Integrated energy collaborative management methods, equipment and storage media

Abstract

This application discloses a comprehensive energy collaborative management method, device, and storage medium, comprising: collecting multi-source heterogeneous data of energy equipment in an industrial park; constructing standardized feature inputs based on the multi-source heterogeneous data to obtain feature data; inputting the feature data into an adaptive prediction model for multi-time-scale load prediction; constructing a multi-objective optimization model based on the load prediction results; obtaining real-time equipment information of the energy equipment; inputting the real-time equipment information into the multi-objective optimization model for calculation; generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model; and controlling the operation of the energy equipment using the equipment scheduling strategy. This invention improves system response speed and operating efficiency, significantly reduces energy consumption and carbon emissions, and enhances the system's ability to cope with load fluctuations through a software and hardware collaborative closed-loop control mechanism of multi-source data fusion, load prediction, and multi-objective optimization.

Description

Technical Field

[0001] This application relates to the field of energy management technology, and in particular to a comprehensive energy collaborative management method, device and storage medium. Background Technology

[0002] With the continued advancement of the "dual-carbon" strategy and the development of multi-energy complementary technologies, integrated energy systems are widely used in regional scenarios such as industrial parks and commercial complexes. These systems typically integrate multiple energy forms, including electricity, heat, cooling, and gas, and are characterized by multi-energy flow coupling and the coordinated operation of multiple types of equipment. To achieve optimal energy efficiency, controllable costs, and controllable carbon emissions, the system needs higher scheduling accuracy and response flexibility. However, most current integrated energy scheduling systems still adopt a "centralized optimization + decentralized control" approach, with separated hardware and software modules and delayed response, making it difficult to meet the comprehensive requirements for real-time performance, intelligence, and reliability under high-frequency disturbance scenarios. Therefore, there is an urgent need for a scheduling and management device with hardware and software collaborative operation capabilities, supporting real-time strategy deployment and closed-loop execution of control commands, to improve the intelligent control level and implementation capability of integrated energy systems in practical engineering projects.

[0003] In current technologies, the management of integrated energy systems primarily employs a hierarchical control architecture. This involves strategy planning through a host computer system (such as an Energy Management System, EMS), supplemented by field controllers (such as PLCs and DCSs) to handle equipment start-up, shutdown, and status regulation. These systems often rely on centralized optimization, periodic scheduling, and manually set rules to achieve supply and demand balance and equipment coordination across multiple energy sources. However, this approach has significant limitations: firstly, system response speed depends on the upper-level calculation and command issuance cycle, making it difficult to handle frequently fluctuating load scenarios; secondly, the separation of optimization algorithms and execution logic results in high deployment costs, significant communication latency, and insufficient system flexibility and real-time performance.

[0004] Some studies have attempted to deploy real-time optimization algorithms on the server side, combining data acquisition and state awareness for rolling scheduling. However, deep integration of scheduling strategies and control execution has not yet been achieved. Software and hardware rely on interface connections and lack an integrated linkage mechanism. In addition, most hardware systems commonly used in current industrial applications (such as embedded controllers) employ fixed-value control or priority start / stop strategies, which are insufficient to support dynamic adjustment and reasoning capabilities of scheduling logic.

[0005] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main objective of this application is to provide a method, device, and storage medium for integrated energy collaborative management, which aims to solve the technical problems of real-time optimization scheduling and closed-loop control in existing integrated energy management.

[0007] To achieve the above objectives, this application proposes a comprehensive energy collaborative management method, the method comprising:

[0008] Collect multi-source heterogeneous data of energy equipment in industrial parks, and construct standardized feature inputs based on the multi-source heterogeneous data to obtain feature data;

[0009] The feature data is input into an adaptive prediction model for multi-time-scale load prediction.

[0010] Construct a multi-objective optimization model based on load forecasting results;

[0011] Obtain the real-time equipment information of the energy equipment, and input the real-time equipment information into the multi-objective optimization model for calculation;

[0012] Based on the calculation results of the multi-objective optimization model, an equipment scheduling strategy is generated, and the operation of the energy equipment is controlled by the equipment scheduling strategy.

[0013] In one embodiment, before the step of inputting the feature data into the adaptive prediction model for multi-timescale load prediction, the method further includes:

[0014] Training samples are generated based on historical load data and external environmental characteristics, and a multi-task deep learning prediction model is trained based on the training samples.

[0015] The training results are optimized using a self-supervised learning method, and the adaptive prediction model is obtained based on the optimization results.

[0016] In one embodiment, the step of constructing a multi-objective optimization model based on load forecasting results includes:

[0017] The load forecast results are analyzed, and a basic function, a cost function, and a carbon emission function are generated based on the analysis results.

[0018] The basic function, cost function, and carbon emission function are combined to generate an optimization objective function, and the multi-objective optimization model is constructed based on the optimization objective function;

[0019] The steps of parsing the load forecast results and generating the basic function, cost function, and carbon emission function based on the parsing results include:

[0020] Extract the prediction results based on electricity, heat and cooling load from the analysis results, calculate the total energy input of coupled energy demand in the future time period based on the prediction results, and generate the basic function based on the total energy input;

[0021] The load curve and time-of-use energy price signal are extracted from the analysis results, and the operation and maintenance costs required for equipment start-up, shutdown and operating condition switching are obtained to generate the cost function.

[0022] The load demand and equipment carbon emission characteristics are extracted from the analysis results. Based on the nonlinear relationship between the equipment start-up and shutdown process, load rate and the equipment carbon emission characteristics, a dynamic carbon emission function is generated.

[0023] In one embodiment, the step of fusing the basic function, cost function, and carbon emission function to generate an optimization objective function, and constructing the multi-objective optimization model based on the optimization objective function, includes:

[0024] The weighting coefficients of the cost function and carbon emission function are determined based on the preset operating cost target and carbon emission target, and the weighting coefficients are dynamically adjusted based on real-time energy price signals and carbon emission quotas.

[0025] The cost function and the carbon emission function are normalized, and the weighted summation method is used to merge the normalized cost function and carbon emission function into a comprehensive objective function.

[0026] The multi-objective optimization model is generated using the integrated objective function and the basic function.

[0027] In one embodiment, the step of obtaining real-time device information of the energy equipment and inputting the real-time device information into the multi-objective optimization model for calculation includes:

[0028] Collect real-time device information of the energy equipment, and calculate the deviation value and current health index of the energy equipment based on the real-time device information and a preset calculation threshold;

[0029] The constraints of the multi-objective optimization model are configured based on the deviation value, and the constraints are dynamically adjusted based on the current health index.

[0030] The current health index is converted into a device reliability coefficient and embedded into the device information. The device information is then converted into input parameters for the multi-objective optimization model and input into the multi-objective optimization model for calculation.

[0031] In one embodiment, the step of converting the current health index into a device reliability coefficient and embedding it into the device information, and converting the device information into input parameters for the multi-objective optimization model and inputting them into the multi-objective optimization model for calculation, includes:

[0032] The operating status parameters in the real-time device information are normalized into the state variables of the multi-objective optimization model;

[0033] The deviation value is mapped to the adjustment parameter of the objective function weight vector in the multi-objective optimization model. The adjusted state variable and weight vector are input to the solver of the multi-objective optimization model for iterative calculation. Among them, the energy cost weight of equipment that is determined to be in poor operating condition is reduced, and the carbon emission weight is increased.

[0034] In one embodiment, the step of generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model and controlling the operation of the energy equipment using the equipment scheduling strategy includes:

[0035] The calculation results are analyzed, and the start-up and shutdown status and output setpoint of the energy equipment for each time period are extracted from the calculation results to generate a sequence of equipment start-up and shutdown plan instructions.

[0036] Obtain the equipment operation constraints and system balance requirements of the energy equipment, and verify the feasibility of the equipment start-up and shutdown plan instruction sequence based on the equipment operation constraints;

[0037] After verification, the power balance difference for each time period is calculated according to the system balance requirements, and the equipment scheduling strategy is generated based on the equipment start-stop plan instruction sequence and the calculation result of the power balance difference.

[0038] In one embodiment, the step of generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model and controlling the operation of the energy equipment using the equipment scheduling strategy includes:

[0039] The device control instructions for the energy equipment are generated according to the device scheduling strategy, and an execution timestamp and priority flag are added to the device control instructions.

[0040] The device control command is sent to the energy device, and the execution process of the device control command is verified by real-time detection of the device control command response data.

[0041] In addition, to achieve the above objectives, this application also proposes an integrated energy collaborative management device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the integrated energy collaborative management method described above.

[0042] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the integrated energy collaborative management method described above.

[0043] One or more technical solutions proposed in this application have at least the following technical effects:

[0044] This invention collects multi-source heterogeneous data from energy equipment in an industrial park, constructs standardized feature inputs based on this data to obtain feature data, inputs the feature data into an adaptive prediction model for multi-time-scale load forecasting, constructs a multi-objective optimization model based on the load forecasting results, acquires real-time equipment information from the energy equipment, inputs this information into the multi-objective optimization model for calculation, generates an equipment scheduling strategy based on the calculation results of the multi-objective optimization model, and controls the operation of the energy equipment using the equipment scheduling strategy. Therefore, this application, through multi-source data fusion, load forecasting, and multi-objective optimization, achieves real-time and accurate scheduling of a comprehensive energy system. Through a closed-loop control mechanism combining hardware and software, it improves system response speed and operating efficiency, significantly reduces energy consumption and carbon emissions, and enhances the system's ability to cope with load fluctuations. Attached Figure Description

[0045] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart illustrating the first embodiment of the integrated energy collaborative management method of this application;

[0048] Figure 2 This is a detailed step diagram of step S40;

[0049] Figure 3 This is a detailed step diagram of step S50;

[0050] Figure 4 This is a schematic diagram of the functional modules based on the integrated energy management center;

[0051] Figure 5 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the integrated energy collaborative management method in this application embodiment.

[0052] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0053] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of this application and are not intended to limit this application.

[0054] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0055] The main solution of this application embodiment is as follows: collect multi-source heterogeneous data of energy equipment in industrial parks, construct standardized feature inputs based on the multi-source heterogeneous data to obtain feature data; input the feature data into an adaptive prediction model to perform multi-time-scale load prediction; construct a multi-objective optimization model based on the load prediction results; obtain real-time equipment information of the energy equipment, input the real-time equipment information into the multi-objective optimization model for calculation; generate an equipment scheduling strategy based on the calculation results of the multi-objective optimization model, and control the operation of the energy equipment with the equipment scheduling strategy.

[0056] Current integrated energy system management primarily employs a hierarchical control architecture, where a host computer system (such as an Energy Management System, EMS) performs strategy planning, supplemented by field controllers (such as PLCs and DCSs) to handle equipment start-up, shutdown, and status adjustment. These systems often rely on centralized optimization, periodic scheduling, and manually set rules to achieve supply-demand balance and equipment coordination across multiple energy sources. However, this approach has significant limitations: firstly, system response speed depends on the upper-level calculation and command issuance cycle, making it difficult to handle frequently fluctuating load scenarios; secondly, the separation of optimization algorithms and execution logic results in high deployment costs, significant communication latency, and insufficient system flexibility and real-time performance. Some research attempts to deploy real-time optimization algorithms on the server side, combining data acquisition and status awareness for rolling scheduling, but deep integration of scheduling strategies and control execution has not yet been achieved. Hardware and software rely on interface connections, lacking an integrated linkage mechanism. Furthermore, common hardware systems in current industrial applications (such as embedded controllers) mostly employ fixed-value control or priority start-up / shutdown strategies, making it difficult to support dynamic adjustment and reasoning capabilities for scheduling logic.

[0057] This application provides a solution that achieves real-time and precise scheduling of integrated energy systems through multi-source data fusion, load forecasting, and multi-objective optimization. A closed-loop control mechanism combining hardware and software improves system response speed and operating efficiency, significantly reduces energy consumption and carbon emissions, and enhances the system's ability to cope with load fluctuations.

[0058] Based on this, the embodiments of this application provide a comprehensive energy collaborative management method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the integrated energy collaborative management method of this application. In this embodiment, the integrated energy collaborative management method includes steps S10 to S50:

[0059] Step S10: Collect multi-source heterogeneous data of energy equipment in the industrial park, and construct standardized feature input based on the multi-source heterogeneous data to obtain feature data;

[0060] In this embodiment, relevant data generated by energy equipment in the industrial park during application is collected. Since the energy types indicated by the energy equipment are different, the relevant data generated by the energy equipment is defined as multi-source heterogeneous data. The collection of multi-source heterogeneous data is achieved through edge smart gateways deployed at each energy equipment node in the industrial park.

[0061] Specifically, the edge intelligent gateway has a built-in multi-protocol adapter, which supports the synchronous acquisition of analog signals of operating power, temperature, and pressure of the energy equipment, as well as switch status signals, meteorological monitoring data, and equipment vibration acoustic signals of the energy system formed by multiple energy equipment. After the multi-source heterogeneous data is initially time-stamped and compressed by the edge intelligent gateway, it is transmitted to the central data processing unit for processing via industrial Ethernet. The central data processing unit adopts a streaming data processing architecture to perform real-time parsing and cache management of the input data stream, i.e., the multi-source heterogeneous data. During the parsing process, the multi-source heterogeneous data is classified into two categories, continuous monitoring data and discrete event data, according to the data source and physical meaning of the multi-source heterogeneous data, and uniformly converted into a standardized data frame format with timestamps.

[0062] The converted multi-source heterogeneous data is identified using a Z-score-based statistical anomaly detection algorithm. Anomalies that significantly deviate from the historical distribution are marked based on the identification results. Subsequently, a dynamic threshold filtering mechanism is applied to remove pulse interference caused by instantaneous sensor failures from the standardized data frame format. Furthermore, missing data is filled using time series prediction. In this filling process, linear interpolation is used for rapid repair of short-term missing segments, while for long-term missing segments, an ARIMA model is introduced to reconstruct the data based on its own temporal patterns to ensure data continuity.

[0063] The cleaned multi-source heterogeneous data is fed into the feature extraction module for feature extraction. Specifically, the cleaned multi-source heterogeneous data is classified, and each type of classified multi-source heterogeneous data is preprocessed. For power data, the mean, variance, and gradient of change within the sliding window are calculated; for temperature and pressure data, the steady-state value and fluctuation coefficient are extracted; and for switch data, the distribution of state switching frequency and duration is statistically analyzed.

[0064] Furthermore, the feature extraction module integrates the multi-source heterogeneous data to calculate higher-order feature indicators. Specifically, it combines power and temperature data from the multi-source heterogeneous data to calculate the deviation rate between the real-time efficiency and rated efficiency of energy equipment; it integrates meteorological data and load data from the multi-source heterogeneous data to construct a cooling and heating load index; and it correlates vibration signals and operating power from the multi-source heterogeneous data to generate an equipment health status score. After normalization, the calculated higher-order feature indicators are aggregated into a feature matrix of a unified dimension according to a time window. Finally, principal component analysis is used to reduce dimensionality and remove redundancy, forming a standardized feature vector suitable for model input, i.e., the feature data.

[0065] Step S20: Input the feature data into the adaptive prediction model to perform multi-time-scale load prediction;

[0066] In this embodiment, the extracted feature data is processed by a multi-scale temporal segmentation module, dividing it into three types of input sequences: long-term, short-term, and ultra-short-term, corresponding to three decision levels: day-ahead scheduling, intraday rolling adjustment, and real-time control, respectively. The adaptive prediction model adopts a layered fusion deep learning architecture, consisting of a spatial feature extraction layer, a temporal dependency modeling layer, and a multi-task output layer. The spatial feature extraction layer uses a graph attention network (GAT) to model the topological connections of energy equipment in the industrial park, dynamically capturing the spatial coupling characteristics of energy flow by calculating the attention weights between nodes. The temporal dependency modeling layer employs a coupled architecture of an ensemble learning method (XGBoost) and a self-attention model (Transformer). XGBoost provides strong feature representation for the model by performing high-dimensional feature engineering and importance filtering on historical sequences; the Transformer utilizes its multi-head self-attention mechanism to focus on the global dependencies of the input sequences in parallel, effectively capturing complex correlations between different time points. Based on this, the long-term, short-term and ultra-short-term sequences after feature enhancement are sequentially input into the adaptive prediction model for joint inference, and the corresponding multi-scale prediction results are generated through the task output layer.

[0067] Specifically, the multi-task output layer comprises three parallel prediction sub-networks, corresponding to the prediction tasks of electricity, heat, and cooling loads, respectively. Each sub-network employs a combination of a fully connected layer and a gating mechanism to generate load prediction values ​​and corresponding probability distributions for multiple future time points. The adaptive prediction model introduces a meta-learning mechanism to dynamically adjust model parameters: it periodically detects prediction errors and changes in the energy network topology; when the deviation exceeds a set threshold or a device switching event is detected, it triggers a model fine-tuning process, updating the node embedding vectors in the graph attention network and the weight parameters of the temporal network online based on the latest data.

[0068] During the prediction process, the adaptive prediction model receives preprocessed standardized feature vectors, namely long-term, short-term, and ultra-short-term input sequences. Based on these input sequences, spatial feature aggregation is performed in a graph attention network to generate node representations containing topological information. Subsequently, the node representations are modeled along with temporal data to generate structured prediction results through a multi-task output layer.

[0069] In addition, before the step of inputting the feature data into the adaptive prediction model for multi-time-scale load forecasting, the method further includes:

[0070] Training samples are generated based on historical load data and external environmental characteristics, and a multi-task deep learning prediction model is trained based on the training samples.

[0071] The training results are optimized using a self-supervised learning method, and the adaptive prediction model is obtained based on the optimization results.

[0072] In this embodiment, before inputting the feature data into the adaptive prediction model for multi-timescale load prediction, the training and optimization of the multi-task deep learning prediction model are first completed. The process of generating training samples based on historical load data and external environmental features specifically includes: collecting at least two years of historical load data and corresponding meteorological and date-type data to construct an original sample set; using a sliding window method to perform time-series segmentation on the original sample set to generate sample sequences with fixed-length input sequences and output labels; and performing data augmentation operations on the sample sequences by adding Gaussian noise, random masks, and time-series interpolation methods to expand sample diversity. The final generated training sample set includes an input feature matrix and corresponding multi-label outputs for electricity, heat, and cooling loads.

[0073] Step S30: Construct a multi-objective optimization model based on the load forecast results;

[0074] In this embodiment, a multi-objective optimization model is constructed using the output of the multi-objective adaptive model, namely the load forecast results. The construction of this multi-objective optimization model is essentially based on deep analysis and multi-dimensional feature fusion of the load forecast results.

[0075] Specifically, firstly, key parameters related to the electricity, heat, and cooling load forecast curves are extracted from the load forecast results. These key parameters include peak load, load change gradient, duration, and energy grade demand. Based on these key parameters, a basic function, a cost function, and a carbon emission function are obtained. This process can be defined as the steps in constructing a multi-objective optimization model based on the load forecast results, including:

[0076] The load forecast results are analyzed, and a basic function, a cost function, and a carbon emission function are generated based on the analysis results.

[0077] The basic function, cost function, and carbon emission function are combined to generate an optimization objective function, and the multi-objective optimization model is constructed based on the optimization objective function;

[0078] The steps of parsing the load forecast results and generating the basic function, cost function, and carbon emission function based on the parsing results include:

[0079] Extract the prediction results based on electricity, heat and cooling load from the analysis results, calculate the total energy input of coupled energy demand in the future time period based on the prediction results, and generate the basic function based on the total energy input;

[0080] The load curve and time-of-use energy price signal are extracted from the analysis results, and the operation and maintenance costs required for equipment start-up, shutdown and operating condition switching are obtained to generate the cost function.

[0081] The load demand and equipment carbon emission characteristics are extracted from the analysis results. Based on the nonlinear relationship between the equipment start-up and shutdown process, load rate and the equipment carbon emission characteristics, a dynamic carbon emission function is generated.

[0082] In this embodiment, the construction of the multi-objective optimization model is based on the refined analysis and multi-dimensional reconstruction of the load forecast results. First, based on the load forecast results, spatiotemporal feature analysis based on the classification of electricity, heat, and cooling loads is performed, and the electricity curve, heat curve, and cooling load curve for each time period are extracted.

[0083] Subsequently, a dynamic full life cycle cost modeling method is used to calculate the cost function. In this calculation process, a basic energy cost curve needs to be generated based on the load curve and time-of-use energy price signal in the load forecast results. Then, an equipment start-up and shutdown cost model is introduced. The equipment start-up and shutdown model comprehensively considers energy loss, lifespan depreciation and maintenance requirements during the equipment start-up and shutdown process to obtain the maintenance cost coefficient of the energy equipment. At the same time, a sub-model of equipment operating condition switching cost is established to quantify the maintenance cost coefficient of the energy equipment under different load rates. Based on the quantification results, a comprehensive cost function is finally formed that includes energy purchase cost, operation and maintenance cost and equipment depreciation cost.

[0084] Finally, a carbon emission function is constructed using a dynamic carbon emission factor method based on equipment operating status. This construction process involves first establishing an equipment-level carbon emission characteristic database, which contains instantaneous carbon emission intensity curves for different energy devices during start-up and shutdown processes and under different load rates. Based on the load demand and equipment carbon emission characteristics in the load prediction results, the corresponding instantaneous carbon emission intensity curves are obtained from the carbon emission characteristic database. Then, an equipment operating status simulation algorithm is used to generate the expected operating conditions for each energy device from these instantaneous carbon emission intensity curves. Finally, the expected operating conditions are coupled with the corresponding dynamic carbon emission factor to generate a nonlinear carbon emission function that takes into account the transient processes of equipment start-up and shutdown and changes in load rate.

[0085] Finally, a hierarchical weighted aggregation strategy is used to fuse the aforementioned basic functions, cost functions, and carbon emission functions to obtain the multi-objective optimization model. In this fusion process, the basic functions are first used as hard constraints to ensure the system's energy supply and demand balance. Then, an objective weight allocation method based on entropy weighting is used, combined with real-time energy price signals and carbon emission quotas, to dynamically determine the weight coefficients of the cost and carbon emission functions. Finally, a weighted summation method is used to transform the multi-objective problem into a single-objective optimization problem, while retaining an adjustable weight mechanism to generate a Pareto optimal solution set. The final model includes constraints such as system power balance, equipment operation limitations, and pipeline power flow safety, forming a complete mixed-integer nonlinear programming model.

[0086] The final multi-objective optimization model contains a complete system of constraints: in addition to basic physical constraints such as system power balance and equipment ramp rate, it also includes constraints on equipment operation reliability, energy network power flow, and robustness based on prediction uncertainty, providing a complete mathematical framework for subsequent optimization calculations.

[0087] Furthermore, the step of fusing the basic function, cost function, and carbon emission function to generate an optimization objective function, and constructing the multi-objective optimization model based on the optimization objective function, includes:

[0088] The weighting coefficients of the cost function and carbon emission function are determined based on the preset operating cost target and carbon emission target, and the weighting coefficients are dynamically adjusted based on real-time energy price signals and carbon emission quotas.

[0089] The cost function and the carbon emission function are normalized, and the weighted summation method is used to merge the normalized cost function and carbon emission function into a comprehensive objective function.

[0090] The multi-objective optimization model is generated using the integrated objective function and the basic function.

[0091] In this embodiment, a dynamic weighted fusion method based on multi-objective decision theory is used to construct the multi-objective optimization model. First, a dynamic decision-making mechanism for weight allocation coefficients is established, constructing a weight decision-maker based on two-layer optimization. The upper-layer optimization aims to balance operating cost and carbon emission targets, while the lower-layer optimization generates weight adjustment strategies based on real-time energy price signals and dynamic changes in carbon emission quotas. Specifically, by monitoring the time-of-use electricity price fluctuation trend in the real-time electricity market and the proportion of remaining carbon emission quotas in the system, a fuzzy inference system is used to generate preliminary weight allocation coefficients. Then, a feedback correction mechanism is used to fine-tune the weights, ensuring that the weight coefficients simultaneously reflect short-term economic pressures and long-term carbon emission constraints.

[0092] A normalization method based on historical extreme values ​​and dynamic updates is used to normalize the cost function and carbon emission function. Specifically, for the cost function, the maximum and minimum values ​​of recent operating costs are used as normalization boundaries, and exponential smoothing is used to dynamically update these boundary values ​​to adapt to changes in system operating conditions. For the carbon emission function, a relative normalization method based on carbon emission quota ratios is employed to convert actual carbon emissions into a completion index relative to the quota. The normalized cost function and carbon emission function are then fused using a weighted summation method, where the weighting coefficients are updated in real time according to the aforementioned dynamic decision-making mechanism, forming a comprehensive objective function.

[0093] Furthermore, when constructing the complete multi-objective optimization model, the fundamental functions are treated as constraints. Specifically, the total energy input calculated based on grade coupling is used as an equality constraint to ensure the energy supply and demand balance of the system; simultaneously, physical constraints such as equipment operation limitations and pipeline transmission capacity are used as inequality constraints. The final generated optimization model is a dynamic optimization problem with time-varying weight coefficients, mathematically expressed as solving an equipment scheduling scheme that minimizes the comprehensive objective function under the constraints defined by the fundamental functions. This model also introduces a robust optimization mechanism, generating an optimization decision scheme with anti-interference capabilities by considering the uncertainty of load forecasting.

[0094] Step S40: Obtain the real-time equipment information of the energy equipment, and input the real-time equipment information into the multi-objective optimization model for calculation;

[0095] In this embodiment, the real-time equipment information of the energy devices is collected and preprocessed through distributed edge computing nodes. These edge computing nodes are deployed at each energy device site and collect multi-dimensional status parameters, including output power, operating efficiency, temperature, and vibration frequency, as raw data in real time via industrial IoT protocols. The collected raw data undergoes local preprocessing at the edge, employing a Kalman filter algorithm to eliminate measurement noise and a time series alignment mechanism to ensure the synchronization of multi-source data within the raw data. The preprocessed equipment status information is then transmitted to the central optimization decision-making unit via high-speed industrial Ethernet.

[0096] A fuzzy inference-based evaluation method is used to compare and analyze the real-time equipment information with preset performance thresholds. Specifically, an equipment health status evaluation model is established and the preset performance thresholds are configured. The real-time operating parameters are compared with the equipment's rated parameters and historical operating baselines within the configured preset performance thresholds in a multi-dimensional manner to generate a comprehensive evaluation index that includes performance deviation, health index, and predicted remaining service life. This comprehensive evaluation index is used to dynamically reconstruct the constraint system of the multi-objective optimization model. Specifically, for energy equipment in good health, an adaptive constraint relaxation strategy is adopted to expand the feasible domain of its operating parameters to improve optimization flexibility. For energy equipment with performance degradation, a chance-constraint programming method based on reliability theory is introduced to convert its operating constraints into probabilistic form and add equipment availability guarantee conditions.

[0097] The revised constraints and real-time equipment parameters are input into an improved multi-objective optimization solver. This solver employs a decomposition-based multi-objective evolutionary algorithm and incorporates a specific constraint handling mechanism and particle update strategy tailored to the problem characteristics. A rolling optimization framework is introduced during the solution process, executing optimization calculations forward within a fixed time window. Each optimization update updates the model parameters and constraints based on the latest equipment status information. The scheduling scheme generated by the optimization calculations is simultaneously fed back to the equipment health management module, forming a closed-loop optimization control system.

[0098] Step S50: Generate an equipment scheduling strategy based on the calculation results of the multi-objective optimization model, and control the operation of the energy equipment using the equipment scheduling strategy.

[0099] In this embodiment, the calculation results of the multi-objective optimization model are processed by a distributed decision engine to generate an equipment scheduling strategy. Within the distributed decision engine, a multi-criteria decision analysis is performed on the Pareto optimal solution set from the calculation results to select the optimal scheduling scheme. The selection process comprehensively considers economic operating costs, carbon emission levels, and system reliability indicators to generate an optimal decision that balances multiple objectives.

[0100] When generating equipment-level control instructions based on the selected optimal scheduling scheme, a hierarchical instruction compilation mechanism is adopted. Specifically, this mechanism includes an upper-level instruction compiler and a lower-level instruction converter. The upper-level instruction compiler decomposes the selected optimal scheduling scheme into equipment group control strategies, and the lower-level instruction converter converts these strategies into specific equipment control parameters. An instruction sequence is formed based on these control parameters. This sequence includes unit start-up and shutdown timings, power setpoint curves, operating mode switching logic, and equipment collaborative operation parameters. Before issuing any equipment instructions from this sequence, virtual verification is performed using a digital twin system. This system constructs a virtual model consistent with the physical system, predicting the instruction execution effect through simulation to ensure the safety and feasibility of the equipment instruction execution.

[0101] The verified device commands are sent to the local controllers of each energy device via a time-sensitive network protocol. During the execution of these commands, a high-precision condition monitoring system collects device response data in real time, including actual output power, operating status, and efficiency parameters. The monitoring data is compared and analyzed in real time with the expected state. A Kalman filter-based state estimator is used to eliminate measurement noise and accurately identify the deviation between the actual and expected system states.

[0102] When a significant deviation is detected, a multi-level response mechanism is triggered: for minor deviations, the local controller adaptively adjusts based on preset rules; for larger deviations, the online optimization module is activated for local re-optimization; when the deviation exceeds a safety threshold or a device fault is detected, a global re-optimization process is immediately triggered, resolving the multi-objective optimization model based on the latest system state and generating updated scheduling instructions. Simultaneously, all deviation events and their handling processes are recorded in the system knowledge base for the self-learning and improvement of the optimization model.

[0103] In addition, the step of generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model and controlling the operation of the energy equipment with the equipment scheduling strategy includes:

[0104] The device control instructions for the energy equipment are generated according to the device scheduling strategy, and an execution timestamp and priority flag are added to the device control instructions.

[0105] The device control command is sent to the energy device, and the execution process of the device control command is verified by real-time detection of the device control command response data.

[0106] In this embodiment, when the device scheduling strategy is converted into specific device control commands by the command generation module, the command generation module first parses the optimal scheduling scheme output by the multi-objective optimization model, decomposing the optimal scheduling scheme into independent control commands for each energy device. These independent control commands include power setpoints, operating mode commands, and start / stop timing parameters. Subsequently, an execution timestamp accurate to milliseconds is added to each control command to ensure improved timing accuracy of coordinated operations between energy devices when executing the device commands. Simultaneously, different priority markers are assigned to each energy device based on its device type and its role in the system, with critical device commands set to high priority to ensure the stable execution of the system's core functions.

[0107] Before issuing the device commands, security verification is performed in a virtual execution environment. This virtual execution environment uses digital twin technology to construct a system simulation model and predict system state changes after command execution. Verified device commands are then issued to the local controllers of each energy device via a time-sensitive network protocol, ensuring the real-time and deterministic nature of command transmission. The command transmission process employs an encryption and authentication mechanism to guarantee the integrity and security of the device commands. Based on the energy management indicated in the above embodiment, one can also view... Figure 4 , Figure 4 This is a schematic diagram of the functional modules based on the integrated energy management center.

[0108] During the execution of the device commands, the response status of each energy device is monitored in real time through a high-frequency data acquisition system. The monitored data includes actual output power, device operating status, and efficiency parameters. This data is processed by a state estimator, and a Kalman filter algorithm is used to eliminate measurement noise, obtaining an accurate picture of the actual device operating status. Subsequently, the acquired actual operating status is compared with the expected status in real time to establish an anomaly detection mechanism based on confidence intervals. When a deviation exceeding the allowable range is detected, a command retransmission mechanism is immediately triggered, or a local optimization adjustment process is initiated. All command execution processes and device response data are recorded in the operation log, providing data support for system performance analysis and optimization.

[0109] You can also view Figure 2 , Figure 2 This is a detailed step diagram of step S40, based on the diagram shown. Figure 2 The step of obtaining real-time equipment information of the energy equipment and inputting the real-time equipment information into the multi-objective optimization model for calculation includes steps S41-43:

[0110] Step S41: Collect real-time device information of the energy equipment, and calculate the deviation value and current health index of the energy equipment based on the real-time device information and a preset calculation threshold;

[0111] Step S42: Configure the constraints of the multi-objective optimization model based on the deviation value, and dynamically adjust the constraints based on the current health index;

[0112] Step S43: The current health index is converted into a device reliability coefficient and embedded into the device information, and the device information is converted into an input parameter of the multi-objective optimization model and input into the multi-objective optimization model for calculation.

[0113] In this embodiment, real-time equipment information is collected by intelligent sensing units deployed at each energy equipment node. This real-time equipment information includes multi-dimensional parameters such as output power, operating efficiency, vibration spectrum, temperature distribution, and insulation status. The collected real-time equipment information is preprocessed by edge computing nodes, and a Kalman filter algorithm is used to eliminate measurement noise. Time alignment is used to ensure data consistency.

[0114] The preprocessed real-time equipment information is input into a health assessment model. This model compares the real-time parameters in the equipment information with the equipment's rated parameters and historical operating baselines, calculating multiple deviation values, including output power deviation rate, efficiency deviation coefficient, and vibration characteristic difference. Based on these deviation values, a deep belief network (DBN) is used to construct the health assessment model. Through multi-layer nonlinear transformations, multi-source features are fused, ultimately outputting a current health index ranging from 0 to 1 to characterize the energy equipment's current health. This current health index comprehensively reflects the degree of performance degradation of the energy equipment.

[0115] The constraints of the multi-objective optimization model are dynamically configured based on the calculated deviation values ​​and health index. For energy equipment with deviation values ​​within the allowable range, a constraint relaxation strategy is adopted to expand the upper and lower limit constraint ranges proportionally to its health index. For example, the power upper limit constraint of an energy equipment with a health index of 0.8 is relaxed to 105% of the rated value. For energy equipment with deviation values ​​exceeding the limit or health indices below the threshold, a chance-constrained programming method is introduced to convert its operating constraints into probabilistic forms, such as requiring that the probability of its output power meeting the requirement is not less than the health index value. Simultaneously, a dynamic constraint adjustment mechanism is established to update the relaxation coefficients and probability confidence levels of the constraint parameters for each energy equipment in real time according to the changing trend of the health index.

[0116] Subsequently, the health index is transformed into equipment reliability coefficients and embedded into the optimization model. Specifically, based on the mapping relationship between the health index and equipment failure rate, the instantaneous failure probability of each energy device is calculated, thereby deriving the availability coefficient of the energy device. The availability coefficient is embedded into the equipment information and transformed into input parameters of a multi-objective optimization model. Specifically, a maintenance cost adjustment term based on the reliability coefficient is introduced into the cost function, and the carbon emission increment of reserve capacity caused by equipment failure risk is included in the carbon emission function. Finally, the equipment information of the energy devices containing the reliability coefficient is input into an improved multi-objective optimization solver, and an evolutionary algorithm with a time-varying constraint handling mechanism is used for rolling optimization calculation to ensure that the optimization result not only meets the current operating requirements but also fully considers the impact of equipment health status on the long-term operational reliability of the system.

[0117] Furthermore, the step of converting the current health index into a device reliability coefficient and embedding it into the device information, and converting the device information into input parameters for the multi-objective optimization model and inputting them into the multi-objective optimization model for calculation, includes:

[0118] The operating status parameters in the real-time device information are normalized into the state variables of the multi-objective optimization model;

[0119] The deviation value is mapped to the adjustment parameter of the objective function weight vector in the multi-objective optimization model. The adjusted state variable and weight vector are input to the solver of the multi-objective optimization model for iterative calculation. Among them, the energy cost weight of equipment that is determined to be in poor operating condition is reduced, and the carbon emission weight is increased.

[0120] In this embodiment, the operating status parameters in the real-time equipment information undergo standardized preprocessing. The Min-Max normalization method is used to transform each dimension of the parameters to the [0,1] interval, forming the standard state variables of the multi-objective optimization model. During normalization, a differentiated processing strategy is adopted for different types of operating status parameters. Based on this strategy, power-related operating status parameters are normalized according to rated capacity, temperature-related operating status parameters are normalized according to allowable operating range, and vibration-related operating status parameters are normalized according to safety threshold to obtain the state variables for each operating status parameter. The normalized state variables retain the physical meaning and dimensional relationships of the original parameters, providing standardized input data for subsequent optimization calculations.

[0121] Furthermore, the deviation value is mapped to adjustment parameters of the objective function weight vector through a fuzzy logic controller. A dual-input fuzzy inference system based on the degree of deviation and the health index is pre-established, and a corresponding fuzzy rule base is set. In the fuzzy rule base, when the health index of the energy equipment is lower than a first preset health threshold and the power deviation rate exceeds a first power deviation threshold, it is determined to be in poor operating condition, and a weight adjustment instruction is generated. The energy cost weight coefficient of the energy equipment is reduced by the first energy cost weight threshold, for example, by 20%-30%, while the carbon emission weight coefficient of the energy equipment is increased by the first carbon emission weight threshold, for example, by 15%-25%. A smooth transition mechanism is adopted in this weight adjustment process to avoid oscillations in the optimization results caused by sudden weight changes.

[0122] The adjusted state variables and weight vectors are input together into the solver of the multi-objective optimization model for iterative calculation. The solver employs an improved NSGA-II algorithm, dynamically adjusting the population search direction during iteration and prioritizing the optimization of objective functions with higher weight coefficients. For energy equipment operating poorly, the energy cost weight is reduced to guide the optimization algorithm to decrease task allocation to that equipment; simultaneously, the carbon emission weight is increased to strengthen the penalty for the environmental impact of such equipment, achieving a rebalancing of economic efficiency and environmental protection. The final output is a multi-objective optimization scheme that takes into account the actual operating conditions of the equipment.

[0123] You can also view Figure 3 , Figure 3 This is a detailed step diagram of step S50, based on the diagram shown. Figure 3 The step of generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model and controlling the operation of the energy equipment with the equipment scheduling strategy includes steps S51-53:

[0124] Step S51: Analyze the calculation results and extract the start-stop status and output setting value of the energy equipment in each time period from the calculation results to generate a sequence of equipment start-stop plan instructions;

[0125] Step S52: Obtain the equipment operation constraints and system balance requirements of the energy equipment, and verify the feasibility of the equipment start-up and shutdown plan instruction sequence based on the equipment operation constraints;

[0126] Step S53: After the verification is completed, calculate the power balance difference for each time period according to the system balance requirements, and generate the equipment scheduling strategy based on the equipment start-stop plan instruction sequence and the calculation result of the power balance difference.

[0127] In this embodiment, the output results of the multi-objective optimization model are subjected to in-depth analysis and instruction sequence generation. Specifically, a multi-dimensional deconstruction algorithm is first used to analyze and optimize the output results, and the operating status parameters of all energy devices in future time periods are extracted from the optimized output results. The operating status parameters include unit start-up and shutdown status, output setpoint, operating mode, and collaborative work instructions. When generating the equipment start-up and shutdown plan instruction sequence based on the analysis results of the output results, a timing optimization and arrangement method is used to establish an instruction timing conflict detection mechanism to ensure that the minimum time interval requirement is met between equipment start-up and shutdown instructions; and a load smoothing algorithm is used to optimize the output setpoint curve to avoid drastic power fluctuations; at the same time, equipment collaborative operation parameters are generated, including the load distribution ratio and synchronous operation accuracy requirements among multiple energy devices. The generated instruction sequence contains complete time tags, instruction priorities, and execution condition judgment logic, forming a set of equipment control instructions that can be directly issued.

[0128] Based on the generated instruction sequence, a multi-constraint instruction feasibility verification system is constructed to verify the feasibility of the device instructions in the instruction sequence. Specifically, detailed operating constraint parameters for each energy device are obtained from the device management database. These detailed operating constraint parameters include minimum start-up and shutdown time, maximum ramp rate, upper and lower output limits, allowable operating area, and coupling constraints between devices. Simultaneously, real-time system balance requirements, including power balance tolerance, frequency stability range, and voltage safety boundary, are obtained from the system monitoring platform. Based on the detailed operating constraint parameters and the implemented system balance requirements as constraints, a multi-layer verification model is established. In this multi-layer verification model, the first layer performs device-level constraint verification to verify whether each device instruction in the instruction sequence violates its independent operating limitations; the second layer performs system-level constraint verification, using power flow calculation to verify whether the instruction sequence meets the system's safe operation requirements; and the third layer performs economic verification to assess whether the execution cost of the instruction sequence is within budget. During the verification process, digital twin technology is used for virtual execution simulation to predict system state changes after the instruction sequence is executed.

[0129] After verification, a refined scheduling strategy is generated. Based on the verified instruction sequence, a power balance model constructed using a real-time power balance algorithm is employed to calculate the power balance difference for each time period. When processing the instruction sequence through the power balance model, line losses, transformer losses, and reactive power effects are taken into account. State estimation techniques are used to improve calculation accuracy, and Kalman filtering is used to eliminate measurement noise. Based on the calculated power balance difference, a coordinated control mechanism is activated: for small power differences, the output setpoints of operating equipment are adjusted to achieve balance; for large power differences, a standby capacity call-up strategy is activated, including rapid activation of standby units, adjustment of interruptible loads, and call-up of energy storage devices. The final generated equipment scheduling strategy includes two schemes: the primary scheme is directly generated based on optimization results, while the standby scheme is designed considering system uncertainties to ensure safe and stable system operation under various operating conditions. All scheduling strategies are issued and executed after security authentication, and a comprehensive feedback and adjustment mechanism is established.

[0130] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the integrated energy collaborative management method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0131] This application provides an integrated energy collaborative management device, the integrated energy collaborative management device including a controller, the controller including: 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, the instructions being executed by the at least one processor to enable the at least one processor to execute the integrated energy collaborative management method in the above embodiment one.

[0132] The following is for reference. Figure 5 It shows a schematic diagram of a controller suitable for implementing embodiments of this application. Figure 5As shown, the controller may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for controller operation. The processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. The communication device 1009 allows the controller to communicate wirelessly or wiredly with other devices to exchange data.

[0133] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0134] The integrated energy collaborative management device provided in this application, employing the integrated energy collaborative management method described in the above embodiments, can solve the technical problems of real-time optimization scheduling and closed-loop control in existing integrated energy management. Compared with the prior art, the beneficial effects of the integrated energy collaborative management provided in this application are the same as those of the integrated energy collaborative management method provided in the above embodiments, and other technical features in this integrated energy collaborative management are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0135] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0136] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. In some variant embodiments, it can also be applied to electrically operated reclining furniture such as electric sofas, and is not limited to integrated energy collaborative management. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0137] This application provides a storage medium, which is a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the integrated energy collaborative management method in the above embodiments.

[0138] The computer-readable storage medium provided in this application can be any medium that tangibly stores program code that can be executed by a computing device (such as a processor, controller, etc.). Such media include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0139] The aforementioned computer-readable storage medium may be included in the integrated energy coordination management controller; or it may exist independently and not be assembled into the controller.

[0140] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the controller, enable the controller to implement the technical content of the integrated energy collaborative management method embodiment shown above.

[0141] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages—such as Java, C++, Python, and Smalltalk—as well as conventional procedural programming languages—such as C, Python, etc., or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0142] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0143] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0144] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., computer programs) for executing the above-described integrated energy collaborative management method, which can solve the technical problems of real-time optimization scheduling and closed-loop control in existing integrated energy management. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the integrated energy collaborative management method provided in the above embodiments, and will not be repeated here.

Claims

1. A comprehensive energy collaborative management method, characterized in that, The integrated energy collaborative management method includes the following steps: Collect multi-source heterogeneous data of energy equipment in industrial parks, and construct standardized feature inputs based on the multi-source heterogeneous data to obtain feature data; The feature data is input into an adaptive prediction model for multi-time-scale load prediction. The process of constructing a multi-objective optimization model based on load forecast results includes: analyzing the load forecast results and generating a basic function, a cost function, and a carbon emission function based on the analysis results; fusing the basic function, cost function, and carbon emission function to generate an optimization objective function, and constructing the multi-objective optimization model based on the optimization objective function; wherein, the step of analyzing the load forecast results and generating the basic function, cost function, and carbon emission function based on the analysis results includes: extracting forecast results based on electricity, heat, and cooling loads from the analysis results; calculating the total energy input of coupled energy demand in the future time period based on the forecast results; generating the basic function based on the total energy input; extracting load curves and time-of-use energy price signals from the analysis results; obtaining the operation and maintenance costs required for equipment start-up, shutdown, and operating condition switching; and extracting load demand and equipment carbon emission characteristics from the analysis results, generating the dynamic carbon emission function based on the nonlinear relationship between equipment start-up and shutdown process, load rate, and equipment carbon emission characteristics. The process involves acquiring real-time equipment information of the energy equipment and inputting this information into a multi-objective optimization model for calculation. Specifically, the process includes collecting real-time equipment information of the energy equipment and calculating the deviation value and current health index of the energy equipment based on this information and a preset calculation threshold. Constraints of the multi-objective optimization model are configured based on the deviation value, and these constraints are dynamically adjusted based on the current health index. The current health index is then converted into an equipment reliability coefficient and embedded into the equipment information. This equipment information is then converted into input parameters for the multi-objective optimization model and input into the model for calculation. The steps of converting the current health index into a device reliability coefficient and embedding it into the device information, and converting the device information into input parameters of the multi-objective optimization model and inputting them into the multi-objective optimization model for calculation, include: normalizing the operating status parameters in the real-time device information into state variables of the multi-objective optimization model; mapping the deviation value into adjustment parameters of the objective function weight vector in the multi-objective optimization model; and inputting the adjusted state variables and weight vector into the solver of the multi-objective optimization model for iterative calculation, wherein the energy cost weight of the device determined to be in poor operating condition is reduced, and the carbon emission weight is increased; Based on the calculation results of the multi-objective optimization model, an equipment scheduling strategy is generated, and the operation of the energy equipment is controlled by the equipment scheduling strategy. The step of dynamically adjusting the constraints based on the current health index includes: For energy equipment in good health, an adaptive constraint relaxation strategy is adopted to expand the feasible domain of its operating parameters to improve optimization flexibility; for energy equipment with performance degradation, an opportunity constraint programming method based on reliability theory is introduced to convert its operating constraints into probabilistic form and add equipment availability guarantee conditions.

2. The integrated energy collaborative management method as described in claim 1, characterized in that, Before the step of inputting the feature data into the adaptive prediction model for multi-time-scale load forecasting, the method further includes: Training samples are generated based on historical load data and external environmental characteristics, and a multi-task deep learning prediction model is trained based on the training samples. The training results are optimized using a self-supervised learning method, and the adaptive prediction model is obtained based on the optimization results.

3. The integrated energy collaborative management method as described in claim 1, characterized in that, The step of integrating the basic function, cost function, and carbon emission function to generate an optimization objective function, and constructing the multi-objective optimization model based on the optimization objective function, includes: The weighting coefficients of the cost function and carbon emission function are determined based on the preset operating cost target and carbon emission target, and the weighting coefficients are dynamically adjusted based on real-time energy price signals and carbon emission quotas. The cost function and the carbon emission function are normalized, and the weighted summation method is used to merge the normalized cost function and carbon emission function into a comprehensive objective function. The multi-objective optimization model is generated using the integrated objective function and the basic function.

4. The integrated energy collaborative management method as described in claim 1, characterized in that, The step of generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model, and controlling the operation of the energy equipment using the equipment scheduling strategy, includes: The calculation results are analyzed, and the start-up and shutdown status and output setpoint of the energy equipment for each time period are extracted from the calculation results to generate a sequence of equipment start-up and shutdown plan instructions. Obtain the equipment operation constraints and system balance requirements of the energy equipment, and verify the feasibility of the equipment start-up and shutdown plan instruction sequence based on the equipment operation constraints; After verification, the power balance difference for each time period is calculated according to the system balance requirements, and the equipment scheduling strategy is generated based on the equipment start-stop plan instruction sequence and the calculation result of the power balance difference.

5. The integrated energy collaborative management method as described in claim 1, characterized in that, The step of generating an equipment scheduling strategy based on the calculation results of the multi-objective optimization model, and controlling the operation of the energy equipment using the equipment scheduling strategy, includes: The device control instructions for the energy equipment are generated according to the device scheduling strategy, and an execution timestamp and priority flag are added to the device control instructions. The device control command is sent to the energy device, and the execution process of the device control command is verified by real-time detection of the device control command response data.

6. A comprehensive energy collaborative management device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the integrated energy collaborative management method as described in any one of claims 1 to 5.

7. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the integrated energy collaborative management method as described in any one of claims 1 to 5.

Images