equipment in the energy management system.

[0095] For details of steps S601-S603, please refer to the above text. Figure 3 The detailed descriptions of steps S301-S303 are not repeated here.

[0096] Step S604: Based on the operating status data, identify at least one target device among multiple controlled devices that has an overload risk, and calculate the resource penalty coefficient based on the real-time utilization rate of the target device.

[0097] In this application embodiment, to prevent strategy failure, equipment damage, or user experience degradation caused by excessive competition for equipment resources or local overload while achieving multi-objective collaborative optimization, the device adaptive control method provided in this application also introduces a resource penalty coefficient mechanism. Controlled devices in the energy management system (such as charging piles, energy storage converters, high-power air conditioners, etc.) all have physical capacity or safe operating limits. During multi-device collaborative scheduling, if the optimization algorithm only pursues the optimal global objective while ignoring the real-time load pressure of individual controlled devices, some controlled devices may continuously operate at high loads during specific periods (such as peak electricity consumption), increasing the risk of failure and potentially affecting the achievement of other objectives due to resource contention (for example, charging piles operating at full power may excessively consume energy storage capacity, endangering backup power security). Therefore, this application, by real-time sensing of device load status, marks devices approaching or exceeding safety thresholds and adjusts their scheduling priority in the optimization model through a quantified penalty coefficient. This automatically and smoothly achieves load balancing and peak-shifting scheduling in subsequent optimization processes, ensuring that the system pursues multi-objective optimization while also considering equipment safety and operational reliability.

[0098] Specifically, based on operational status data, at least one target device among multiple controlled devices is identified as having an overload risk. The resource penalty coefficient is calculated based on the real-time utilization rate of this target device, including: calculating a first value based on a first difference between the real-time utilization rate and a utilization rate safety threshold, and a second difference between a preset penalty benchmark value and the utilization rate safety threshold, wherein the utilization rate safety threshold is greater than or equal to 0 and less than 1; calculating a second value based on the first value and a penalty correction coefficient; and calculating the resource penalty coefficient based on the penalty benchmark value and the second value.

[0099] In this embodiment of the application, the resource penalty coefficient can be calculated using the following formula (5). : (5).

[0100] in, The real-time utilization rate of the controlled device r; The utilization safety threshold; This is the preset penalty baseline value; This is the penalty correction factor.

[0101] In some embodiments, when the controlled device is a charging station... Average occupancy rate of charging stations between 18:00 and 22:00; utilization rate safety threshold. The value can be set according to the characteristics of the controlled device; for example, 0.8 is used for charging piles. The penalty correction coefficient can be an expert's experience value, with a default of 0.75, to avoid excessive adjustment of the strategy and affecting the user experience.

[0102] For example, in one implementation scenario, suppose the home energy management system monitors the real-time utilization rate of electric vehicle charging stations during the evening peak hours (6:00 PM - 7:00 PM). The system's preset utilization safety threshold is 0.92 (meaning the average power utilization rate reaches 92% of the rated power during this period). The base penalty value is 1.0, and the penalty correction factor is 0.75. The calculation process for the corresponding resource penalty factor is as follows: Calculate the first difference d1 based on the real-time utilization rate and the utilization rate safety threshold. - = 0.92 - 0.8 = 0.12; Based on the preset penalty benchmark value and the second difference d2 between the utilization safety threshold, d2 = 1 - = 1 - 0.8 = 0.2. Based on the first and second differences, the first value A1 = d1 / d2 = 0.12 / 0.2 = 0.6 is calculated. Based on the first value and the penalty correction coefficient, the second value A2 = 0.6 * 0.75 = 0.45 is calculated. Based on the penalty baseline value and the second value, the resource penalty coefficient is calculated. = 1.0 + 1.2 * 0.45 = 1.54. This penalty coefficient means that the scheduling cost of this charging pile in the current time period will be increased by about 1.54 times, thus significantly reducing its probability of being prioritized in multi-objective optimization and guiding charging tasks to migrate to other time periods.

[0103] Step S605: Perform multi-objective optimization using multi-objective functions, dynamic weights, and resource penalty coefficients to generate a control strategy.

[0104] In the embodiments of this application, existing multi-objective optimization methods typically employ strict boundary constraints when dealing with equipment operation constraints, requiring that the operating parameters of the controlled equipment remain completely within a preset safety range. This approach may render the optimization problem unsolvable due to overly stringent constraints when facing dynamic and uncertain energy scenarios.

[0105] The adaptive control method for equipment provided in this application employs a dynamically adjusted constraint processing mechanism. By quantifying the risk of equipment overload into an adjustable term in the objective function through a resource penalty coefficient, the optimization algorithm can coordinate the trade-offs between different objectives by adjusting the numerical impact of this term when the equipment state approaches or slightly exceeds the safety threshold. This mechanism enables the system to establish a dynamic balance between ensuring safe equipment operation and meeting the user-defined multi-dimensional energy management objectives, generating feasible optimization control strategies that fully consider the actual load state of the equipment under most operating conditions. Simultaneously, a linkage is established between the resource penalty coefficient and the dynamic weight, allowing constraint processing to adaptively adjust according to the priority of the overall system optimization objectives. For example, when the objective of "ensuring backup power safety" is assigned a very high dynamic weight, even if the charging pile utilization rate is high, its corresponding resource penalty coefficient may be appropriately adjusted to avoid excessive restrictions on charging functions affecting users' emergency travel needs.

[0106] Specifically, the control strategy is generated by performing multi-objective optimization using a multi-objective function, dynamic weights, and resource penalty coefficients. This includes: adjusting the scheduling priority of each controlled device based on the resource penalty coefficients; performing multi-objective optimization using a non-dominated sorting genetic algorithm based on the multi-objective function and the scheduling priority to obtain a Pareto optimal solution set; scoring each Pareto optimal solution in the Pareto optimal solution set according to the dynamic weight of each energy management objective, and determining the Pareto optimal solution with the highest score as the control strategy.

[0107] For example, in one application scenario, the system identifies that the real-time utilization rate of an electric vehicle charging station reaches 0.92 during the evening peak hours (18:00-20:00), exceeding the preset utilization rate safety threshold of 0.8. Therefore, it marks the charging station as a target device and calculates the resource penalty coefficient. The penalty coefficient is 1.3. The system then adjusts the scheduling priority of the charging piles based on this penalty coefficient, for example, adjusting its basic scheduling priority value from 0.8 to 0.8 × 1.3 = 1.04 (the larger the value, the lower the priority and the later it is scheduled). In the subsequent multi-objective optimization process, when generating candidate control strategies, the algorithm tends to schedule the charging tasks of the charging piles during off-peak hours or reduce their charging power. This is because the scheme scheduled during high utilization periods will be at a disadvantage in the selection process due to its lower scheduling priority. After iterative optimization, the algorithm outputs a Pareto optimal solution set, which contains multiple strategy schemes that achieve different balances among objectives such as "reducing electricity costs", "increasing the proportion of green electricity", and "ensuring backup power security". The system then uses the dynamic weights at the current moment (such as electricity cost weight 0.5, green electricity weight 0.3, and backup power weight 0.2) to perform a weighted comprehensive score on each strategy in this solution set, and finally selects the strategy with the best comprehensive score as the final control strategy for execution. This strategy likely reflects the staggered scheduling of charging tasks, thereby effectively alleviating grid pressure during peak hours and reducing potential risks caused by overload of controlled equipment without excessively affecting users' charging needs.

[0108] Step S606: Control the operation of the controlled equipment according to the control strategy.

[0109] In this embodiment of the application, the details of step S606 can be found above. Figure 3 The detailed description of step S305 is already provided and will not be repeated here.

[0110] To facilitate understanding of the device adaptive control method provided in the above embodiments, the following examples are provided for illustration.

[0111] Example 1: Electricity cost control and green electricity utilization scenario.

[0112] In this embodiment, in response to the user's target input operation, at least two energy management objectives corresponding to the energy management system are determined, including: monthly total electricity expenditure not exceeding 500 yuan, and the self-consumption ratio of photovoltaic green electricity not less than 70%. The control device, in response to the target input operation, determines the above two energy management objectives and constructs a multi-objective function based on these objectives, including an electricity cost sub-function and a green electricity ratio sub-function. In the initial stage of the energy management system's operation, initial weights are assigned to each objective, with the initial weight for the electricity cost objective being 0.6 and the initial weight for the green electricity ratio objective being 0.4. During system operation, the operating status data of the controlled equipment is continuously monitored. On the 10th day of operation, monitoring data shows that the health status parameter of the energy storage system has decreased from the initial 95% to 83%. Based on this status data, the energy management system automatically calculates and incorporates the energy management objective related to "extending energy storage lifespan" into the optimization considerations, assigning it an initial weight of 0.2. Simultaneously, the original weights are normalized and adjusted, with the adjusted weights for the electricity cost objective being 0.5, the green electricity ratio objective being 0.3, and the energy storage lifespan objective being 0.2.

[0113] On the 15th day of operation, the energy management system detected a sudden change in the external grid electricity price, with a real-time price increase of 25%. This event met the preset trigger conditions, and the energy management system immediately initiated the strategy update process. During the update process, the energy management system reused the first five non-dominated solutions saved in the historical optimization iterations as part of the initial population to accelerate convergence. Simultaneously, based on the current high electricity price, the dynamic weight of the electricity cost objective was increased to 0.7. A new control strategy was generated through rapid optimization using the DWA-NSGA-II algorithm. This strategy mainly includes: increasing the maximum allowable depth of discharge of the energy storage system from the original 70% to 85% to release more stored energy during peak electricity price periods; and adjusting the air conditioner temperature setpoint to 26℃ during peak electricity consumption periods to reduce the load power during these periods. According to monthly operation statistics, the actual electricity cost was 482 yuan, which is 103.7% of the target of 500 yuan; the actual green electricity self-consumption rate was 72%, which is 102.9% of the target of 70%; in addition, through optimized scheduling, the average monthly charge and discharge cycle of the energy storage system was reduced by 18 times, and its service life is estimated to be extended by about 15%.

[0114] Example 2: Collaborative scenario of safe backup power and low-carbon operation.

[0115] In this embodiment, the user-defined energy management objectives are: maintaining a minimum energy storage backup capacity of 5 kWh and reducing monthly carbon emissions by 20%. The energy management system executes the device adaptive control method provided in this application, and the specific process is as follows: After determining the aforementioned objectives, the control equipment constructs a corresponding multi-objective function. During the operation of the energy management system, three consecutive days of rainy weather occurred. The energy management system monitored that the error between the actual power generation of the photovoltaic system and the predicted value reached 22%, exceeding the preset threshold of 15%. Based on this equipment status data, the system dynamically adjusts and optimizes the weights. Due to the increased uncertainty in photovoltaic output, the weight of "increasing the proportion of green electricity" (related to the low-carbon objective) is reduced to 0.2 (initially 0.3); simultaneously, to prioritize ensuring energy supply security, the weight of the objective of "ensuring backup power capacity" is increased to 0.8.

[0116] Based on the updated dynamic weights, the system derives a new dynamic control strategy through an optimization algorithm. This strategy mainly includes two adjustments: first, limiting the discharge depth of the energy storage system to no more than 60% (the original strategy was 80%) to ensure that the backup power capacity is always maintained above a safe level; second, adjusting the operating hours of high-power heating equipment such as heat pumps to the off-peak hours of the power grid (00:00-08:00) to reduce power extraction from the grid during peak hours, thereby reducing carbon emissions.

[0117] By implementing the dynamically adjusted strategy described above, the system successfully maintained a backup power capacity retention rate of 98.3% throughout the entire operating cycle. Simulation calculations show that without this strategy recalculation, the backup power capacity retention rate is expected to drop to 90%. Simultaneously, monthly carbon emissions were reduced by 21.5%, exceeding the user-set 20% emission reduction target. The entire optimization and adjustment process was completed automatically by the system; the user was unaware of the strategy switch, resulting in a seamless experience.

[0118] Example 3: Multi-objective fuzzy balance and long-term adaptive scenario.

[0119] In this embodiment, the user inputs a fuzzy target of "balancing electricity costs and energy storage lifespan" via natural language, without setting a specific value. The energy management system executes the device adaptive control method provided in this application, and the specific process is as follows: After parsing the user's intent, the system assigns default initial weights of 0.5 to both energy management objectives: "reducing electricity costs" and "extending energy storage lifespan." The system then enters a long-term operation and self-learning state.

[0120] The system first performs automatic weight allocation and learning: After running continuously for one month, the system analyzes historical operating data and finds that when the daily charge-discharge frequency of the energy storage system exceeds 3 times, its cumulative impact on battery life increases significantly. Based on this learning result, the system automatically adjusts the weight allocation rules: In subsequent operation, if the daily charge-discharge frequency is detected to exceed 3 times, the weight of the "extending energy storage life" target is automatically increased by 0.1.

[0121] Secondly, the system performs resource congestion adjustment: On a certain operating day, the system detected that the utilization rate of electric vehicle charging piles remained consistently high during peak electricity price periods (18:00-22:00), indicating resource competition. Based on the real-time utilization rate of the charging piles, the system calculated a resource penalty coefficient of 1.3. In subsequent optimizations, this penalty coefficient was used to adjust the scheduling priority of charging piles, prompting the system to schedule charging tasks more frequently during peak photovoltaic output periods (10:00-15:00), thereby achieving peak shifting and improving green electricity consumption.

[0122] After long-term operation (e.g., one year), the long-term effects of this solution include: the total annual charge-discharge cycle count of the energy storage system decreased from 365 in the baseline scenario to 292, a reduction of approximately 20%, effectively mitigating battery aging; simultaneously, through optimized scheduling, the total annual electricity cost was reduced by 420 yuan compared to the baseline scenario, a decrease of approximately 8%. The entire optimization process was completed autonomously by the system based on real-time status and learned rules, requiring no manual configuration or intervention from the user throughout the year.

[0123] Figure 7 A schematic structural block diagram of the device adaptive control apparatus 70 provided in the fourth aspect of the present application is shown, as follows: Figure 7 As shown in this embodiment, the device adaptive control device 70 can be divided into multiple functional modules according to the functions it performs. These functional modules may include: a determination module 701, a construction module 702, a generation module 703, and a control module 704.

[0124] The determination module 701 is used to determine at least two energy management objectives corresponding to the energy management system in response to the user's target input operation; the construction module 702 is used to construct a multi-objective function based on the energy management objectives; the determination module 701 is also used to determine the dynamic weights corresponding to each energy management objective based on the operating status data of the controlled equipment in the energy management system; the generation module 703 is used to perform multi-objective optimization based on the multi-objective function and the dynamic weights to generate a control strategy; and the control module 704 is used to control the operation of the controlled equipment according to the control strategy.

[0125] In some embodiments, the determining module 701 is further configured to determine, based on the operating status data, an equipment status matrix associated with each of the at least two energy management objectives; and to obtain the dynamic weight of each energy management objective based on the equipment status matrix.

[0126] In some embodiments, the determining module 701 is further configured to obtain the critical lower limit value and critical upper limit value corresponding to the device state matrix; obtain a normalized state value based on the device state matrix, the critical lower limit value and the critical upper limit value; obtain a state coefficient based on the normalized state value and the adjustment coefficient; and obtain the dynamic weight based on the state coefficient and the initial weight corresponding to each energy management target.

[0127] In some embodiments, the generation module 703 is further configured to identify at least one target device among the plurality of energy devices that has an overload risk based on the operating status data, and calculate a resource penalty coefficient based on the real-time utilization rate of the target device; and perform multi-objective optimization using the multi-objective function, the dynamic weights and the resource penalty coefficient to generate the control strategy.

[0128] In some embodiments, the generation module 703 is further configured to calculate and determine a first value based on a first difference between the real-time utilization rate and the utilization rate safety threshold, and a second difference between a preset penalty benchmark value and the utilization rate safety threshold, wherein the utilization rate safety threshold is greater than or equal to 0 and less than 1; calculate a second value based on the first value and the penalty correction coefficient; and calculate the resource penalty coefficient based on the penalty benchmark value and the second value.

[0129] In some embodiments, the generation module 703 is further configured to adjust the scheduling priority of each of the controlled devices based on the resource penalty coefficient; perform multi-objective optimization using a non-dominated sorting genetic algorithm based on the multi-objective function and the scheduling priority to obtain a Pareto optimal solution set; score each Pareto optimal solution in the Pareto optimal solution set according to the dynamic weight of each of the energy management objectives, and determine the Pareto optimal solution corresponding to the highest score as the control strategy.

[0130] In some embodiments, the control module 704 is further configured to continuously monitor the operating status data of the controlled device; when the change in the operating status data meets a preset triggering condition, update the dynamic weight based on the current operating status data; and update the control strategy based on the updated dynamic weight.

[0131] It should be noted that the information interaction and execution process between the above-mentioned devices / units / modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0132] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0133] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0134] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0135] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0136] In the embodiments provided in this application, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the apparatus / network device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0137] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0138] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An adaptive control method for equipment, applied to control equipment in an energy management system, characterized in that, The method includes: In response to the user's target input operation, at least two energy management targets corresponding to the energy management system are determined; Construct a multi-objective function based on the aforementioned energy management objectives; The dynamic weights corresponding to each energy management objective are determined based on the operating status data of the controlled equipment in the energy management system. Multi-objective optimization is performed based on the multi-objective function and the dynamic weights to generate a control strategy; The controlled equipment is operated according to the control strategy.

2. The adaptive control method for equipment as described in claim 1, characterized in that, The determination of the dynamic weights corresponding to each energy management objective based on the operating status data of the controlled equipment in the energy management system includes: Based on the operational status data, determine the equipment status matrix associated with each of the at least two energy management objectives; The dynamic weight of each energy management objective is obtained based on the device state matrix.

3. The adaptive control method for equipment as described in claim 2, characterized in that, The step of obtaining the dynamic weight of each energy management objective based on the device state matrix includes: Obtain the critical lower limit and critical upper limit values ​​corresponding to the device state matrix; A normalized state value is obtained based on the device state matrix, the lower critical limit, and the upper critical limit; The state coefficients are obtained based on the normalized state values ​​and the adjustment coefficients. The dynamic weights are obtained based on the state coefficients and the initial weights corresponding to each energy management objective.

4. The adaptive control method for equipment as described in claim 1, characterized in that, The method further includes: Based on the operating status data, at least one target device among the multiple energy devices is identified as having an overload risk, and a resource penalty coefficient is calculated based on the real-time utilization rate of the target device. Multi-objective optimization is performed based on the multi-objective function and the dynamic weights to generate a control strategy, including: The control strategy is generated by performing multi-objective optimization using the multi-objective function, the dynamic weights, and the resource penalty coefficient.

5. The adaptive control method for equipment as described in claim 4, characterized in that, The calculation of the resource penalty coefficient based on the real-time utilization rate of the target device includes: Based on the first difference between the real-time utilization rate and the utilization rate safety threshold, and the second difference between the preset penalty benchmark value and the utilization rate safety threshold, a first value is calculated and determined, wherein the utilization rate safety threshold is greater than or equal to 0 and less than 1. The second value is obtained by calculating based on the first value and the penalty correction coefficient; The resource penalty coefficient is obtained by calculating based on the penalty benchmark value and the second value.

6. The adaptive control method for equipment as described in claim 4 or 5, characterized in that, The step of performing multi-objective optimization using the multi-objective function, the dynamic weights, and the resource penalty coefficient to generate the control strategy includes: The scheduling priority of each controlled device is adjusted based on the resource penalty coefficient. Based on the multi-objective function and the scheduling priority, a non-dominated sorting genetic algorithm is used to perform multi-objective optimization to obtain a Pareto optimal solution set. Each Pareto optimal solution in the Pareto optimal solution set is scored according to the dynamic weight of each energy management objective, and the Pareto optimal solution with the highest score is determined as the control strategy.

7. The adaptive control method for equipment as described in claim 1 or 4, characterized in that, The method further includes: Continuously monitor the operating status data of the controlled equipment; When the changes in the operating status data meet the preset triggering conditions, the dynamic weights are updated based on the current operating status data; The control strategy is updated based on the updated dynamic weights.

8. An adaptive control device for equipment, characterized in that, The device includes: The determination module is used to determine at least two energy management objectives corresponding to the energy management system in response to the user's target input operation; A construction module is used to construct a multi-objective function based on the energy management objectives; The determining module is further configured to determine the dynamic weights corresponding to each of the energy management objectives based on the operating status data of the controlled equipment in the energy management system; The generation module is used to perform multi-objective optimization based on the multi-objective function and the dynamic weights, and generate a control strategy. The control module is used to control the operation of the controlled device according to the control strategy.

9. An energy management system, characterized in that, The energy management system includes a control device, which includes: At least one processor; and, A memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the device adaptive control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions that are executed by a processor in an electronic device to implement the device adaptive control method as described in any one of claims 1 to 7.

Images