Campus micro-energy multi-scene scheduling method, device, equipment and medium

Abstract

The invention relates to a campus micro-energy multi-scene scheduling method, apparatus and device, and a medium. The method comprises the steps of obtaining target operation data; performing scene recognition on the target operation data according to a preset multi-scene zero-carbon operation strategy model to obtain a current operation scene; determining a power control target of each functional module in the combined device according to the current operation scene and the target operation data, and generating a corresponding power control instruction sequence; and responding to the power control instruction sequence to control each function module to carry out corresponding adjustment. The method has the effect of improving the campus energy scheduling efficiency.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of distributed energy management, and particularly relates to a campus micro-energy multi-scenario scheduling method, device, equipment and medium. BACKGROUND

[0002] At present, with the in-depth promotion of green low carbon and the acceleration of new power system construction, the proportion of micro-energy systems such as campuses and communities in the power system is increasing, and zero carbon operation is gradually becoming an important target in the process of energy transformation. The campus environment has the characteristics of strong energy use regularity, diverse load types, high safety requirements, and the like. It is necessary to connect photovoltaic power generation, energy storage devices, electric heating, battery replacement facilities and other energy units under the premise of ensuring the reliability of teaching and scientific research and daily life energy use, to realize the local consumption and orderly use of distributed energy, which puts forward higher requirements for the operation safety, economy and user comfort of the micro-energy system.

[0003] The existing operation strategy of the micro-energy system mostly adopts simple rule control based on experience or scheduling around a single operation target. When configuring photovoltaic power generation and energy storage devices, fixed charging and discharging thresholds or fixed operation curves are usually preset, which cannot be dynamically adjusted according to the intermittency of photovoltaic power generation, the random fluctuation of load demand and the differentiated energy use demand under different operation scenarios in the campus, resulting in limited local consumption of photovoltaic power, and a large proportion of dependence on grid power purchase during operation, which is difficult to balance the coordination between green electricity utilization, user comfort and operation cost, and under the working conditions of power grid failure or extreme weather, most of them only perform simple off-grid switching and load shedding, and lack of fine supply protection strategies for key loads.

[0004] The existing technical solutions in the above have the following defects: the existing micro-energy system operation method relies on static rules and local single target scheduling strategies for a long time, and it is difficult to respond to photovoltaic output fluctuation, load change and operation scenario switching in time, and the operation effect is difficult to meet the comprehensive needs of the campus micro-energy system in zero carbon utilization, comfort protection and safety resilience, and therefore there is room for improvement. SUMMARY

[0005] The purpose of the present application is to provide a campus micro-energy multi-scenario scheduling method, device, equipment and medium, to solve the technical problem that the existing micro-energy system operation method relies on static rules and local single target scheduling strategies for a long time, and it is difficult to respond to photovoltaic output fluctuation, load change and operation scenario switching in time, and the operation effect is difficult to meet the comprehensive needs of the campus micro-energy system in zero carbon utilization, comfort protection and safety resilience.

[0006] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: In a first aspect, the present application provides a campus micro-energy multi-scenario scheduling method, comprising: acquiring target operation data; performing scene identification on the target operation data according to a preset multi-scenario zero-carbon operation strategy model to obtain a current operation scene; determining power control targets of each functional module in a combined device according to the current operation scene and the target operation data, and generating a corresponding power control instruction sequence; controlling the each functional module to make corresponding adjustments in response to the power control instruction sequence.

[0007] By adopting the above technical solution, the multi-source state information of power generation, energy storage, power consumption and environment can be mastered under a unified time reference, thereby providing complete and accurate operation basis for subsequent decision-making; by performing scene identification on the target operation data according to the preset multi-scenario zero-carbon operation strategy model, it can be automatically judged that the current operation scene is low carbon, comfortable or extreme, thereby avoiding the lag and deviation caused by artificial experience-based judgment; by determining the power control targets of each functional module according to the current operation scene and the target operation data and generating a corresponding power control instruction sequence, the coordinated power distribution of each link of generation, storage, consumption and exchange can be completed under a unified framework, thereby realizing multi-objective coordinated optimization under the premise of meeting constraints; by controlling each functional module to make corresponding adjustments in response to the power control instruction sequence, the optimization results can be implemented to the equipment side for execution in real time, thereby realizing dynamic zero-carbon operation of the campus micro "generation-storage-consumption-exchange" combined device.

[0008] In an example, the present application can be further configured to acquire the target operation data, comprising: acquiring multi-source operation data in real time through photovoltaic components, energy storage devices, power consumption loads and environment sensors under a unified time reference; generating photovoltaic power generation prediction data and load prediction data based on weather forecasts and load historical data to obtain a predicted operation data set on a preset time scale; aligning the multi-source operation data with the predicted operation data set in time to obtain the target operation data.

[0009] By adopting the technical scheme, the multi-source operation data of the photovoltaic module, the energy storage device, the power load and the environment sensor are collected under the unified time reference, so that the basic measurement information reflecting the current energy flow and environment state can be obtained, and the calculation error caused by asynchronous data can be avoided; the photovoltaic power generation prediction data and the load prediction data are generated based on the weather forecast and the load historical data on the preset time scale, so that the supply and demand trend in the future period can be known in advance, and the forward-looking input for the rolling optimization is provided; the target operation data is obtained by time alignment of the multi-source operation data and the prediction operation data set, so that the current situation and the prediction can be mapped to the same time axis, and the structured and consistent input data set for the subsequent scene identification and power optimization is provided.

[0010] The application can be further configured in an example as follows: the construction of the multi-scenario zero-carbon operation strategy model comprises: obtaining historical operation data, configuring a set of operation scenes for representing different operation targets according to the historical operation data and the zero-carbon operation demand, and classifying the historical operation data based on the set of operation scenes to obtain a training data set corresponding to each operation scene; organizing the training data set into an observation sequence in time sequence, calculating a target function of a scene identification model based on the observation sequence, and adjusting the model parameters through iteration until the change amount of the target function is less than a preset threshold to obtain scene identification model parameters; configuring power control strategy parameters corresponding to each operation scene according to the set of operation scenes and the scene identification model parameters, and integrating the set of operation scenes, the scene identification model parameters and the power control strategy parameters to generate the multi-scenario zero-carbon operation strategy model.

[0011] By adopting the technical scheme, the multi-source operation data of the photovoltaic module, the energy storage device, the power load and the environment sensor are collected under the unified time reference, so that the basic measurement information reflecting the current energy flow and environment state can be obtained, and the calculation error caused by asynchronous data can be avoided; the photovoltaic power generation prediction data and the load prediction data are generated based on the weather forecast and the load historical data on the preset time scale, so that the supply and demand trend in the future period can be known in advance, and the forward-looking input for the rolling optimization is provided; the target operation data is obtained by time alignment of the multi-source operation data and the prediction operation data set, so that the current situation and the prediction can be mapped to the same time axis, and the structured and consistent input data set for the subsequent scene identification and power optimization is provided.

[0012] In one example, the present invention can be further configured as follows: the step of performing scenario identification on the target operation data according to a preset multi-scenario zero-carbon operation strategy model to obtain the current operation scenario includes: The target operation data is input into the scenario recognition model in the multi-scenario zero-carbon operation strategy model to obtain the state probability of each operation scenario. The running scenarios are compared based on the state probabilities, and the running scenario with the highest state probability is determined as the current running scenario.

[0013] By adopting the above technical solution, the state probability of each operating scenario is obtained by inputting the target operating data into the scenario recognition model in the multi-scenario zero-carbon operation strategy model. This allows for the quantification of the probability that the current moment belongs to different scenarios, thus providing a continuous and smooth basis for scenario switching. By comparing each operating scenario based on the state probability and determining the operating scenario with the highest state probability as the current operating scenario, the operating mode that best matches the current operating condition can be automatically selected, thereby ensuring that the subsequent power control strategy is consistent with the actual scenario.

[0014] In one example, the present invention can be further configured as follows: inputting the target operating data into the scene recognition model in the multi-scenario zero-carbon operating strategy model to obtain the state probability of each operating scenario includes: The target operational data is divided according to a preset observation dimension to obtain an observation sequence; The observation sequence is input into a hidden Markov model to solve for the state probability distribution of each of the operating scenarios. Extract the state probability corresponding to each running scenario from the state probability distribution.

[0015] By adopting the above technical solution, the target operation data is divided into observation sequences according to preset observation dimensions. This allows complex multi-source measurements to be organized into temporal features that meet the model input requirements, thereby improving the modeling efficiency and generalization ability of the scene recognition model. By inputting the observation sequences into a hidden Markov model to solve for the state probability distribution of each operation scenario, the hidden scene state can be inferred under the premise of considering the state transition rules and observation characteristics, thereby improving the stability of multi-scene recognition. By extracting the state probabilities corresponding to each operation scenario from the state probability distribution, refined probability information can be provided to the upper-level scene decision module, thereby achieving precise control over the timing and results of scene switching.

[0016] In one example, the present invention can be further configured as follows: determining the power control target of each functional module in the combined device based on the current operating scenario and the target operating data, and generating a corresponding power control instruction sequence, includes: When the current operating scenario is a low-carbon scenario, the photovoltaic power generation capacity of the future target time window is determined based on the target operating data. The upper and lower limits of the state of charge of the energy storage device are adjusted according to the photovoltaic power generation capacity to obtain a feasible range of the state of charge. The corresponding power control command sequence is generated according to the feasible range of the state of charge. When the current operating scenario is a comfort scenario, thermal comfort index and visual comfort index are calculated based on the target operating data. Under the preset external power purchase constraint, the corresponding power control command sequence is determined with the goal of minimizing the deviation between the thermal comfort index and the visual comfort index. In the case of an extreme scenario, the load is classified according to the target operating data, and a virtual queue of power supply shortage is established for the important loads in the classification results. An evaluation function is constructed based on the virtual queue and the preset limited available power constraint to determine the power control target of each load and generate the corresponding power control instruction sequence.

[0017] By adopting the above technical solutions, under the current low-carbon operating scenario, the photovoltaic power generation capacity of the future target time window is determined based on the target operating data. The upper and lower limits of the state of charge (SOC) of the energy storage device are adjusted accordingly to obtain the feasible SOC range and generate a corresponding power control command sequence. This maximizes the absorption of photovoltaic power while ensuring the safety boundary of energy storage, thereby reducing curtailment and grid-connected power, and improving the local consumption level of green electricity. Furthermore, under the current comfortable operating scenario, thermal comfort and visual comfort indices are calculated based on the target operating data. Under the preset external power purchase constraint, the two types of comfort deviations are minimized. By defining a power control command sequence for the target, it can take into account comfort requirements such as temperature, humidity, and light environment under limited external power purchase conditions, thereby improving the user experience while controlling energy consumption. By classifying the load according to the target operating data and establishing a virtual queue of power supply shortage for important loads under the current extreme operating scenario, and constructing an evaluation function based on the virtual queue and limited available power constraints to determine the power control target of each load and generate a command sequence, it can prioritize lifeline loads and important loads when power supply capacity is limited, thereby extending the power supply duration of critical businesses and improving the resilience of micro energy systems under extreme operating conditions.

[0018] In one example, the present invention can be further configured such that the campus micro-energy multi-scenario scheduling method also includes: Based on the power control command sequence, operational feedback data is collected; Based on the operational feedback data, the model parameters of the multi-scenario zero-carbon operation strategy model are adaptively adjusted to obtain the updated multi-scenario zero-carbon operation strategy model.

[0019] By adopting the above technical solution and collecting operational feedback data based on power control command sequences, indicators such as photovoltaic absorption rate, comfort compliance rate, and key load supply status during actual execution can be fed back to the strategy layer, thus providing a quantitative basis for operational performance evaluation and strategy correction. By adaptively adjusting the model parameters of the multi-scenario zero-carbon operation strategy model based on operational feedback data, an updated multi-scenario zero-carbon operation strategy model is obtained. This enables the operation strategy to be continuously optimized with environmental changes and load characteristic evolution, thereby maintaining the adaptability and comprehensive performance advantages of the multi-scenario zero-carbon operation method in long-term operation.

[0020] In a second aspect, the present invention provides a campus micro-energy multi-scenario scheduling device, the device comprising: The data acquisition module is used to acquire target runtime data; The scene recognition module is used to identify the target operation data according to a preset multi-scenario zero-carbon operation strategy model to obtain the current operation scenario; The power target module is used to determine the power control target of each functional module in the combined device based on the current operating scenario and the target operating data, and generate the corresponding power control instruction sequence. An execution control module is used to control each functional module to make corresponding adjustments in response to the power control command sequence.

[0021] By adopting the above technical solutions and acquiring target operating data, it is possible to grasp the status information of multiple sources such as power generation, energy storage, power consumption, and the environment under a unified time benchmark, thereby providing a complete and accurate operational basis for subsequent decision-making. By identifying the target operating data according to the preset multi-scenario zero-carbon operation strategy model, it is possible to automatically determine whether the current operating scenario is low-carbon, comfortable, or extreme, thereby avoiding the lag and bias caused by manual experience-based judgment. By determining the power control targets of each functional module based on the current operating scenario and target operating data and generating corresponding power control command sequences, it is possible to complete the coordinated power allocation of each link of generation, storage, consumption, and exchange under a unified framework, thereby achieving multi-objective coordinated optimization under the premise of meeting constraints. By responding to the power control command sequence to control each functional module to make corresponding adjustments, the optimization results can be implemented to the equipment side in real time, thereby realizing the dynamic zero-carbon operation of the campus micro "generation-storage-consumption-exchange" combined device.

[0022] In a third aspect, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the campus micro-energy multi-scenario scheduling method described above.

[0023] In a fourth aspect, the present invention provides a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the campus micro-energy multi-scenario scheduling method described above. Attached Figure Description

[0024] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of a campus micro-energy multi-scenario scheduling method in an embodiment of the present invention; Figure 2 This is a structural block diagram of the campus micro-energy multi-scenario scheduling device according to an embodiment of the present invention; Figure 3 This is a structural block diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0025] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0026] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this invention is for describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0027] Example 1 like Figure 1 As shown, this invention discloses a multi-scenario scheduling method for campus micro-energy, which specifically includes the following steps: S10: Obtain target runtime data.

[0028] Specifically, under the cloud-edge-device collaborative architecture, power generation, energy storage, power consumption, power swapping, and environmental status information within a scheduling cycle are collected and preprocessed on a unified time base. Operational data from different sources and at different time scales are archived and packaged in a unified format to form target operational data.

[0029] S20: Based on the preset multi-scenario zero-carbon operation strategy model, the target operation data is identified to obtain the current operation scenario.

[0030] Specifically, the target operating data is input into a preset multi-scenario zero-carbon operation strategy model. The probability of occurrence of each operating scenario is calculated based on the degree of matching of the characteristic patterns of different operating scenarios within the model. The probability vectors representing candidate scenarios such as low-carbon scenarios, comfort scenarios, and extreme scenarios are output. Then, the current operating scenario for power control is determined based on the probability vectors.

[0031] S30: Based on the current operating scenario and target operating data, determine the power control target of each functional module in the combined unit and generate the corresponding power control command sequence.

[0032] Specifically, after identifying the current operating scenario, an optimization strategy corresponding to the scenario is selected from the preset multi-scenario zero-carbon operation strategy model. The target operating data is substituted into the power allocation and constraint model of the strategy to solve the power control targets of the photovoltaic power generation module, energy storage module, power load module and battery swapping module. The power control command sequence is then organized into a time series according to the scheduling cycle and sampling time step.

[0033] S40: In response to the power control command sequence, control each functional module to make corresponding adjustments.

[0034] Specifically, the power control command sequence is sent down to the device control layer level by level through the command link from the cloud to the edge. The target power and status set value obtained at each control time step are parsed to drive the photovoltaic grid-connected interface, energy storage converter unit, heating and lighting equipment and battery swapping interface to complete the start-up, shutdown, power adjustment or working mode switching. Combined with the real-time backed status information, abnormal operating conditions are protected and necessary temporary corrections are made.

[0035] In one embodiment, step S10, namely acquiring the target running data, includes: S11: Under a unified time reference, multi-source operation data is collected in real time through photovoltaic modules, energy storage devices, power loads and environmental sensors.

[0036] Specifically, the instantaneous voltage and current of each photovoltaic string are collected by voltage and current sensors deployed on the DC side and inverter side of the photovoltaic modules, and the photovoltaic output power is calculated. The DC bus voltage, battery current, and state of charge (SOC) are collected by the energy storage device battery stack and bidirectional converter unit to obtain the charging and discharging power and energy storage status. The active power, power factor, and switch status are collected by smart energy meters installed on the far-infrared heating equipment, adjustable light and color temperature lighting fixtures, and battery swapping cabinet branches. At the same time, environmental sensing devices such as indoor and outdoor temperature sensors, humidity sensors, illuminance sensors, and cameras are used to collect environmental information such as temperature, relative humidity, illuminance, and the number of people in the room. The above measurements are collected synchronously at a sampling period of seconds or minutes under a unified time reference, and multi-source operation data sequences are formed according to timestamps.

[0037] S12: At a preset time scale, photovoltaic power generation forecast data and load forecast data are generated based on weather forecasts and historical load data to obtain a set of forecast operation data.

[0038] Specifically, at the day-ahead level, with a 24-hour cycle, the solar irradiance and ambient temperature prediction sequences for each time period within the next day are obtained based on the accessed high-precision weather forecast data. Combining the historical correlation between photovoltaic power and irradiance and temperature over multiple days, the photovoltaic power generation prediction values ​​for each time period in the future are obtained through fitting models. At the same time, similar day analysis and time series modeling are performed on the historical campus load data. Based on factors such as weekday or holiday type, timestamp, and temperature, the power demand of various types of electricity loads in the future period is predicted. The obtained photovoltaic power generation prediction curve and load prediction curve are discretized according to the time granularity consistent with the operation data to form a prediction operation data set containing multiple prediction variables.

[0039] S13: Time-align the multi-source running data with the predicted running data set to obtain the target running data.

[0040] Specifically, using a unified sampling interval within the scheduling cycle as a time index, multi-source operational data and predicted operational data are resampled and interpolated. When there are differences in sampling frequency or timestamp offsets, various types of data are mapped to a unified time grid through linear interpolation or zero-order hold-up methods. Missing data are supplemented using data from adjacent time points or short-term predictions. Then, at each discrete time point, the corresponding measured photovoltaic power, energy storage status, power consumption, grid purchase and sale of electricity, environmental variables, and photovoltaic and load prediction values ​​for several future time periods are combined into a multi-dimensional feature vector, which is then stacked in chronological order to form the target operational data sequence.

[0041] In one embodiment, step S20, namely the construction of the multi-scenario zero-carbon operation strategy model, includes: S201: Obtain historical operation data, configure a set of operation scenarios to represent different operation objectives based on the historical operation data and zero-carbon operation requirements, and then classify the historical operation data based on the set of operation scenarios to obtain the training dataset corresponding to each operation scenario.

[0042] Specifically, multi-dimensional time series data, including photovoltaic power, load power, grid power purchase and sales, energy storage SOC, equipment operating status, indoor environmental parameters, and power supply status of critical loads, are extracted from long-term campus operation records. Based on the zero-carbon operation requirements, such as maximizing local photovoltaic consumption, ensuring thermal and visual comfort, or extending the power supply duration of critical loads, and combined with grid connection status and extreme weather markers, historical data are divided into several operation scenarios, such as low-carbon scenarios, comfortable scenarios, and extreme scenarios. Each time slice is labeled with a corresponding scenario, and the labeled samples under the same scenario are organized in chronological order to form a training sample set, thus forming a training dataset corresponding to each operation scenario.

[0043] S202: Organize the training dataset into an observation sequence in chronological order, calculate the objective function of the scene recognition model based on the observation sequence, and iteratively adjust the model parameters until the change in the objective function is less than a preset threshold to obtain the scene recognition model parameters.

[0044] Specifically, the observations such as photovoltaic power, load power, grid status, timestamp, and occupancy rate contained in the training dataset of each operating scenario are arranged sequentially according to the sampling time to form multiple observation sequences. The scene recognition model is abstracted into a probabilistic model with hidden states such as low-carbon scene, comfortable scene, and extreme scene. The log-likelihood of the training observation sequence under given model parameters is used as the objective function. In each iteration, the posterior probability of each time step belonging to different hidden states and the expected state transition probability between adjacent time steps are calculated using the current parameters. Based on this, the initial state probability distribution, state transition probability matrix, and observation probability parameters are updated. The iteration stops when the difference between the objective function values ​​obtained from two adjacent iterations is less than the preset convergence threshold. The converged parameter set is used as the scene recognition model parameters.

[0045] S203: Based on the set of operating scenarios and the parameters of the scenario recognition model, configure the power control strategy parameters corresponding to each operating scenario, and integrate the set of operating scenarios, the parameters of the scenario recognition model, and the parameters of the power control strategy to generate a multi-scenario zero-carbon operating strategy model.

[0046] Specifically, after obtaining the scene recognition model parameters, a quadratic programming objective function with weights α, β, and γ, primarily based on photovoltaic self-consumption and energy storage flexibility, and upper and lower limits of the safe state of charge (SOC) for energy storage are configured for low-carbon scenarios. min SOC maxParameters such as a, b, and c are used to configure the fitting coefficients a, b, and c of temperature and humidity deviations in the thermal comfort model and the weights ω of illuminance and color temperature deviations in the visual comfort model for comfort scenarios. The weights λ1, λ2, and λ3 for balancing comfort and purchased electricity costs in the multi-objective optimization function are further determined. For extreme scenarios, L1, L2, and L3 grading thresholds and initial virtual queue lengths are set according to the importance of the load. The utility function form and control parameter V in the drift plus penalty framework are selected. The above set of operating scenarios, scenario identification model parameters, and power control strategy parameters corresponding to each scenario are solidified and saved in a unified data structure to form a multi-scenario zero-carbon operation strategy model that can be directly called during the operation phase.

[0047] In one embodiment, step S20, namely, identifying the target operation data according to a preset multi-scenario zero-carbon operation strategy model to obtain the current operation scenario, includes: S21: Input the target operation data into the scenario recognition model in the multi-scenario zero-carbon operation strategy model to obtain the state probability of each operation scenario.

[0048] Specifically, the target operation data sequence is input into the scene recognition model in the multi-scenario zero-carbon operation strategy model step by step according to the sampling time. At each time step, the state probability of each operation scenario, such as low-carbon scenario, comfortable scenario and extreme scenario, is calculated based on the input observation and trained parameters to obtain a state probability vector sequence composed of multiple probability components.

[0049] S22: Compare the running scenarios based on the state probabilities and determine the running scenario with the highest state probability as the current running scenario.

[0050] Specifically, the state probability vectors obtained at consecutive time steps are analyzed. At each time step, the state probabilities corresponding to low-carbon scenarios, comfortable scenarios, and extreme scenarios are compared. When the probability advantage is obvious and the preset smooth switching conditions are met, the running scenario with the highest state probability is determined as the current running scenario. When the probability change is not obvious, the scenario of the previous moment is maintained to avoid frequent switching, and the moment of scenario switching is recorded.

[0051] In one embodiment, step S21 involves inputting the target operation data into the scenario recognition model within the multi-scenario zero-carbon operation strategy model to obtain the state probability of each operation scenario, including: S211: Divide the target running data according to the preset observation dimensions to obtain the observation sequence.

[0052] Specifically, based on the pre-set observation set of the multi-scenario zero-carbon operation strategy model, the target operation data for each time step is broken down into multiple observation components such as photovoltaic power, load power, grid connection status, timestamp, and personnel presence rate. The observation vectors for each time step are then organized into an observation sequence O={o1,o2,…,o…} according to chronological order. T}, where each o t A corresponding observation vector containing the aforementioned multidimensional variables provides input for implicit state inference.

[0053] S212: Input the observation sequence into the Hidden Markov Model to solve the state probability distribution of each operating scenario.

[0054] Specifically, the observation sequence O is input into a Hidden Markov Model with low-carbon scenario, comfortable scenario and extreme scenario as the hidden state set S={s1,s2,s3} and photovoltaic power, load power, grid status, timestamp and personnel presence rate as the observation set. Given the initial state probability, state transition probability matrix and observation probability parameters, the conditional probability of the observation sequence at each time step t and the posterior probability of each hidden state are calculated using the forward recursion equal probability inference method, so as to obtain the state probability distribution of each operating scenario in S.

[0055] S213: Extract the state probabilities corresponding to each running scenario from the state probability distribution.

[0056] Specifically, in the state probability distribution at each time step, the probability components corresponding to the low-carbon scenario, the comfortable scenario, and the extreme scenario are read in the order of the hidden state set S, and denoted as π respectively. low (t), π comfort (t), π extreme (t), and form a state probability sequence {π} in chronological order. low (t),π comfort (t),π extreme (t)}.

[0057] In one embodiment, step S30, namely, determining the power control target of each functional module in the combined device based on the current operating scenario and target operating data, and generating a corresponding power control command sequence, includes: S31: Under the current low-carbon operating scenario, determine the photovoltaic power generation capacity for the future target time window based on the target operating data, adjust the upper and lower limits of the state of charge of the energy storage device according to the photovoltaic power generation capacity to obtain the feasible range of the state of charge, and generate the corresponding power control command sequence according to the feasible range of the state of charge.

[0058] Specifically, in a low-carbon scenario, a target time window consisting of T time steps is selected. A quadratic programming model with intraday rolling optimization is established using photovoltaic power forecasts and load forecasts as inputs, and an objective function is constructed. Where t is the time index, representing a discrete time point within a scheduling cycle; T is the total time range for optimized scheduling; This represents the electrical power (kW) that the device purchases from the external power grid at time t. A value greater than 0 indicates that electricity has been purchased. The electrical power (kW) delivered by the device to the external power grid at time t; a value greater than 0 indicates electricity sales. Let t be the state of charge (%) of the energy storage module at time t, representing the relative remaining capacity of the battery; The reference state of charge (%) for the energy storage module is usually set to 50% to allow for flexible adjustment for sudden charging and discharging demands. The actively discarded photovoltaic power (kW) at time t is the last resort when photovoltaic output is excessive and cannot be absorbed or stored; α, β, γ are weighting coefficients, all positive numbers, and their magnitudes are related as follows: α β>γ reflects the priority of the optimization objectives: eliminating grid interaction > maintaining flexible energy storage > avoiding curtailment of solar power; at the same time, power balance constraints are imposed. and dynamic constraints of energy storage ,in, The actual output power (kW) of the photovoltaic power generation module at time t. Let be the discharge power (kW) of the energy storage battery at time t. Let t be the charging power (kW) of the energy storage battery. The total load power (kW) of electrical modules such as heating and lighting at time t. Let t be the charging power (kW) of the battery swapping module. The charging efficiency (%) of the energy storage battery represents the efficiency of converting electrical energy into chemical energy. The discharge efficiency (%) of the energy storage battery represents the efficiency of converting chemical energy into electrical energy. The time interval for scheduling (h); The rated capacity (kWh) of the energy storage battery is set, and safety and dynamic feasible domain constraints are imposed on the energy storage power and state of charge. ,in, The minimum state of charge (%) for energy storage; The upper limit of the energy storage state of charge (%). The maximum charging power of the energy storage (kW); The maximum discharge power of the energy storage (kW); This is a charging status indicator variable; 1 indicates charging, and 0 indicates no charging. The variable is a discharge state indicator, where 1 indicates discharge and 0 indicates no discharge. This is then used to predict the photovoltaic power output over the next two hours. Assess the available charging energy and the SOC min (t) is defined as the minimum permissible state of charge (SOC) under the premise of ensuring basic power supply needs. max (t) through Dynamic updates are performed, among which, This is the absolute lower limit of the state of charge (%) of energy storage, usually set to 20% to protect battery life; This is the absolute upper limit of the state of charge (%) of energy storage, usually set to 90% to prevent overcharging; The limit of energy storage state of charge (%) at time t; The upper limit of the state of charge of the energy storage at time t (%). Given the current state of charge (in %) of the energy storage, the feasible range of SOC for each time step within the time window is given, and the optimized energy storage charging and discharging power, grid power purchase and sale power, and other decision quantities are mapped into a power control command sequence for photovoltaic power allocation and energy storage control.

[0059] S32: When the current operating scenario is a comfort scenario, calculate the thermal comfort index and visual comfort index based on the target operating data. Under the preset external power consumption constraint, determine the corresponding power control command sequence with the goal of minimizing the deviation between the thermal comfort index and the visual comfort index.

[0060] Specifically, in a comfort scenario, based on the indoor temperature T in the target operating data... in (t), set temperature T set Relative humidity RH(t), set relative humidity RH set And variables such as the metabolic rate M of human activity estimated by the camera, according to Calculate the thermal comfort index PMV(t) at each time step, and simultaneously use the actual illuminance E. actual (t), Optimal Illuminance E optimal Actual correlated color temperature (CCT) actual (t) and optimal correlated color temperature (CCT) optimal Calculate the visual comfort index Based on this, a multi-objective optimization function is constructed. ,in, , , These are weighting coefficients used to balance the importance of the three objectives of thermal comfort, visual comfort, and the cost of purchased electricity, and to impose a constraint on the power of purchased electricity. By limiting the proportion of purchased electricity to the total load, and using PMV deviation, VCI deviation, and the ratio of current photovoltaic output to total load as input fuzzy variables, the system performs inference through a pre-built fuzzy rule base, outputting the power adjustment of heating equipment, the adjustment of lighting brightness and color temperature, and the priority of energy storage discharge power, thereby generating a power control command sequence that meets the comfort target and the constraint of purchased electricity.

[0061] S33: In the case of an extreme scenario, the load is classified according to the target operating data, and a virtual queue of power supply shortage is established for the important loads in the classification results. Based on the virtual queue and the preset limited available power constraints, an evaluation function is constructed to determine the power control target of each load and generate the corresponding power control instruction sequence.

[0062] Specifically, in extreme scenarios, the load is first divided into L1 lifeline loads, L2 critical loads, and L3 comfort loads according to a pre-defined strategy. For each type of L2 critical load, a virtual queue Q for power supply shortage is defined. i (t), using Qi(t+1)=max[Qi(t)] μi(t),0]+Ai(t), where, For time t, the queue of power supply shortage for the i-th type L2 critical load is defined. The queue length intuitively reflects the cumulative unmet power demand (kWh) of the load, i.e. the degree of "starvation". The power (kW) actually allocated to the i-th type L2 level critical load at time t. To determine the new electricity demand (kW) generated by the i-th type L2 critical load at time t, a Lyapunov function is then constructed. and drift items ,in, For all load virtual queue state vectors, combine with the utility function Non-negative control parameter V forms a drift plus penalty function Solve the real-time optimization problem within each control cycle. ,in, The aim is to prioritize power allocation to the "hungerest" loads, thereby stabilizing the queue; Aiming to maximize system utility, under constraints The total available power that can be allocated to L2 level loads is determined below, where, Let be the total power (kW) of the L1 level load at time t, and based on the obtained μ... i (t) Determine the power supply of critical loads L2, and decide whether to retain or cut off comfort loads L3 based on the remaining power, thereby generating a power control command sequence that prioritizes critical loads under off-grid or limited power supply conditions.

[0063] In one embodiment, the campus micro-energy multi-scenario scheduling method further includes: S50: Collects operational feedback data based on power control command sequences.

[0064] Specifically, during the continuous issuance and execution of power control command sequences, key operational variables such as photovoltaic output power, energy storage charging and discharging power and SOC, power and switching status of various electrical loads, charging power of battery swapping modules, power purchase and sale from the grid, indoor temperature and humidity, illuminance and related color temperature are collected from the device at a time resolution consistent with or finer than the scheduling cycle. At the same time, the on-site photovoltaic absorption rate in low-carbon scenarios, the compliance rate of thermal comfort and visual comfort indicators in comfortable scenarios, and the continuous power supply duration of key loads at all levels in extreme scenarios are statistically analyzed. After filtering and denoising the raw sampled data and removing abnormal data at the edge, the processed operational feedback data is uploaded to the cloud via the communication link to provide a basis for strategy evaluation and model adaptive updates.

[0065] S60: Based on the operational feedback data, the model parameters of the multi-scenario zero-carbon operation strategy model are adaptively adjusted to obtain the updated multi-scenario zero-carbon operation strategy model.

[0066] Specifically, a performance evaluation system is constructed in the cloud based on operational feedback data, including photovoltaic absorption rate, comfort compliance rate, operational economic indicators, and resilience indicators. The system quantifies the operational effects under different time periods and scenarios, and uses the evaluation results as reward signals for reinforcement learning or heuristic search. Online or offline parameter updates are performed on the low-carbon scenario objective function weights α, β, and γ, the comfort scenario multi-objective optimization weights λ1, λ2, and λ3, the control parameter V in the extreme scenario drift plus penalty framework, as well as the fuzzy rule base parameters and scenario transition probability matrix. At the same time, some local model parameters are fine-tuned at the edge by combining the latest local operational data. This allows the updated multi-scenario zero-carbon operation strategy model to better adapt to the load fluctuations and scenario switching of campus micro-energy in subsequent operations, thereby achieving continuous adaptive optimization of the zero-carbon operation strategy.

[0067] Example 2 like Figure 2 As shown, based on the same inventive concept as the above embodiments, the present invention also provides a campus micro-energy multi-scenario scheduling device, comprising: The data acquisition module is used to acquire target runtime data; The scene recognition module is used to identify the current operating scene by analyzing the target operating data based on a preset multi-scenario zero-carbon operation strategy model. The power target module is used to determine the power control target of each functional module in the combined device based on the current operating scenario and target operating data, and generate the corresponding power control command sequence. The execution control module is used to control each functional module to make corresponding adjustments in response to the power control command sequence.

[0068] Optionally, the data acquisition module includes: The operation acquisition submodule is used to collect multi-source operation data in real time through photovoltaic modules, energy storage devices, power loads and environmental sensors under a unified time reference. The prediction generation submodule is used to generate photovoltaic power generation prediction data and load prediction data based on weather forecasts and historical load data at a preset time scale, so as to obtain a set of predicted operation data. The alignment processing submodule is used to perform time alignment between multi-source running data and the predicted running data set to obtain the target running data.

[0069] Optional, the construction of multi-scenario zero-carbon operation strategy models includes: The scenario configuration module is used to acquire historical operation data, configure a set of operation scenarios to represent different operation goals based on the historical operation data and zero-carbon operation requirements, and then classify the historical operation data based on the set of operation scenarios to obtain the training dataset corresponding to each operation scenario. The training optimization module is used to organize the training dataset into an observation sequence in chronological order, calculate the objective function of the scene recognition model based on the observation sequence, and iteratively adjust the model parameters until the change in the objective function is less than a preset threshold to obtain the scene recognition model parameters. The strategy integration module is used to configure the power control strategy parameters corresponding to each operating scenario based on the set of operating scenarios and the parameters of the scenario recognition model, and to integrate the set of operating scenarios, the parameters of the scenario recognition model and the parameters of the power control strategy to generate a multi-scenario zero-carbon operating strategy model.

[0070] Optionally, the scene recognition module includes: The scene recognition submodule is used to input the target operation data into the scene recognition model in the multi-scenario zero-carbon operation strategy model to obtain the state probability of each operation scenario; The scenario determination submodule is used to compare each running scenario based on the state probability and determine the running scenario with the highest state probability as the current running scenario.

[0071] Optionally, the scene recognition submodule includes: The observation construction unit is used to divide the target operation data according to the preset observation dimensions to obtain the observation sequence; The probability solving unit is used to input the observation sequence into the Hidden Markov Model and solve for the state probability distribution of each operating scenario. The probability extraction unit is used to extract the state probabilities corresponding to each running scenario from the state probability distribution.

[0072] Optionally, the power target module includes: The low-carbon control submodule is used to determine the photovoltaic power generation capacity of the future target time window based on the target operating data when the current operating scenario is a low-carbon scenario, adjust the upper and lower limits of the state of charge of the energy storage device according to the photovoltaic power generation capacity, obtain the feasible range of the state of charge, and generate the corresponding power control command sequence according to the feasible range of the state of charge. The comfort control submodule is used to calculate thermal comfort index and visual comfort index based on target operating data when the current operating scenario is a comfort scenario. Under the preset external power consumption constraint, the module aims to minimize the deviation of thermal comfort index and visual comfort index and determine the corresponding power control command sequence. The extreme control submodule is used to classify the load according to the target operating data when the current operating scenario is an extreme scenario, and to establish a virtual queue of power supply shortage for important loads in the classification results. Based on the virtual queue and the preset limited available power constraints, an evaluation function is constructed to determine the power control target of each load and generate the corresponding power control command sequence.

[0073] Optionally, the campus micro-energy multi-scenario scheduling method also includes: The feedback acquisition module is used to acquire operational feedback data based on the power control command sequence; The model update module is used to adaptively adjust the model parameters of the multi-scenario zero-carbon operation strategy model based on the operation feedback data, so as to obtain the updated multi-scenario zero-carbon operation strategy model.

[0074] Example 3 like Figure 3 As shown, the present invention also provides an electronic device 100 for implementing a multi-scenario scheduling method for campus micro-energy; The electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on at least one processor 102, and at least one communication bus 104.

[0075] The memory 101 can be used to store the computer program 103. The processor 102 implements the steps of the campus micro-energy multi-scenario scheduling method of Embodiment 1 by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101.

[0076] The memory 101 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0077] At least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 102 may be a microprocessor or any conventional processor. Processor 102 is the control center of electronic device 100, connecting various parts of electronic device 100 via various interfaces and lines.

[0078] The memory 101 in the electronic device 100 stores multiple instructions to implement a multi-scenario scheduling method for campus micro-energy, and the processor 102 can execute multiple instructions to achieve the following: Obtain target runtime data; Based on the preset multi-scenario zero-carbon operation strategy model, the target operation data is used to identify the current operation scenario; Based on the current operating scenario and target operating data, determine the power control target of each functional module in the combined unit and generate the corresponding power control instruction sequence; The system responds to the power control command sequence and controls each functional module to make corresponding adjustments.

[0079] Example 4 If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they 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 of the present invention can also 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: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).

[0080] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0081] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0082] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0083] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0084] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for multi-scenario scheduling of campus micro energy, characterized in that, The method comprises: acquiring target operation data; performing scene recognition on the target operation data according to a preset multi-scene zero-carbon operation strategy model to obtain a current operation scene; determining power control targets of each functional module in a combination device according to the current operation scene and the target operation data, and generating a corresponding power control instruction sequence; controlling the functional modules to perform corresponding adjustments in response to the power control instruction sequence.

2. The multi-scene scheduling method of the campus micro energy source according to claim 1, wherein the acquiring target operation data comprises: collecting multi-source operation data in real time through photovoltaic components, energy storage devices, power consumption loads, and environmental sensors under a unified time reference; generating photovoltaic power generation prediction data and load prediction data based on weather forecasts and load historical data on a preset time scale to obtain a prediction operation data set; aligning the multi-source operation data with the prediction operation data set in time to obtain the target operation data.

3. The multi-scene scheduling method of the campus micro energy source according to claim 1, wherein the construction of the multi-scene zero-carbon operation strategy model comprises: acquiring historical operation data, configuring a set of operation scenes for representing different operation targets according to the historical operation data and zero-carbon operation requirements, and then classifying the historical operation data based on the set of operation scenes to obtain a training data set corresponding to each operation scene; organizing the training data set into observation sequences in chronological order, calculating an objective function of a scene recognition model based on the observation sequences, and adjusting model parameters through iteration until the change amount of the objective function is less than a preset threshold to obtain scene recognition model parameters; configuring power control strategy parameters corresponding to each operation scene according to the set of operation scenes and the scene recognition model parameters, and integrating the set of operation scenes, the scene recognition model parameters, and the power control strategy parameters to generate the multi-scene zero-carbon operation strategy model.

4. The multi-scene scheduling method of the campus micro energy source according to claim 1, wherein the performing scene recognition on the target operation data according to a preset multi-scene zero-carbon operation strategy model to obtain a current operation scene comprises: inputting the target operation data into a scene recognition model in the multi-scene zero-carbon operation strategy model to obtain state probabilities of each operation scene; comparing each operation scene according to the state probabilities to determine an operation scene with the largest state probability as the current operation scene.

5. The multi-scene scheduling method of the campus micro energy source according to claim 4, wherein the inputting the target operation data into a scene recognition model in the multi-scene zero-carbon operation strategy model to obtain state probabilities of each operation scene comprises: dividing the target operation data according to a preset observation dimension to obtain observation sequences; inputting the observation sequences into a hidden Markov model to solve state probability distributions of each operation scene; extracting the state probabilities of each operation scene from the state probability distributions.

6. The method of claim 1, wherein the power control target of each functional module in the combined device is determined based on the current running scenario and the target running data, and a corresponding power control instruction sequence is generated, including: in the case of the current running scenario being a low-carbon scenario, determining the photovoltaic power generation capacity of a future target time window based on the target running data, adjusting the upper and lower limits of the state of charge of the energy storage device according to the photovoltaic power generation capacity to obtain a feasible state of charge interval, and generating a corresponding power control instruction sequence according to the feasible state of charge interval; in the case of the current running scenario being a comfort scenario, calculating a thermal comfort index and a visual comfort index according to the target running data, and determining a corresponding power control instruction sequence under the condition of a preset external power purchase power constraint, with the minimum deviation of the thermal comfort index and the visual comfort index as the target; and in the case of the current running scenario being an extreme scenario, grading the load according to the target running data, establishing a virtual queue of power supply shortage for important loads in the grading result, constructing an evaluation function based on the virtual queue and a preset limited available power constraint, determining the power control target of each load, and generating a corresponding power control instruction sequence. The method further includes: collecting running feedback data based on the power control instruction sequence; and adaptively adjusting the model parameters of the multi-scenario zero-carbon running strategy model according to the running feedback data to obtain an updated multi-scenario zero-carbon running strategy model. The device includes: a data acquisition module configured to acquire target running data; a scenario identification module configured to identify the current running scenario of the target running data according to a preset multi-scenario zero-carbon running strategy model; a power target module configured to determine the power control target of each functional module in the combined device based on the current running scenario and the target running data, and generate a corresponding power control instruction sequence; and an execution control module configured to control the corresponding adjustment of the functional modules in response to the power control instruction sequence. The computer readable storage medium stores at least one instruction, which is executed by the processor to implement the steps of the multi-scenario scheduling method of the campus micro energy source as claimed in any one of claims 1 to 7.

7. The campus micro-energy multi-scenario scheduling method according to claim 1, characterized in that, ​ ​ ​ 8. A campus micro-energy multi-scenario scheduling device, characterized in that, ​ ​ ​ ​ ​ 9. An electronic device, comprising: ​ 10. A computer-readable storage medium, characterized in that, ​

Images