Deep learning-based intelligent analysis method for electric power energy load of deep sea aquaculture

Abstract

The invention relates to the technical field of electric power energy load analysis, in particular to an intelligent analysis method for electric power energy load of deep and far sea aquaculture based on deep learning. The method comprises the following steps: acquiring electric power multi-energy equipment data and user power consumption data of the deep and far sea culture electric power energy system; based on the electric power multi-energy equipment data and the user power consumption data, establishing an optimized operation light-diesel-storage electric power energy integration model to collect electric power multi-energy real-time operation load data; acquiring multi-modal operation sensing element data of the deep and far sea culture electric power energy system; establishing a trend prediction model of operation load transformation based on the multi-modal operation perception element data, and generating an operation load transformation trend prediction model; and transmitting the electric power multi-energy real-time operation load data to the operation load transformation trend prediction model to analyze the electric power energy operation load trend characteristic data. Accurate prediction of the operation load of the deep and far sea culture electric power energy system is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy load analysis, and in particular to an intelligent analysis method for deep-sea aquaculture electric energy load based on deep learning. Background Art

[0002] In the current development of the energy sector, the application scope of deep learning technology continues to expand, playing an increasingly important role in many key areas such as energy resource management, energy consumption forecasting, and intelligent scheduling. Among the many aspects of energy management, load forecasting and intelligent scheduling are of paramount importance. They are crucial for improving energy utilization, reducing energy consumption, promoting energy conservation, and achieving environmental protection. The coordinated operation of multiple devices and energy sources in deep-sea aquaculture power energy management requires more sophisticated load management. Leveraging algorithmic models to dynamically predict load trends, optimize energy scheduling, and provide early warning of load peaks or abnormal fluctuations, this technology will promote the development of intelligent and green deep-sea aquaculture. However, existing intelligent analysis technologies for deep-sea aquaculture power energy loads ignore the strong nonlinear characteristics of power operation due to the strong coupling between the subsystems of the platform energy management system. Furthermore, they fail to account for the unique operating conditions of aquaculture platforms, such as feeding and fishing. Consequently, they fail to reflect the high variability of loads, resulting in insufficient accuracy in aquaculture power energy load forecasting. Summary of the Invention

[0003] Based on this, the present invention provides an intelligent analysis method for deep-sea aquaculture power energy load based on deep learning to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a deep-sea aquaculture power energy load intelligent analysis method based on deep learning includes the following steps: Step S1: Acquire the power multi-energy equipment data and user power consumption data of the deep-sea aquaculture power energy system; perform power consumption characteristic analysis on the user power consumption data to generate user power consumption characteristic data; Step S2: establishing a demand-target operation mapping relationship specific to the power multi-energy device based on the power multi-energy device data, and generating a demand-target operation power energy integration model; establishing a solar-diesel-storage power energy integration model for optimizing the user's power consumption characteristics based on the user's power consumption characteristic data and the demand-target operation power energy integration model, and generating an optimized operation solar-diesel-storage power energy integration model; Step S3: Acquire multimodal operation perception element data of the deep-sea aquaculture power energy system; perform a characteristic analysis of the factors influencing the power multi-energy operation load based on the multimodal operation perception element data to generate characteristic data of the factors influencing the power multi-energy operation load; establish a trend prediction model for the operation load transformation based on the characteristic data of the factors influencing the power multi-energy operation load to generate a trend prediction model for the operation load transformation; Step S4: Collect the real-time operating load data of multiple power sources by optimizing the operation of the solar-diesel-storage power energy integration model; transmit the real-time operating load data of multiple power sources to the operating load transformation trend prediction model to perform power energy operating load trend characteristic analysis and generate power energy operating load trend characteristic data; and execute the power energy load coordinated control operation of deep-sea aquaculture based on the power energy operating load trend characteristic data.

[0005] Furthermore, step S1 includes the following steps: Step S11: Acquire the power multi-energy equipment data and user power consumption data of the deep-sea aquaculture power energy system; Step S12: Analyze the power consumption period according to the user power consumption data to generate power consumption period data, and perform time series slicing of the power load according to the time difference of the user power consumption data based on the power consumption period data to generate time series slicing data of the user power load according to the time difference of the period; Step S13: extracting load time series characteristics based on the time series slice data of the user's electricity load in different periods to obtain the time series characteristic data of the user's electricity load in different periods; Step S14: performing a cluster analysis of user electricity consumption behaviors based on the time series characteristic data of user electricity loads in different periods to generate user electricity consumption behavior cluster data; Step S15: performing power usage pattern correlation feature analysis based on the user power usage behavior clustering data to generate power usage pattern correlation feature data; Step S16: performing a user power consumption characteristic analysis on the time series characteristic data of the user power load in different periods by using the power consumption pattern association characteristic data to generate user power consumption characteristic data.

[0006] Furthermore, the electric multi-energy equipment data in step S11 includes photovoltaic electric energy equipment data, diesel electric energy equipment data and energy storage electric energy equipment data.

[0007] Furthermore, step S2 includes the following steps: Step S21: performing power multi-energy topology structure analysis based on power multi-energy equipment data to generate power multi-energy topology structure data; Step S22: performing power multi-energy connection relationship analysis on the power multi-energy device data to generate power multi-energy connection relationship data; Step S23: performing network topology modeling of the power multi-energy connection based on the power multi-energy connection relationship data and the power multi-energy topology structure data to generate a power multi-energy connection network topology model; Step S24: performing energy-specific mechanism operation analysis based on the electric multi-energy connection network topology model to generate energy-specific mechanism operation data; Step S25: performing mechanism operation constraint identification on the energy-specific mechanism operation data to generate constrained energy-specific mechanism operation data; Step S26: Analyzing the power multi-energy optimization operation objective function based on the power multi-energy connection network topology model to generate the power multi-energy optimization operation objective function; Step S27: Transmitting the constraint energy specific mechanism operation data and the power multi-energy optimization operation objective function to the power multi-energy connection network topology model for mapping the operation constraint conditions and the optimization operation objective function to generate a demand target operation power energy integration model; Step S28: Performing a boundary condition analysis of multiple power usage scenarios based on the user's power usage characteristic data to generate boundary condition data of multiple power usage scenarios; Step S29: Transmitting the boundary condition data of the electricity consumption multi-scenario to the demand target operation power energy integration model to perform intelligent optimization processing of the operation parameters of the solar-diesel-storage power energy integration, and generating an optimized operation solar-diesel-storage power energy integration model.

[0008] Furthermore, step S29 includes the following steps: Transmitting the boundary condition data of multiple power consumption scenarios to the demand target operation power energy integration model to perform power energy integration simulation operation characteristic analysis of each scenario and generate power energy integration simulation operation characteristic data; According to the power energy integration simulation operation characteristic data, the power energy integration optimization configuration parameters of each scenario are analyzed to generate the power energy integration optimization configuration parameters. The power energy integration optimization configuration parameters are used to intelligently optimize the operation parameters of the solar-diesel-storage power energy integration of the demand target operation power energy integration model to generate the optimized operation solar-diesel-storage power energy integration model.

[0009] Furthermore, step S3 includes the following steps: Step S31: Acquire multimodal operation perception element data of the deep-sea aquaculture power energy system, wherein the multimodal operation perception element data includes environmental perception data, power multi-energy operation load data, and power multi-energy equipment status data; Step S32: preprocessing the standard multimodal operation perception element data to obtain the standard multimodal operation perception element data; Step S33: performing load correlation data analysis of electric multi-energy operation based on the standard multi-modal operation perception element data to generate electric multi-energy operation load correlation data, and performing electric multi-energy operation load influencing factor characteristic analysis on the electric multi-energy operation load correlation data to generate electric multi-energy operation load influencing factor characteristic data; Step S34: performing an operation load cognitive rule feature analysis based on the preset power multi-energy load operation cognitive architecture and the power multi-energy load operation influencing factor feature data to generate operation load cognitive rule feature data; Step S35: performing Gaussian mixture analysis on the operating load cognitive rule feature data to generate the optimal number of operating load components, and establishing a Gaussian mixture model architecture of the operating load characteristics by configuring a preset Gaussian mixture algorithm based on the optimal number of operating load components to obtain a Gaussian mixture operating load characteristic model architecture; Step S36: Based on the characteristic data of the power multi-energy operation load influencing factors and the characteristic data of the operation load cognitive rules, a trend prediction model of the operation load transformation is established for the Gaussian mixture operation load characteristic model architecture to generate an operation load transformation trend prediction model.

[0010] Furthermore, step S36 includes the following steps: Step S361: extracting the operating load transformation time series hierarchical features from the operating load recognition rule feature data to obtain operating load transformation time series hierarchical feature data; Step S362: performing fuzzy feature analysis and training optimization processing on the operating load transformation trend based on the characteristic data of the power multi-energy operating load influencing factors and the characteristic data of the operating load transformation time series level to generate optimized operating load transformation trend fuzzy feature data; Step S363: Transmitting the optimized operating load transformation trend fuzzy feature data to the Gaussian mixture operating load feature model architecture to establish an operating load transformation trend prediction model, thereby generating an operating load transformation trend prediction model.

[0011] Furthermore, step S361 includes the following steps: Based on the characteristic data of the load influencing factors of the power multi-energy operation, the local time series hierarchical specificity analysis of the load influencing factors is performed to generate the local time series hierarchical specificity data of the load influencing factors. The local time series hierarchical specificity data of the load influencing factors is then used to design the network rules for the operation load time series hierarchical memory extraction. The operating load time series level memory extraction network rule is used to extract the operating load transformation time series level feature of the operating load cognitive rule feature data to obtain the operating load transformation time series level feature data.

[0012] Furthermore, step S362 includes the following steps: The fuzzy characteristic logic conversion of the operating load transformation trend is performed using the characteristic data of the power multi-energy operating load influencing factors as input data and the corresponding operating load transformation time series level characteristic data as output data to obtain the operating load transformation trend fuzzy characteristic data; The fuzzy trend feature data of the operating load transformation is subjected to fuzzy rule membership mapping through a preset operating load transformation fuzzy logic rule library to generate operating load transformation trend fuzzy feature data of the membership mapping, and the fuzzy feature data of the operating load transformation trend of the membership mapping is trained and optimized through a back propagation algorithm to generate optimized operating load transformation trend fuzzy feature data.

[0013] Furthermore, the step S4 of collecting the real-time load data of the multi-energy source by optimizing the operation of the solar-diesel-storage power energy integration model includes: Based on the optimized operation of the solar-diesel-storage power energy integration model, the power multi-energy operation optimization operation is performed on the deep-sea aquaculture power energy system; the power multi-energy real-time operation load data is collected according to the power multi-energy operation optimization operation.

[0014] Beneficial effects of the present application: The present invention systematically obtains power multi-energy equipment data and user electricity consumption data, combines multi-dimensional analysis methods to accurately characterize user electricity consumption characteristics, conducts targeted electricity consumption period analysis, and slices user electricity consumption data into time series according to different periods such as feeding, daily operations, and fishing, accurately capturing the differentiated characteristics of loads in each period, and solving the problem of vague load description caused by ignoring the particularity of the working conditions of the aquaculture platform in the analysis.

[0015] By extracting load time series features and clustering user electricity usage behaviors, we uncovered the associated characteristics of electricity usage patterns across different user groups, transforming raw data into a cognition of patterns and providing a refined basis for user needs in subsequent energy system modeling. By explicitly incorporating data from multiple energy devices, including photovoltaics, diesel, and energy storage, and covering the typical energy mix of deep-sea aquaculture platforms, we ensured a comprehensive analysis. The resulting user electricity usage characteristic data truly reflects the high variability of loads, laying a data foundation for optimal matching of energy systems and effectively improving the precision of matching energy supply with user needs.

[0016] Focusing on the modeling and optimization of integrated photovoltaic-diesel-storage energy systems, this project, through multi-level topological analysis, mechanism analysis, and parameter optimization, has constructed an energy integration model tailored to the realities of deep-sea aquaculture platforms, boasting outstanding technical features and advantages. During the model construction phase, from analyzing the multi-energy topology of the power grid to modeling the network topology, the project fully replicated the connectivity and coupling characteristics of the photovoltaic, diesel, and energy storage subsystems, overcoming the limitations of traditional linearized models that ignore strong nonlinear characteristics and more realistically reflecting the system's dynamic behavior.

[0017] Through energy-specific mechanism operation analysis and constraint identification, the operating boundaries and interaction rules of each energy device were clarified. Combined with the optimization objective function, the generated demand-target operation model possesses multi-constraint and multi-objective optimization capabilities. Furthermore, multi-scenario boundary condition analysis was conducted in conjunction with user electricity consumption data. Through simulation operation and parameter optimization, the final optimized operation model was adapted to the load demands under different electricity usage scenarios, achieving the optimization of the energy system from static design to dynamic adaptation. This provides core technical support for solving problems such as untimely energy regulation and low supply-demand matching accuracy in deep-sea aquaculture platforms.

[0018] By integrating multi-dimensional data with intelligent algorithms, a high-precision trend prediction model with distinctive technical features is constructed. At the data level, environmental perception data (such as light intensity and wind speed), multi-energy operational load data, and equipment status data are explicitly incorporated to form a multimodal operational perception element system, comprehensively covering the key factors influencing load fluctuations. In the data processing phase, pre-processing is used to ensure data quality, and in-depth analysis of load-related data is conducted to extract the characteristics of influencing factors, providing a solid foundation for model construction. The model architecture integrates a pre-defined multi-energy load operational cognitive architecture with a Gaussian mixture model, combined with time-series hierarchical memory extraction network rules and fuzzy feature logic transformation, to achieve multi-level analysis and modeling of load change trends. The time-series hierarchical feature extraction designs rules specific to the local time series of load influencing factors, while the fuzzy feature analysis enhances the model's adaptability to nonlinear and highly variable loads through membership mapping and back-propagation optimization. The resulting operational load change trend prediction model accurately captures the dynamic load variations of aquaculture platforms under different operating conditions.

[0019] The introduction of multimodal operational perception factor data fully considers the impact of multiple factors such as the environment and equipment status on the load, avoids deviations caused by ignoring key variables in the prediction, and improves the comprehensiveness of the prediction. Data preprocessing and influencing factor feature analysis ensure the reliability and relevance of the input data, provide a high-quality training foundation for the model, and reduce noise interference. The modeling method that integrates time-series hierarchical memory extraction, fuzzy logic, and Gaussian mixture models can deeply explore the dynamic evolution mechanism of the load, especially the nonlinear change characteristics of the load of the aquaculture platform during different periods such as feeding and fishing, achieving more accurate trend prediction and solving the problem of being unable to reflect the high variability of the load. The optimized prediction model can provide accurate load prediction for the energy management system and provide data support for subsequent energy coordinated control, which will help improve the timeliness and accuracy of energy regulation and control, and ensure the stable and efficient operation of the energy system of the aquaculture platform.

[0020] Based on the optimized operation of the photovoltaic-diesel-storage electric energy integration model, multi-energy operation optimization operations are performed, which can give full play to the characteristics of different energy sources such as photovoltaics, diesel, and energy storage, and realize the reasonable distribution of energy by combining their coupling relationship. By collecting real-time operation load data and combining the operation load change trend prediction model for trend analysis, the dynamic changes of load can be grasped in real time, providing an accurate basis for subsequent regulation, and solving the problem of untimely energy regulation caused by load forecast lag. Load collaborative control is performed based on trend feature data, which can realize dynamic matching between the energy supply side and the load side. The load side can adjust the power consumption plan according to its own characteristics and scheduling information to achieve flexible control effect, effectively cope with the fluctuations caused by the strong nonlinearity of the system and the high penetration rate of new energy, ensure the safe and stable operation of the power energy system of the breeding platform, and improve the overall economic and environmental benefits.

[0021] Therefore, in the deep-sea aquaculture power energy load intelligent analysis method based on deep learning of the present invention, the coupling relationship and strong nonlinear characteristics of each energy subsystem are fully analyzed, breaking through the limitation of the traditional linear model that ignores the strong nonlinear characteristics, making the model more in line with the actual operation scenario, and by analyzing the user's power consumption mode according to the unique working conditions of the aquaculture platform such as feeding and fishing, the power consumption period is analyzed, and the load differences under different working conditions are accurately captured, solving the problem that the high variability of the load cannot be reflected due to the failure to consider the working condition differences. In addition, the dynamic evolution mechanism of the load is deeply excavated in combination with deep learning technology, which improves the load prediction accuracy and effectively solves the problem of insufficient prediction accuracy caused by the prior art due to the failure to fully consider the system characteristics and working condition differences. By optimizing the combination of the operation model and the trend prediction model to achieve collaborative control, it can respond to load changes in real time, cope with fluctuations caused by the strong nonlinearity of the system, and ensure the stable and efficient operation of the energy system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a schematic diagram of the steps of a method for intelligent analysis of power energy load for deep-sea aquaculture based on deep learning according to the present invention; Figure 2 for Figure 1 Detailed implementation steps of step S1 in FIG. Figure 3 for Figure 1 Detailed implementation steps of step S2 in FIG. Figure 4 for Figure 1 Detailed implementation steps of step S3 in FIG. The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0023] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0024] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0025] To achieve this, please refer to Figures 1 to 4 The present invention provides a method for intelligent analysis of power energy load in deep-sea aquaculture based on deep learning. In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a flowchart of a method for intelligently analyzing power energy loads for deep-sea aquaculture based on deep learning according to the present invention. The method comprises the following steps: Based on this, the present invention provides an intelligent analysis method for deep-sea aquaculture power energy load based on deep learning to solve at least one of the above technical problems.

[0026] To achieve the above objectives, a deep-sea aquaculture power energy load intelligent analysis method based on deep learning includes the following steps: Step S1: Acquire the power multi-energy equipment data and user power consumption data of the deep-sea aquaculture power energy system; perform power consumption characteristic analysis on the user power consumption data to generate user power consumption characteristic data; In this embodiment of the present invention, a sensor network is deployed on a deep-sea aquaculture platform to collect real-time data on various power and energy devices, including the voltage, current, and conversion efficiency of photovoltaic panels; the output power, fuel consumption, and operating temperature of diesel generators; and the charge and discharge current, remaining capacity, and cycle count of energy storage batteries. Furthermore, user electricity consumption data, including real-time power, operating hours, and start / stop times, is collected for user electrical equipment within the platform, such as bait dispensers, aerators, monitoring equipment, and lighting systems. This user electricity consumption data is segmented into periods of feeding, daily operations, and fishing, generating period data. Based on this period data, a sliding time window method is used to slice the user electricity data into 10-minute intervals to generate period-specific time-series slices of user electricity load. Fourier transforms are used to extract periodic load characteristics, and wavelet transforms are used to extract mutation characteristics, generating period-specific time-series feature data for user electricity load. The K-means clustering algorithm is used to cluster the time-series feature data, grouping devices with similar power usage patterns together to generate clustered data on user electricity usage behavior. An association rule mining algorithm analyzes the power consumption relationships between different clustered data, such as the changing patterns in the power of the aerator pump during the operation of a baitcaster, to generate power consumption pattern correlation feature data. Finally, combined with this correlation feature data, a comprehensive analysis of the time series feature data by period is conducted to determine the power consumption patterns of each period and type of equipment, generating user power consumption characteristic data.

[0027] Step S2: establishing a demand-target operation mapping relationship specific to the power multi-energy device based on the power multi-energy device data, and generating a demand-target operation power energy integration model; establishing a solar-diesel-storage power energy integration model for optimizing the user's power consumption characteristics based on the user's power consumption characteristic data and the demand-target operation power energy integration model, and generating an optimized operation solar-diesel-storage power energy integration model; In this embodiment of the present invention, a topology mapping method is used to analyze the connection methods and positional relationships of devices such as photovoltaic arrays, diesel generators, energy storage batteries, inverters, and distribution boxes based on collected data on multi-energy devices, generating multi-energy topological structure data. Circuit principle analysis is used to determine the power transmission direction, wire type, and rated current capacity between each device, generating multi-energy connection relationship data. Based on these two types of data, a graph theory modeling method is used to construct a multi-energy connection network topology model consisting of nodes (devices) and edges (connection relationships). For photovoltaic devices in this model, the relationship between light intensity and output power is analyzed based on the principle of photoelectric conversion. For diesel generators, the relationship between speed and output power is analyzed based on the operating principle of internal combustion engines. For energy storage batteries, the relationship between charge and discharge rate and capacity change is analyzed based on electrochemical principles, generating energy-specific mechanism operation data. Constraints such as the power upper limit, efficiency range, and operating temperature range of each device are annotated based on device nameplate parameters to generate constrained energy-specific mechanism operation data. A multi-objective optimization function is established with the goals of minimizing fuel consumption, maximizing photovoltaic utilization, and extending the life of energy storage batteries to generate the multi-energy optimization operation objective function. Constraint data and objective functions are input into the network topology model. A mapping algorithm is used to establish a relationship between the constraints and the objective function, generating a demand-targeted power energy integration model. Based on user electricity consumption data, boundary conditions such as maximum load, duration, and voltage stability requirements for scenarios like feeding, daily operations, and fishing are analyzed to generate boundary condition data for multiple power scenarios. This boundary condition data is input into the demand-targeted power model. Through simulation of energy flows in each scenario, the optimal output ratio of photovoltaics, diesel, and energy storage is analyzed. Optimization parameters such as inverter conversion efficiency and energy storage charge and discharge thresholds are determined. Model parameters are adjusted to generate an optimized photovoltaic-diesel-storage power energy integration model.

[0028] Step S3: Acquire multimodal operation perception element data of the deep-sea aquaculture power energy system; perform a characteristic analysis of the factors influencing the power multi-energy operation load based on the multimodal operation perception element data to generate characteristic data of the factors influencing the power multi-energy operation load; establish a trend prediction model for the operation load transformation based on the characteristic data of the factors influencing the power multi-energy operation load to generate a trend prediction model for the operation load transformation; In this embodiment of the present invention, meteorological sensors are installed on the aquaculture platform to collect environmental perception data such as light intensity, wind speed, temperature, and humidity. A power monitoring device collects real-time active power, reactive power, voltage, and current data on multi-energy energy loads. An equipment status monitoring module collects multi-energy energy equipment status data, such as photovoltaic panel cleanliness, diesel generator fault codes, and energy storage battery internal resistance, to form multimodal operational perception factor data. The collected data is then subjected to the Laida criterion to remove outliers, linear interpolation to fill missing values, and maximum-minimum normalization to uniformly map the data to the [0,1] interval, generating standardized multimodal operational perception factor data. The Pearson correlation coefficient method is used to analyze the correlations between environmental factors, equipment status, and operational loads in the standardized data, such as the correlation coefficient between light intensity and photovoltaic output, and the correlation coefficient between wind speed and wind turbine output (if any), to generate multi-energy energy load correlation data. Principal component analysis is then used to extract key factors influencing loads, such as light intensity, equipment operating temperature, and feeder power, to generate characteristic data on factors influencing multi-energy energy loads. Based on a pre-defined multi-energy load operation cognitive architecture consisting of a perception layer, a fusion layer, and a decision-making layer, the influencing factor feature data is input into the fusion layer for feature fusion. Rule-based reasoning at the decision-making layer determines rules such as "diesel generator load increases during high temperatures" and "photovoltaic output increases during strong sunlight," generating operational load cognitive rule feature data. The Bayesian Information Criterion is used to determine the optimal number of components in the Gaussian mixture model for the rule feature data. The mean and covariance parameters of the Gaussian mixture algorithm are configured based on this number of components to construct the Gaussian mixture operational load feature model architecture. Based on the influencing factor feature data, the variation patterns of different factors on hourly, daily, and weekly timescales are analyzed. A hierarchical structure of a temporal hierarchical memory extraction network is designed, with the input layer receiving raw data, the hidden layer extracting hourly features, and the output layer extracting daily features. This network is used to process the operational load cognitive rule feature data to generate operational load transformation temporal hierarchical feature data. Using influencing factor feature data as input and time-series hierarchical feature data as output, a Takagi-Sugeno-Kang fuzzy generator performs fuzzy logic conversion, establishing fuzzy rules such as "If light intensity is high and temperature is suitable, the load will increase," to generate fuzzy feature data for operating load change trends. Using a backpropagation algorithm, the membership function parameters of the fuzzy rules are adjusted to minimize the output error. This generates optimized fuzzy feature data for operating load change trends, which are then input into a Gaussian mixture model architecture to complete the construction of a trend prediction model for operating load change.

[0029] Step S4: Collect the real-time operating load data of multiple power sources by optimizing the operation of the solar-diesel-storage power energy integration model; transmit the real-time operating load data of multiple power sources to the operating load transformation trend prediction model to perform power energy operating load trend characteristic analysis and generate power energy operating load trend characteristic data; and execute the power energy load coordinated control operation of deep-sea aquaculture based on the power energy operating load trend characteristic data.

[0030] In this embodiment of the present invention, based on an optimized solar-diesel-storage integrated power energy model, an energy management controller controls the MPPT (maximum power point tracking) strategy of the photovoltaic inverter, the start / stop thresholds of the diesel generator, and the charge / discharge strategy of the energy storage battery in real time to optimize the operation of multiple energy sources. During this optimization process, Hall sensors installed at the output terminals of each energy device collect real-time photovoltaic output, diesel generator power, and energy storage battery charge / discharge power, and aggregate these data to generate real-time load data for the multiple energy sources. This real-time load data is input into a load change trend prediction model. The model analyzes the data's compatibility with historical load patterns and combines it with current environmental data (such as the rate of change of light intensity and wind speed fluctuations) to output load trend characteristics for the next 1, 3, and 24 hours, including load increase / decrease trends, fluctuation amplitudes, and peak occurrence times. Based on trend data, if the load is predicted to increase within the next hour and the PV output is sufficient, the energy management controller issues a command to place the energy storage battery in a floating charge state, while the diesel generator remains in standby mode. If the load is predicted to decrease and the PV output is excessive, the energy storage battery is switched to a charging state, and the inverter output voltage is adjusted to efficiently store excess energy. If the load is predicted to rise sharply and the PV output is insufficient, the diesel generator is immediately started, simultaneously releasing the energy storage battery to ensure that the total output matches the load demand. Load-side equipment such as bait dispensers and aerators receive dispatch signals from the control center and adjust their operating hours based on their own operating cycles. For example, monitoring equipment reduces its sampling frequency during non-emergency periods, achieving coordinated control of power energy loads.

[0031] Further, as an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in FIG. 1 , in this embodiment, step S1 includes the following steps: Step S11: Acquire the power multi-energy equipment data and user power consumption data of the deep-sea aquaculture power energy system; In an embodiment of the present invention, sensors are installed on the photovoltaic arrays, diesel generators, energy storage battery packs and other power and multi-energy equipment of the deep-sea aquaculture platform. Among them, voltage sensors and current sensors are deployed at the photovoltaic panels to collect the output voltage, current and power generation efficiency in real time; power sensors and fuel flow sensors are installed on the diesel generators to record the output active power, reactive power and hourly fuel consumption; the energy storage battery packs are connected to the power monitoring module to collect the charge and discharge current, remaining capacity and number of cycles. The above data together constitute the power and multi-energy equipment data. At the same time, metering devices are installed on the power supply lines of user electrical equipment such as bait dispensers, aeration pumps, water quality monitors, and lighting equipment to collect the start time, operating time, real-time power and stop time of each device to form user electricity consumption data. All data are collected to the data acquisition terminal through wired transmission. The transmission frequency is set to once per minute to ensure the continuity and timeliness of the data and provide complete original data support for subsequent analysis.

[0032] Step S12: Analyze the power consumption period according to the user power consumption data to generate power consumption period data, and perform time series slicing of the power load according to the time difference of the user power consumption data based on the power consumption period data to generate time series slicing data of the user power load according to the time difference of the period; In an embodiment of the present invention, electricity usage periods are divided into three categories: feeding period, daily operation period, and fishing period, based on the operating time stamps of each device in the user's electricity usage data. The feeding period is determined based on the fixed operating hours of the bait feeding machine; the daily operation period is the period excluding the feeding and fishing periods; and the fishing period is set as a specific time period on the last three days of each month according to the aquaculture plan, thereby generating electricity usage period data. Based on the electricity usage period data, a fixed time interval segmentation method is used to process the user's electricity usage data. The electricity usage data for the feeding period, daily operation period, and fishing period are sliced ​​using a 5-minute time slice unit. Each slice contains the power data of all electrical devices within that time period, generating time series slice data for the user's electricity load by period. This slicing process based on period differences can clearly distinguish load changes in different aquaculture operation stages, avoiding feature ambiguity caused by the mixing of data from different periods.

[0033] Step S13: extracting load time series characteristics based on the time series slice data of the user's electricity load in different periods to obtain the time series characteristic data of the user's electricity load in different periods; In an embodiment of the present invention, a time series analysis method is used to extract the load time series characteristics for the time series slice data of the user's electricity load in different periods. For the slice data of the feeding period, the maximum power, minimum power, average power and power change rate in each slice are calculated, wherein the power change rate is obtained by the ratio of the power difference between two adjacent slices to the time interval, reflecting the load fluctuation when the bait feeder starts and stops; for the slice data of the daily operation period, the periodic characteristics of the power data are analyzed, and the operating cycle of the lighting equipment, circulating water pump and other equipment is determined by calculating the similarity of the power curves in adjacent periods; for the slice data of the fishing period, the number of power peaks and the duration are counted to capture the load characteristics of high-power equipment such as the net-lifting machine when they are in operation. The maximum value, minimum value, average power, change rate, periodic parameters, number of peaks and duration extracted above are summarized to form the time series characteristic data of the user's electricity load in different periods, which fully presents the dynamic change law of the load in different periods.

[0034] Step S14: performing a cluster analysis of user electricity consumption behaviors based on the time series characteristic data of user electricity loads in different periods to generate user electricity consumption behavior cluster data; In an embodiment of the present invention, a cluster analysis method is used to classify user electricity consumption behaviors based on the time series characteristic data of user electricity load in different periods. The power average value, change rate, and peak duration of each period are selected as clustering characteristic indicators, and by calculating the Euclidean distance of different electrical equipment on these indicators, devices with closer distances are grouped into one category. For example, bait casters and net lifters both exhibit the characteristics of high power peaks and fast change rates when working, and are clustered into one category; lighting equipment and small monitoring equipment have stable power and slow changes, and are clustered into another category; aeration pumps and circulating water pumps have medium power and long continuous operation time, and are clustered into the third category. Through this clustering process, user electricity consumption behavior clustering data is generated, and the behavioral patterns of different types of electrical equipment can be distinguished, laying the foundation for the subsequent analysis of the electricity consumption correlation of various types of equipment.

[0035] Step S15: performing power usage pattern correlation feature analysis based on the user power usage behavior clustering data to generate power usage pattern correlation feature data; In an embodiment of the present invention, based on the clustering data of user electricity consumption behavior, an association rule analysis method is used to mine the associated characteristics of electricity consumption patterns between different clusters. The clustering data of the feeding period is analyzed, and the power changes of the aerator when the bait feeder is started are counted. The probability and amplitude of the increase in the power of the aerator after the bait feeder is started are calculated to determine the collaborative operation relationship between the two in the feeding operation; the clustering data of the daily operation period is analyzed to observe the correlation between the lighting equipment and the power of the monitoring equipment, such as the law that the power of the monitoring equipment is stable due to the improvement of the image acquisition clarity after the lighting equipment is turned on; the clustering data of the fishing period is analyzed to record the changes in the discharge power of the energy storage battery when the net machine is running, and to clarify the impact of the operation of high-power equipment on the energy storage system. These associations are presented in the form of characteristic parameters, such as the probability of collaborative operation, the ratio of power change amplitude, etc., to generate electricity consumption pattern association feature data, revealing the interactive rules of electricity consumption of different types of equipment in different periods.

[0036] Step S16: performing a user power consumption characteristic analysis on the time series characteristic data of the user power load in different periods by using the power consumption pattern association characteristic data to generate user power consumption characteristic data.

[0037] In an embodiment of the present invention, combined with the power consumption pattern correlation characteristic data, a comprehensive analysis is performed on the user power load time series characteristic data in different periods to generate user power consumption characteristic data. During the feeding period, based on the correlation characteristics of the bait feeding machine and the aeration pump, the power change rate in the time series characteristics of the different periods is corrected, and the synergistic operation coefficient of the two is added to more accurately reflect the load fluctuation characteristics of the feeding operation; during the daily operation period, based on the correlation law of the lighting equipment and the monitoring equipment, the periodic parameters in the time series characteristics are adjusted and the linkage cycle between the equipment is incorporated; during the fishing period, based on the correlation characteristics of the net lifting machine and the energy storage system, the corresponding relationship between the peak duration and the energy storage discharge rate in the time series characteristics is supplemented. Through this integrated analysis, the load time series characteristics of different periods are combined with the correlation characteristics between the equipment to form user power consumption characteristic data containing parameters such as the load mean value, fluctuation range, equipment synergy coefficient, and energy storage dependence of each period, which comprehensively depicts the power consumption patterns and characteristics of the deep-sea aquaculture platform in different operation stages.

[0038] Furthermore, the electric multi-energy equipment data in step S11 includes photovoltaic electric energy equipment data, diesel electric energy equipment data and energy storage electric energy equipment data.

[0039] Further, as an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment, step S2 includes the following steps: Step S21: performing power multi-energy topology structure analysis based on power multi-energy equipment data to generate power multi-energy topology structure data; In this embodiment of the present invention, a topology analysis method is used to map the system's physical connection framework based on the installation locations, physical parameters, and functional identifiers of photovoltaic panels, diesel generators, energy storage batteries, inverters, and other devices in the power multi-energy device data. For the photovoltaic array, the arrangement order of each photovoltaic panel group is marked according to the series and parallel combination. The cable routing and fixing method between the photovoltaic panels and the combiner box are measured and recorded, and the installation spacing and connection port types between the combiner box and the photovoltaic inverter are clarified. For the diesel generator, its relative position to the main distribution cabinet is marked, and the cable routing and support structure from the generator output to the distribution cabinet are recorded. For the energy storage battery pack, the arrangement layout within the battery compartment is marked according to module grouping, and the cable length and interface specifications between the battery pack and the bidirectional converter are measured. The spatial coordinates of each device are determined using a laser rangefinder. Combined with the model parameters on the device nameplate, the power multi-energy topology data is generated, including the device number, coordinate location, connection port type, and cable specifications. This fully presents the physical distribution and connection framework of the three types of energy devices on the aquaculture platform, laying the spatial foundation for subsequent electrical relationship analysis.

[0040] Step S22: performing power multi-energy connection relationship analysis on the power multi-energy device data to generate power multi-energy connection relationship data; In this embodiment of the present invention, the electrical parameters and operating records in the data of multi-energy devices are analyzed through circuit principles to determine the connection relationships between each device. For photovoltaic systems, the DC voltage level between the output of the photovoltaic panel and the input of the combiner box is measured. The circuit continuity is tested with a multimeter to determine the unidirectional transmission relationship of current from the photovoltaic panel to the combiner box, and the maximum transmission current between the combiner box and the inverter is recorded. For diesel generators, a phase detector is used to determine the three-phase connection phase between its output and the main circuit breaker. The line impedance between the generator and the bus is measured to determine the capacity limit of power transmission from the generator to the bus. For energy storage systems, a charge and discharge tester is used to detect the current flow between the energy storage battery and the bidirectional converter. The bidirectional transmission characteristics of current from the converter to the battery during charging and from the battery to the converter during discharging are determined, and the voltage conversion range between the converter and the DC bus is recorded. These electrical relationships are quantified into multi-energy connection relationship data that includes transmission direction, voltage level, current limit, and phase matching. This clearly reflects the interaction rules of each device at the electrical level and ensures the safety and compatibility of energy transmission.

[0041] Step S23: performing network topology modeling of the power multi-energy connection based on the power multi-energy connection relationship data and the power multi-energy topology structure data to generate a power multi-energy connection network topology model; In this embodiment of the present invention, a graph-theoretic modeling approach is used to construct a network topology model, combining data on the connection relationships of multiple power energy sources with topological structure data. Devices are represented as nodes, and connections between devices as edges. Each node is assigned attributes such as device type (PV / diesel / energy storage), rated power, and operating status. Each edge is assigned attributes such as transmission direction, electrical parameters, and loss coefficient. For PV nodes, the edge attributes between them and the combiner box node are defined as "DC transmission, voltage XXV, maximum current XXA." For diesel generator nodes, the edge attributes between them and the main distribution cabinet node are defined as "AC three-phase, voltage XXXV, frequency XXHz." For energy storage nodes, the edge attributes between them and the bidirectional converter node are defined as "bidirectional DC, conversion efficiency XX%." By mapping nodes and edges, an overall network structure is constructed, encompassing the PV subnetwork (PV panels - combiner box - inverter), the diesel generator subnetwork (generator - main circuit breaker - busbar), and the energy storage subnetwork (battery - bidirectional converter - busbar), forming a topological model for the multi-energy connection network. The model visualizes the electrical connection paths and parameter limitations of each subsystem, which can intuitively reflect the energy transmission paths and interaction rules within the system, providing a structured model basis for subsequent mechanism analysis.

[0042] Step S24: performing energy-specific mechanism operation analysis based on the electric multi-energy connection network topology model to generate energy-specific mechanism operation data; In an embodiment of the present invention, based on a multi-energy network topology model for electric power, energy-specific mechanism operation analysis is conducted for three types of energy devices: photovoltaic, diesel, and energy storage. For photovoltaic systems, based on the principle of photoelectric conversion, a light intensity sensor continuously collects light data at different time periods, synchronously records the output power of the photovoltaic panels, plots the relationship between light intensity and output power, analyzes the influence of temperature on conversion efficiency, and determines the volt-ampere characteristics and power output curve of the photovoltaic array under different light and temperature conditions. For diesel generators, based on the operating principle of internal combustion engines, the output power is adjusted by changing the load resistance. The fuel consumption, speed, and exhaust temperature at different power levels are recorded, and a relationship curve between power and fuel consumption rate is plotted. The impact of speed fluctuations on output voltage stability is analyzed to determine the power response characteristics and efficiency variation of the generator. For energy storage systems, based on electrochemical principles, a charge and discharge tester is used to set different charge and discharge currents, record the changes in battery voltage, capacity, and internal resistance, plot the relationship between charge and discharge depth and capacity retention rate, analyze the influence of cycle number on battery performance degradation, and determine the charge and discharge characteristics and life cycle variation of the energy storage battery. These analysis results are quantified into characteristic parameters, such as the equation for how photovoltaic conversion efficiency changes with light and temperature, the characteristic equation for diesel generator fuel consumption, the capacity attenuation coefficient of energy storage batteries, etc., to generate energy-specific mechanism operation data and fully present the internal operating mechanisms of various energy equipment.

[0043] Step S25: performing mechanism operation constraint identification on the energy-specific mechanism operation data to generate constrained energy-specific mechanism operation data; In this embodiment of the present invention, energy-specific mechanism operation data is combined with the physical characteristics of the equipment and safe operation specifications to identify mechanism operation constraints. For example, for photovoltaic systems, the maximum output power limit is marked according to the module nameplate parameters, the operating temperature range (-25°C to 65°C) is marked according to environmental adaptability standards, and the voltage fluctuation threshold (±3% of the rated value) is marked according to circuit protection requirements to ensure that the photovoltaic array operates within a safe range. For diesel generators, the minimum stable operating power (no less than 25% of the rated power) is marked according to the unit operating manual, the continuous operation time (no more than 8 hours) is marked according to mechanical fatigue limits, and the start interval (no less than 10 minutes) is marked according to the starting system characteristics to prevent damage to the generator due to low load or frequent starts and stops. For energy storage systems, the maximum charge and discharge current is marked according to battery safety standards (no more than 1.5 times the rated current), the SOC (state of charge) range (15%-90%) is marked according to capacity retention requirements, and the maximum number of cycles (no more than 3,000 times) is marked according to the cycle life design to prevent performance degradation of energy storage batteries due to overcharge, overdischarge, or excessive cycling. These constraints are associated with the characteristic parameters in the energy-specific mechanism operation data to generate constrained energy-specific mechanism operation data, clarify the boundary conditions that various energy equipment must comply with during operation, and provide a constraint basis for subsequent energy system optimization.

[0044] Step S26: Analyzing the power multi-energy optimization operation objective function based on the power multi-energy connection network topology model to generate the power multi-energy optimization operation objective function; In an embodiment of the present invention, an objective function for optimizing the operation of a multi-energy power system is constructed based on a multi-energy power network topology model and the operational requirements of deep-sea aquaculture platforms. With reducing energy costs as the core objective, the objective function incorporates diesel generator fuel consumption, photovoltaic utilization, and energy storage battery maintenance costs. Weight coefficients for each parameter are determined through cost accounting to form a cost optimization sub-function. With improving system stability as a key objective, the objective function incorporates voltage fluctuation amplitude, frequency deviation, and the number of power outages. The allowable ranges for each parameter are determined based on power system operating standards to form a stability optimization sub-function. With extending equipment life as a secondary objective, the number of energy storage battery cycles and the number of diesel generator starts and stops are incorporated into the objective function. The influence coefficients of each parameter are determined based on an equipment life assessment model to form a life optimization sub-function. The three sub-functions are integrated into a comprehensive optimization objective function through weighted summation. The function variables are set as photovoltaic output, diesel generator output power, and energy storage charge and discharge power, and the value ranges of each variable are clearly defined (based on the operating data of the constrained energy-specific mechanism). The generated power multi-energy optimization operation objective function can quantitatively reflect the comprehensive optimization direction of the energy system in terms of cost, stability, and equipment life, and provide a quantitative basis for the subsequent energy scheduling strategy formulation. The power multi-energy optimization operation objective function can adaptively adjust the power multi-energy optimization target according to the needs of managers.

[0045] Step S27: Transmitting the constraint energy specific mechanism operation data and the power multi-energy optimization operation objective function to the power multi-energy connection network topology model for mapping the operation constraint conditions and the optimization operation objective function to generate a demand target operation power energy integration model; In this embodiment of the present invention, the operational data of the constrained energy-specific mechanism and the objective function for optimizing the operation of a multi-energy power grid are input into a multi-energy power grid topology model. A mapping algorithm is then used to establish a dynamic association between the constraints and the objective function. A constraint module is incorporated into the model to convert constraints such as the maximum photovoltaic output power, the minimum stable diesel generator power, and the energy storage SOC range into mathematical inequalities, which are then embedded into the objective function's solution domain. An optimization module is also incorporated to solve the objective function using a gradient descent method. During the solution process, variables are verified in real time to ensure they meet the constraints. If the calculated photovoltaic output exceeds the maximum limit, the output is automatically reduced; if the diesel generator output falls below the minimum power, an adjustment mechanism is triggered. By iteratively matching the variables with the constraints, the optimal solution of the objective function is consistently within the constraints, forming a demand-targeted integrated power energy model that incorporates coordinated operational rules for photovoltaics, diesel generators, and energy storage. This model can output energy allocation plans that meet the lowest cost and highest stability under the constraints, such as a coordinated strategy that prioritizes photovoltaic output, supplements diesel generators for peak load regulation, and uses energy storage to smooth fluctuations. This fully demonstrates the operational logic of the energy system under the dual influence of constraints and optimization.

[0046] Step S28: Performing a boundary condition analysis of multiple power usage scenarios based on the user's power usage characteristic data to generate boundary condition data of multiple power usage scenarios; In this embodiment of the present invention, a scenario analysis method is used to determine boundary conditions for multiple power usage scenarios based on load characteristics at different times in user power consumption data. For the feeding scenario, the maximum load value and load duration in minutes are determined based on the peak load and duration of the bait dispenser's operation. The voltage fluctuation tolerance (not exceeding ±2% of the rated value) is set in conjunction with the requirements for coordinated operation of the aerator pump. For daily operations, the average load level and minimum power supply duration are determined based on the continuous operation characteristics of lighting, monitoring, and other equipment. The frequency deviation range (±0.5Hz) is set based on the equipment's voltage tolerance. For fishing scenarios, the peak load multiple and peak duration in seconds are determined based on the instantaneous start-up characteristics of high-power equipment such as net haulers. The backup power supply response time (not exceeding 5 seconds) is set in accordance with safe operation requirements. Furthermore, marine environmental data is combined to supplement environmental boundary conditions such as light intensity ranges and wind speed limits for different scenarios. For example, photovoltaic output is treated as zero in nighttime scenarios, and diesel generator backup starts are initiated in high winds. These parameters are summarized into multi-scenario boundary condition data for electricity consumption, including load parameters, time parameters, electrical parameters, and environmental parameters, to clearly define the operating thresholds that the energy system must meet in each scenario.

[0047] Step S29: Transmitting the boundary condition data of the electricity consumption multi-scenario to the demand target operation power energy integration model to perform intelligent optimization processing of the operation parameters of the solar-diesel-storage power energy integration, and generating an optimized operation solar-diesel-storage power energy integration model.

[0048] In this embodiment of the present invention, boundary condition data from multiple electricity scenarios is input into a demand-driven power energy integration model, and simulation optimization is used to adjust the solar-diesel-storage operational parameters. For the feeding scenario, the model simulates energy flow during peak load periods. By adjusting the PV inverter's MPPT tracking accuracy, the diesel generator's startup delay, and the energy storage discharge rate, the model ensures that the combined output of the three accurately matches the peak load while minimizing the number of diesel generator starts. For the daily operation scenario, the model simulates energy distribution under stable load conditions. By optimizing the energy storage charging threshold, excess PV power is stored, while the diesel generator remains in standby mode and only starts when PV output is insufficient, reducing fuel consumption. For the fishing scenario, transient peak impacts are simulated. By adjusting the energy storage battery discharge rate and the diesel generator's rapid response coefficient, the energy storage is prioritized to release power to smooth the peak, while the diesel generator simultaneously follows up to ensure voltage stability. Through parameter iteration across multiple simulation scenarios, the optimal combination of PV utilization, diesel generation efficiency, and energy storage charge and discharge depth is determined for each scenario, generating an optimized solar-diesel-storage power integration model. The model can automatically switch parameter configurations according to the scenario, such as increasing the priority of energy storage discharge during feeding and increasing the weight of photovoltaic utilization during daily use, to achieve efficient adaptation of the energy system in different scenarios.

[0049] Furthermore, step S29 includes the following steps: Transmitting the boundary condition data of multiple power consumption scenarios to the demand target operation power energy integration model to perform power energy integration simulation operation characteristic analysis of each scenario and generate power energy integration simulation operation characteristic data; According to the power energy integration simulation operation characteristic data, the power energy integration optimization configuration parameters of each scenario are analyzed to generate the power energy integration optimization configuration parameters. The power energy integration optimization configuration parameters are used to intelligently optimize the operation parameters of the solar-diesel-storage power energy integration of the demand target operation power energy integration model to generate the optimized operation solar-diesel-storage power energy integration model.

[0050] In this embodiment of the present invention, boundary condition data for multiple power scenarios is input into a demand target to run an integrated power energy model. A time-series simulation method is then used to simulate the integrated power energy model for each scenario. For the feeding scenario, the model simulates the photovoltaic output curve as it changes with sunlight, the diesel generator startup response process, and the charge and discharge state switching of the energy storage battery within the specified operating period, power peak, and duration of the bait feeder, based on the boundary conditions. The model also records the actual photovoltaic utilization rate, diesel generator fuel consumption rate, energy storage SOC change amplitude, and system bus voltage fluctuation every 10 minutes. For the daily operation scenario, the coordinated operation of photovoltaics, diesel, and energy storage is simulated over a 24-hour period based on the average load level and duration. The efficiency of the diesel generator during the nighttime dark period, the depth of energy storage discharge, and load matching are specifically recorded. For the fishing scenario, the energy impact response upon startup of high-power equipment is simulated based on the instantaneous peak load and duration in seconds. The maximum discharge current of the energy storage battery, the power ramp rate of the diesel generator, and the system frequency stabilization time are recorded. Through multiple simulation rounds, energy output curves, equipment operating parameters, and system stability indicators for each scenario were summarized to generate power energy integration simulation operational characteristic data, fully demonstrating the dynamic operational characteristics of the solar-diesel-storage system under different scenarios. Based on this power energy integration simulation operational characteristic data, a parameter optimization method was used to analyze the optimal configuration parameters for each scenario. For the feeding scenario, the PV utilization rate at different PV inverter MPPT tracking frequencies, the total fuel consumption at different diesel generator start-up thresholds, and the voltage stability at different energy storage discharge initial SOCs were compared to determine the inverter tracking frequency, generator start-up threshold, and energy storage initial SOC that maximize PV utilization, minimize fuel consumption, and minimize voltage fluctuation. For the daily operation scenario, the energy storage capacity utilization rate corresponding to different energy storage charging cut-off voltages and the energy loss corresponding to different diesel generator standby powers were analyzed to determine the charging cut-off voltage and standby power that balance energy storage life and diesel energy consumption. For the fishing scenario, the peak suppression effect at different energy storage discharge rates and the frequency deviation corresponding to different diesel generator response delays were evaluated to determine the discharge rate and response delay parameters that quickly smooth peaks and ensure frequency stability. These parameters are organized into power energy integration optimization configuration parameters including photovoltaic equipment adjustment parameters, diesel generator control parameters, and energy storage system operating parameters. The demand targets are substituted into the power energy integration model to replace the default parameters in the original model. The intelligent optimization processing of the operating parameters of the photovoltaic-diesel-storage power energy integration is completed, and an optimized operating photovoltaic-diesel-storage power energy integration model is generated to ensure that the model can achieve a balance between efficient energy utilization and stable system operation in all scenarios.

[0051] Further, as an embodiment of the present invention, refer to Figure 4 As shown, Figure 1Detailed step flow diagram of step S3 in the embodiment, step S3 includes the following steps: Step S31: Acquire multimodal operation perception element data of the deep-sea aquaculture power energy system, wherein the multimodal operation perception element data includes environmental perception data, power multi-energy operation load data, and power multi-energy equipment status data; In this embodiment of the present invention, a multi-dimensional perception system is established on a deep-sea aquaculture platform to collect multimodal operational perception data. Environmental perception data is acquired every 10 minutes by meteorological monitoring equipment installed on the upper deck of the platform. This data includes light intensity, offshore wind speed, seawater temperature, and air humidity. Light intensity is captured in real time using a photodiode sensor, wind speed is measured using a propeller-type wind speed sensor, and temperature and humidity are simultaneously recorded using a waterproof temperature and humidity sensor. Power multi-energy load data is collected every two minutes by monitoring devices deployed at the distribution hub. Power measurement is achieved using electromagnetic induction transformers to ensure data accuracy. Power multi-energy device status data is extracted by the device's built-in monitoring unit. Photovoltaic panel status includes panel temperature and conversion efficiency degradation; diesel generator status includes cylinder temperature, fuel pressure, and operating time; and energy storage battery status includes total voltage and number of charge and discharge cycles. The data is collected every minute. All data is transmitted to a local storage unit via shielded cables, forming multimodal operational perception data covering environmental parameters, load parameters, and device status parameters, comprehensively covering the key variables affecting energy system operation.

[0052] Step S32: preprocessing the standard multimodal operation perception element data to obtain the standard multimodal operation perception element data; In this embodiment of the present invention, a standardized preprocessing process is performed on multimodal operational perception factor data to generate standard data. First, outliers are removed. For wind speed and light intensity in environmental data, the Dixon test method is used to identify data points that significantly deviate from the general population. These data points are then replaced with the average of three valid data points before and after the corresponding period. For voltage fluctuations in power load data, a sliding window method (with a window size of five acquisition cycles) is used to identify sudden changes in data and replace them with the median value within the window. Next, missing values ​​are imputed. For missing data in a single period, linear interpolation of two adjacent valid data points is used to fill the gap. For missing data in three or fewer consecutive periods, the arithmetic mean of past data from the same day and the same period is used to fill the gap. Finally, data normalization is performed. Light intensity in environmental data is converted to its maximum measurement range, power in power load data is scaled to its rated power ratio, and parameters such as temperature and pressure in equipment status data are normalized to their safe operating ranges. All data is uniformly mapped to the [0, 1] interval, generating standardized multimodal operational perception factor data and eliminating dimensional differences between different data types.

[0053] Step S33: performing load correlation data analysis of electric multi-energy operation based on the standard multi-modal operation perception element data to generate electric multi-energy operation load correlation data, and performing electric multi-energy operation load influencing factor characteristic analysis on the electric multi-energy operation load correlation data to generate electric multi-energy operation load influencing factor characteristic data; In an embodiment of the present invention, based on standard multimodal operational perception factor data, a method combining statistical analysis and mechanism deduction is used to conduct load correlation analysis and extract influencing factor features. For the light intensity in the environmental perception data and the photovoltaic output in the multi-energy power operation load data, a scatter plot is drawn and the Pearson correlation coefficient is calculated to clarify the degree of linear correlation between the two. At the same time, combined with the photoelectric conversion efficiency curve, the rate of change of photovoltaic output under different illumination intervals is derived. For the wind speed data and the output of the wind turbine (if equipped), a segmented fitting method is used to analyze the output response pattern in the low, medium, and high wind speed intervals to determine the influence coefficient of wind speed on the wind turbine load. For diesel generator speed and output power in the power multi-energy device status data, a nonlinear mapping relationship between speed and power was established based on the internal combustion engine power formula. By continuously measuring power values ​​at different speeds, a correlation curve was generated between the two. For energy storage battery SOC and charge and discharge power, the electrochemical characteristics were combined to analyze the differences in charge and discharge efficiency within the SOC ranges of 20%-50%, 50%-80%, and 80%-90%. The constraints imposed by SOC on energy storage load were determined, thereby generating power multi-energy operating load correlation data. Based on this, the influencing factors of the correlation data were extracted. For light intensity, the daily variation amplitude, continuous peak duration, and hysteresis response time with photovoltaic output were calculated. For wind speed, the interference coefficient of instantaneous fluctuation frequency and direction changes on wind turbine output was extracted. For diesel generators, the correlation characteristics between speed fluctuation rate, cylinder temperature rise rate, and power fluctuation were analyzed. For energy storage batteries, the corresponding relationship between SOC change rate, number of charge and discharge cycles, and capacity decay was extracted. These characteristics are quantified into specific indicators, such as the photovoltaic output adjustment when the light intensity changes by 10,000 lux per hour, and the power deviation value corresponding to a 1% fluctuation in diesel generator speed. This generates characteristic data of factors affecting the multi-energy operation load of electricity, accurately reflecting the driving mechanism of each factor on load changes.

[0054] Step S34: performing an operation load cognitive rule feature analysis based on the preset power multi-energy load operation cognitive architecture and the power multi-energy load operation influencing factor feature data to generate operation load cognitive rule feature data; In this embodiment of the present invention, the pre-configured multi-energy load operation cognitive architecture comprises three layers: a perception layer, a fusion layer, and a decision layer. This layer analyzes the characteristics of operational load cognitive rules based on characteristic data of factors influencing multi-energy load operation. The perception layer receives characteristic data of factors such as light intensity, device temperature, and energy storage SOC. It then divides the continuous data into discrete states based on thresholds. For example, light intensity is divided into "weak (less than 30,000 lux), medium (30,000-70,000 lux), and strong (greater than 70,000 lux)"; and energy storage SOC is divided into "low (less than 30%), medium (30%-70%), and high (greater than 70%)." The fusion layer combines these discrete states to form association patterns, such as "strong light + high energy storage SOC" and "high temperature + heavy diesel generator load." The fusion layer then calculates the direction (increasing / decreasing / stable) and magnitude of load change when each pattern occurs. The decision layer generates cognitive rules based on the pattern statistics from the fusion layer. For example, when the "strong sunlight and feeding period" pattern occurs, the probability and average increase in load are recorded, forming the rule that "strong sunlight + feeding period" leads to "load increase by a set value." When the "wind speed > 10 m / s and nighttime" pattern occurs, the diesel generator startup frequency is counted, forming the rule that "high wind speed + nighttime" leads to "diesel generator load increase." These rules are quantified into characteristic data for operational load cognitive rules, including preconditions, load change conclusions, and confidence levels, to clarify the association between influencing factor combinations and load changes.

[0055] Step S35: performing Gaussian mixture analysis on the operating load cognitive rule feature data to generate the optimal number of operating load components, and establishing a Gaussian mixture model architecture of the operating load characteristics by configuring a preset Gaussian mixture algorithm based on the optimal number of operating load components to obtain a Gaussian mixture operating load characteristic model architecture; In this embodiment of the present invention, a statistical test method is used to determine the optimal number of components for a Gaussian mixture model based on the characteristic data of operational load recognition rules. The rule characteristic data is divided into three independent datasets based on load variation type: increasing, decreasing, and stable. Each dataset contains characteristic parameters such as rule confidence and influencing factor strength. For each dataset, the number of components is set from 1 to 8. The maximum expectation algorithm is used to estimate the Gaussian mixture model parameters (mean, covariance, and weight) for each component number. The Bayesian Information Criterion (BIC) value for each model is calculated. This value comprehensively considers model fit and complexity, with smaller values ​​indicating a better model. By comparing the BIC values ​​corresponding to different numbers of components, the number of components corresponding to the minimum value is selected as the optimal number of components for the operational load. For example, the optimal number of components for the increasing load dataset is 3, for the decreasing load dataset is 4, and for the stable load dataset is 2. The Gaussian mixture algorithm is configured based on the optimal number of components, and a probability distribution of characteristic parameters is assigned to each component. A Gaussian mixture operational load characteristic model architecture is constructed, consisting of an input layer (receiving the rule characteristic data), a mixing layer (corresponding to the optimal number of components), and an output layer (outputting the load variation probability). This model accurately captures the probability distribution characteristics of different load variation types.

[0056] Step S36: Based on the characteristic data of the power multi-energy operation load influencing factors and the characteristic data of the operation load cognitive rules, a trend prediction model of the operation load transformation is established for the Gaussian mixture operation load characteristic model architecture to generate an operation load transformation trend prediction model.

[0057] In this embodiment of the present invention, a Gaussian mixture operating load characteristic model architecture is trained based on the characteristic data of multi-energy operating load influencing factors and the characteristic data of operating load cognitive rules to generate a trend prediction model. Real-time parameters such as the rate of change of sunlight and equipment temperature fluctuations from the influencing factor characteristic data, along with the rule confidence from the cognitive rule characteristic data, serve as model inputs. The load change trend (increase / decrease / stability) for the next 30 minutes is output. The maximum expectation algorithm is used to iteratively optimize the model parameters. By calculating the posterior probability that each sample belongs to a different Gaussian component, the component means and covariances are adjusted to gradually reduce the model's prediction error for historical data. During the training process, the model accuracy is verified every 50 iterations. Training is terminated when the accuracy fluctuation is less than 1% for three consecutive verifications. The resulting operating load change trend prediction model can output the probability distribution of load increase, decrease, and stability in the next 30 minutes based on the real-time input influencing factors and rule characteristics. For example, it can output "58% probability of increase, 32% stability, 10% decrease." It can also distinguish load change patterns in different scenarios, such as feeding and fishing seasons, providing accurate trend information for energy coordinated control.

[0058] Furthermore, step S36 includes the following steps: Step S361: extracting the operating load transformation time series hierarchical features from the operating load recognition rule feature data to obtain operating load transformation time series hierarchical feature data; In an embodiment of the present invention, the time series hierarchical decomposition method is used to extract the time series hierarchical characteristics of the operation load transformation from the operation load cognitive rule feature data. First, the time scale characteristics of each factor in the characteristic data of the multi-energy operation load influencing factors of the power are analyzed. Environmental factors such as light intensity and wind speed change at the hourly level, and equipment temperature, energy storage SOC, etc. change at the minute level. The local time series hierarchical specificity of the load influencing factors is determined, and time series hierarchical division data including hourly, minutely, and secondly levels is generated. Based on this data, the operation load time series hierarchical memory extraction network rules are designed. The network is divided into three layers: the input layer receives the operation load cognitive rule feature data, the hidden layer extracts the features of the corresponding level by dividing into three levels: hours, minutes, and seconds, and the output layer integrates the features of each level. For the "load increase" rule caused by "light intensity + feeding period" in the rule feature data, the hourly load increase starting period, minute-level load increase rate change, and second-level instantaneous peak occurrence time are extracted. These features are quantified into parameters such as timestamp, rate gradient, and peak duration through network rules to generate operating load transformation time series hierarchical feature data, which fully presents the dynamic load change characteristics under different time scales.

[0059] Step S362: performing fuzzy feature analysis and training optimization processing on the operating load transformation trend based on the characteristic data of the power multi-energy operating load influencing factors and the characteristic data of the operating load transformation time series level to generate optimized operating load transformation trend fuzzy feature data; In this embodiment of the present invention, characteristic data of factors influencing the operating load of a multi-energy source is used as input, and characteristic data of the time series of operating load changes is used as output. A fuzzy logic conversion method is used to perform fuzzy characteristic analysis of the operating load change trend. A three-dimensional fuzzy logic rule base is established, with the input variables being the light intensity change rate, diesel generator load fluctuation value, and energy storage charge and discharge rate. The output variable is the fuzzy classification of the load change trend (sharp increase, slow increase, stable, slow decrease, sharp decrease). The input data is fuzzified, and the light intensity change rate is divided into five fuzzy sets: "negatively large, negatively small, zero, positively small, and positively large," corresponding to different membership functions. Using rules in the rule base, such as "If the light intensity change rate is positively large and the diesel generator load fluctuation value is positively small, then the load trend is slowly increasing," fuzzy inference is performed on the input data to generate fuzzy characteristic data of the operating load change trend. The back-propagation algorithm is used to optimize the membership function parameters of the fuzzy rules, calculate the error between the inference results and the actual time series hierarchical feature data, adjust the center value and width of the membership function to control the error within 5%, and generate fuzzy feature data of the optimized operating load transformation trend to improve the matching degree between the feature and the actual load change.

[0060] Step S363: Transmitting the optimized operating load transformation trend fuzzy feature data to the Gaussian mixture operating load feature model architecture to establish an operating load transformation trend prediction model, thereby generating an operating load transformation trend prediction model.

[0061] In this embodiment of the present invention, optimized fuzzy feature data for operational load change trends is input into a Gaussian mixture operational load feature model architecture, and a probability distribution fitting method is used to establish an operational load change trend prediction model. The model's input layer receives the membership values ​​from the fuzzy feature data. The mixture layer assigns probability weights to each fuzzy feature based on the determined optimal number of components (e.g., three components for a load increase). The output layer calculates the posterior probability of each load trend category. For "sharply rising" fuzzy features, the model calculates the probability of them belonging to each Gaussian component and determines the typical load variation curve corresponding to this feature (e.g., a 20% increase within 10 minutes). For "stable" fuzzy features, the probability distributions under each component are fitted to determine the load fluctuation range (e.g., within ±3%). By iteratively adjusting the model's mean and covariance parameters, the model output probability distribution is aligned with the actual distribution of the optimized fuzzy feature data, ultimately generating an operational load change trend prediction model. Based on the input fuzzy feature data, this model outputs the specific probabilities of a sharp rise, slow rise, stability, slow decline, or sharp decline in the load over the next 15 minutes, providing clear trend guidance for energy regulation.

[0062] Furthermore, step S361 includes the following steps: Based on the characteristic data of the load influencing factors of the power multi-energy operation, the local time series hierarchical specificity analysis of the load influencing factors is performed to generate the local time series hierarchical specificity data of the load influencing factors. The local time series hierarchical specificity data of the load influencing factors is then used to design the network rules for the operation load time series hierarchical memory extraction. The operating load time series level memory extraction network rule is used to extract the operating load transformation time series level feature of the operating load cognitive rule feature data to obtain the operating load transformation time series level feature data.

[0063] In an embodiment of the present invention, a time series decomposition method is used to conduct a local time series hierarchical specificity analysis of load influencing factors based on the characteristic data of load influencing factors for multi-energy operation of electricity. For the light intensity factor, the levels are divided into 1-hour, 30-minute, and 10-minute time intervals. The rate of change of light intensity at different levels is calculated to determine its specificity of slowly fluctuating at the hourly level during the day and rising in a step-like manner at the 30-minute level during the feeding period. For the diesel generator load factor, the levels are divided into 5-minute, 1-minute, and 30-second time intervals to analyze its specificity of a sudden rise at the 30-second level during the startup phase and a gentle fluctuation at the 5-minute level during the stable operation phase. For the energy storage SOC factor, the levels are divided into 2-hour, 1-hour, and 20-minute time intervals to determine its specificity of a continuous rise at the 2-hour level during the charging phase and a uniform decrease at the 1-hour level during the discharging phase. These analysis results are quantified into parameters including time intervals, change amplitudes, and durations to generate local time series hierarchical specificity data for load influencing factors. Based on this data, a hierarchical memory extraction network rule for operating load time series was designed. The rule includes hierarchical division criteria (e.g., hourly level corresponds to a ≥60-minute data window), feature extraction indicators (e.g., rate of change, peak frequency), and hierarchical association logic (e.g., minute-level features must match hourly trends), ensuring that the network can extract load features at the corresponding level based on factor specificity. Using the hierarchical memory extraction network rule for operating load time series, the characteristic data of the operating load cognitive rules are hierarchically extracted to obtain hierarchical characteristic data for operating load transformation time series. For the characteristic data of the rule that "load rise" is caused by "light intensity + feeding period," the daily cumulative load increase and the hourly interval of peak occurrence during the rule's effective period are extracted according to the hourly standard in the network rule; the load increase rate and fluctuation amplitude every 30 minutes are extracted according to the 30-minute standard; and the peak load surge and duration at the moment the bait feeder is activated are extracted according to the 10-minute standard. For the characteristic data of "high wind speed + nighttime → increased diesel generator load," we extracted the load ramp rate after diesel generator startup at the 5-minute level; the instantaneous load fluctuation during sudden wind speed changes at the 1-minute level; and the frequency of load oscillations before voltage stabilization at the 30-second level. By summarizing the characteristic parameters extracted at each level, we generated time-series characteristic data for operational load changes, fully presenting the detailed characteristics of load changes at different time scales.

[0064] Furthermore, step S362 includes the following steps: The fuzzy characteristic logic conversion of the operating load transformation trend is performed using the characteristic data of the power multi-energy operating load influencing factors as input data and the corresponding operating load transformation time series level characteristic data as output data to obtain the operating load transformation trend fuzzy characteristic data; The fuzzy trend feature data of the operating load transformation is subjected to fuzzy rule membership mapping through a preset operating load transformation fuzzy logic rule library to generate operating load transformation trend fuzzy feature data of the membership mapping, and the fuzzy feature data of the operating load transformation trend of the membership mapping is trained and optimized through a back propagation algorithm to generate optimized operating load transformation trend fuzzy feature data.

[0065] In an embodiment of the present invention, the light intensity change rate, diesel generator power fluctuation value, and energy storage battery charge and discharge rate in the characteristic data of the electric multi-energy operation load influencing factors are used as input data, and the hourly load change and minute-level rise / fall rate in the corresponding operation load transformation time series level characteristic data are used as output data, and the fuzzy logic conversion method is used to extract the fuzzy features of the operation load transformation trend. For example, the fuzzy classification criteria for input variables are set: the light intensity change rate is divided into “large negative (≤-20% / h), small negative (-20% / h to -5% / h), zero (-5% / h to 5% / h), small positive (5% / h to 20% / h), large positive (≥20% / h)”; the diesel generator power fluctuation value is divided into “violent fluctuation (≥±15%), moderate fluctuation (±5% to ±15%), slight fluctuation (≤±5%)”; the energy storage charge and discharge rate is divided into “fast charging (≥10% / h), slow charging (3% / h to 10% / h), stationary (-3% / h to 3% / h), slow discharge (-10% / h to -3% / h), fast discharge (≤-10% / h)”. The output variable is divided into five fuzzy categories: "sharp rise (≥15% / h), slow rise (5% / h to 15% / h), stable (-5% / h to 5% / h), slow fall (-15% / h to -5% / h), and sharp fall (≤-15% / h). Fuzzy logic operations are used to convert the input data into corresponding output fuzzy categories, generating fuzzy characteristic data for the operating load change trend and establishing a fuzzy association between influencing factors and load trends. Assume that the preset fuzzy logic rule base for operating load change contains 125 rules (5×3×5 input combinations), such as "If the light intensity change rate is positive, the diesel generator power fluctuation is slight, and the energy storage charge and discharge rate is fast, then the load trend is sharply rising." The fuzzy characteristic data for the operating load change trend is input into the rule base, and a membership function is used to calculate the membership value (ranging from 0 to 1) of each input data point to the corresponding fuzzy set. For example, if the light intensity change rate at a certain moment is 25% / h, its membership in the "positive" category is 1, and all other fuzzy sets are 0. Fuzzy reasoning is performed based on the rule base, and the input membership is multiplied by the rule confidence (preset to be 0.8-1.0) to obtain the output membership of each rule. The output of all rules is aggregated using the maximum-minimum synthesis method to generate the fuzzy characteristic data of the load transformation trend of the membership map, which includes the membership distribution of each output fuzzy category. For example, the Takagi-Sugeno-Kang type fuzzifier is used, and the rules in its fuzzy rule base are for: ;in and are the language variables in the antecedent and consequent of the rule respectively; is the input variable language value; is the output exact variable. For a given input x=\left [ {{x}^{*}_{1},...,{x}^{*}_{n}} \right ]^{T} , the exact variable of its output is: {q}^{*}=\frac {{\Sigma}^{l}_{j=1}\left [ {{\Pi}^{n}_{i=1}{u}_{{A}^{j}_{i}}({x}^{*}_{i})\cdot {f}_{j}({x}^{*}_{1},...,{x}^{*}_{n})} \right ]} {{\Sigma}^{l}_{j=1}{\Pi}^{n}_{i=1}{\mu}_{{A}^{j}_{i}}({x}^{*}_{i})}\cdot {x}^{*}_{1},...,{x}^{*}_{n} Represents the actual measured values ​​or specific numerical values ​​of the first to nth input variables, including real-time monitoring data of the rate of change of light intensity in the characteristic data of the influencing factors of the multi-energy operation load, the actual calculated value of the power fluctuation of the diesel generator, and the specific measurement results of the charge and discharge rate of the energy storage battery; It represents the final precise output result of fuzzy reasoning, corresponding to the quantitative value of the operating load transformation trend (such as the load change rate). l represents the total number of rules in the fuzzy rule base, reflecting all possible logical relationships between input variable combinations and output results. j represents the index of the rule, ranging from 1 to 1, and is used to traverse each rule for calculation. n represents the number of input variables, corresponding to the number of influencing factors such as the light intensity change rate and diesel generator power fluctuation value. i represents the index of the input variable, ranging from 1 to n, and is used to distinguish different input variables. It represents the linguistic value corresponding to the i-th input variable in the j-th fuzzy rule. This linguistic value is a qualitative description of the characteristics of the input variable, such as "positive" or "slightly fluctuating". It is used to establish the logical association between the input variable and the output result in the fuzzy reasoning process. Represents the actual value of the i-th input variable Belongs to the corresponding language value in the jth rule The membership degree ranges from 0 to 1, reflecting the matching degree between the input value and the language description. Represents the output function of the jth rule. A backpropagation algorithm is used to train and optimize the membership mapping data. The error between the mapping result and the actual time series hierarchical feature data is calculated. The center value and width of the membership function are adjusted to reduce the error. This generates fuzzy feature data that optimizes the operating load transformation trend, improving the matching accuracy between the fuzzy feature and the actual load changes.

[0066] Furthermore, the step S4 of collecting the real-time load data of the multi-energy source by optimizing the operation of the solar-diesel-storage power energy integration model includes: Based on the optimized operation of the solar-diesel-storage power energy integration model, the power multi-energy operation optimization operation is performed on the deep-sea aquaculture power energy system; the power multi-energy real-time operation load data is collected according to the power multi-energy operation optimization operation.

[0067] In an embodiment of the present invention, based on the optimized operation of the photovoltaic-diesel-storage electric energy integration model, a coordinated operation optimization operation is performed on the photovoltaic equipment, diesel generators, and energy storage batteries of the deep-sea aquaculture electric energy system. The optimized operation of the photovoltaic-diesel-storage electric energy integration model receives the latest user electricity consumption data and performs corresponding electricity optimization. After receiving the power operation load data in the deep-sea aquaculture electric energy system after power optimization, the power sensors installed on the photovoltaic junction box, the output end of the diesel generator, the energy storage battery charge and discharge circuit and the main bus of the system collect the real-time photovoltaic output, the real-time output power of the diesel generator, the real-time charge and discharge power of the energy storage battery, and the total active power, reactive power, voltage, current and other data of the system at a frequency of once per second. These data together constitute the real-time operation load data of the multi-energy power, which fully reflects the real-time load status of each energy device in the optimization operation.

[0068] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0069] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A deep learning-based intelligent analysis method for deep-sea aquaculture power energy load, characterized by: The following steps are involved: Step S1: Acquire the power multi-energy equipment data and user power consumption data of the deep-sea aquaculture power energy system; Analyze the user's electricity consumption characteristics based on the user's electricity consumption data to generate user's electricity consumption characteristic data; Step S2: establishing a demand-target-operation mapping relationship specific to the power multi-energy device based on the power multi-energy device data, and generating a demand-target-operation power energy integration model; Based on the user's electricity consumption characteristics data and the demand target operation power energy integration model, a solar-diesel-storage power energy integration model with optimized operation of the user's electricity consumption characteristics is established to generate an optimized operation solar-diesel-storage power energy integration model; Step S3: Acquire multimodal operation perception element data of the deep-sea aquaculture power energy system; perform characteristic analysis of the influencing factors of the power multi-energy operation load based on the multimodal operation perception element data to generate characteristic data of the influencing factors of the power multi-energy operation load; Establish a trend prediction model for operating load transformation based on the characteristic data of the influencing factors of the multi-energy operating load of the power industry, and generate an operating load transformation trend prediction model; Step S4: Collecting real-time load data of multiple energy sources by optimizing and operating the solar-diesel-storage power energy integration model; Transmitting the real-time operating load data of the electric power multi-energy to the operating load transformation trend prediction model to perform electric power energy operating load trend characteristic analysis and generate electric power energy operating load trend characteristic data; Based on the power energy operation load trend characteristic data, the power energy load collaborative control operation of deep-sea aquaculture is carried out.

2. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire the power multi-energy equipment data and user power consumption data of the deep-sea aquaculture power energy system; Step S12: Analyze the electricity consumption period according to the user's electricity consumption data to generate electricity consumption period data, and perform time series slicing of the electricity load according to the period difference on the user's electricity consumption data using the electricity consumption period data to generate time series slicing data of the user's electricity load according to the period difference; Step S13: extracting load time series characteristics based on the time series slice data of the user's electricity load in different periods to obtain the time series characteristic data of the user's electricity load in different periods; Step S14: performing a cluster analysis of user electricity consumption behaviors based on the time series characteristic data of user electricity loads in different periods to generate user electricity consumption behavior cluster data; Step S15: performing power usage pattern correlation feature analysis based on the user power usage behavior clustering data to generate power usage pattern correlation feature data; Step S16: performing a user power consumption characteristic analysis on the time series characteristic data of the user power load in different periods by using the power consumption pattern association characteristic data to generate user power consumption characteristic data.

3. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 2 is characterized in that: The electric multi-energy equipment data in step S11 includes photovoltaic electric energy equipment data, diesel electric energy equipment data and energy storage electric energy equipment data.

4. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: performing power multi-energy topology structure analysis based on power multi-energy equipment data to generate power multi-energy topology structure data; Step S22: performing power multi-energy connection relationship analysis on the power multi-energy device data to generate power multi-energy connection relationship data; Step S23: performing network topology modeling of the power multi-energy connection based on the power multi-energy connection relationship data and the power multi-energy topology structure data to generate a power multi-energy connection network topology model; Step S24: performing energy-specific mechanism operation analysis based on the electric multi-energy connection network topology model to generate energy-specific mechanism operation data; Step S25: performing mechanism operation constraint identification on the energy-specific mechanism operation data to generate constrained energy-specific mechanism operation data; Step S26: Analyzing the power multi-energy optimization operation objective function based on the power multi-energy connection network topology model to generate the power multi-energy optimization operation objective function; Step S27: Transmitting the constraint energy specific mechanism operation data and the power multi-energy optimization operation objective function to the power multi-energy connection network topology model for mapping the operation constraint conditions and the optimization operation objective function to generate a demand target operation power energy integration model; Step S28: Performing a boundary condition analysis of multiple power usage scenarios based on the user's power usage characteristic data to generate boundary condition data of multiple power usage scenarios; Step S29: Transmitting the boundary condition data of the electricity consumption multi-scenario to the demand target operation power energy integration model to perform intelligent optimization processing of the operation parameters of the solar-diesel-storage power energy integration, and generating an optimized operation solar-diesel-storage power energy integration model.

5. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 4 is characterized in that: Step S29 includes the following steps: Transmitting the boundary condition data of multiple power consumption scenarios to the demand target operation power energy integration model to perform power energy integration simulation operation characteristic analysis of each scenario and generate power energy integration simulation operation characteristic data; According to the power energy integration simulation operation characteristic data, the power energy integration optimization configuration parameters of each scenario are analyzed to generate the power energy integration optimization configuration parameters. The power energy integration optimization configuration parameters are used to intelligently optimize the operation parameters of the solar-diesel-storage power energy integration of the demand target operation power energy integration model to generate the optimized operation solar-diesel-storage power energy integration model.

6. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: Acquire multimodal operation perception element data of the deep-sea aquaculture power energy system, wherein the multimodal operation perception element data includes environmental perception data, power multi-energy operation load data, and power multi-energy equipment status data; Step S32: preprocessing the standard multimodal operation perception element data to obtain the standard multimodal operation perception element data; Step S33: performing load correlation data analysis of electric multi-energy operation based on the standard multi-modal operation perception element data to generate electric multi-energy operation load correlation data, and performing electric multi-energy operation load influencing factor characteristic analysis on the electric multi-energy operation load correlation data to generate electric multi-energy operation load influencing factor characteristic data; Step S34: performing an operation load cognitive rule feature analysis based on the preset power multi-energy load operation cognitive architecture and the power multi-energy load operation influencing factor feature data to generate operation load cognitive rule feature data; Step S35: performing Gaussian mixture analysis on the operating load cognitive rule feature data to generate the optimal number of operating load components, and establishing a Gaussian mixture model architecture of the operating load characteristics by configuring a preset Gaussian mixture algorithm based on the optimal number of operating load components to obtain a Gaussian mixture operating load characteristic model architecture; Step S36: Based on the characteristic data of the power multi-energy operation load influencing factors and the characteristic data of the operation load cognitive rules, a trend prediction model of the operation load transformation is established for the Gaussian mixture operation load characteristic model architecture to generate an operation load transformation trend prediction model.

7. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 6 is characterized in that: Step S36 includes the following steps: Step S361: extracting the operating load transformation time series hierarchical features from the operating load recognition rule feature data to obtain operating load transformation time series hierarchical feature data; Step S362: performing fuzzy feature analysis and training optimization processing on the operating load transformation trend based on the characteristic data of the power multi-energy operating load influencing factors and the characteristic data of the operating load transformation time series level to generate optimized operating load transformation trend fuzzy feature data; Step S363: Transmitting the optimized operating load transformation trend fuzzy feature data to the Gaussian mixture operating load feature model architecture to establish an operating load transformation trend prediction model, thereby generating an operating load transformation trend prediction model.

8. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 7 is characterized in that: Step S361 includes the following steps: Based on the characteristic data of the load influencing factors of the power multi-energy operation, the local time series hierarchical specificity analysis of the load influencing factors is performed to generate the local time series hierarchical specificity data of the load influencing factors. The local time series hierarchical specificity data of the load influencing factors is then used to design the network rules for the operation load time series hierarchical memory extraction. The operating load time series level memory extraction network rule is used to extract the operating load transformation time series level feature of the operating load cognitive rule feature data to obtain the operating load transformation time series level feature data.

9. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 7 is characterized in that: Step S362 includes the following steps: The fuzzy characteristic logic conversion of the operating load transformation trend is performed using the characteristic data of the power multi-energy operating load influencing factors as input data and the corresponding operating load transformation time series level characteristic data as output data to obtain the operating load transformation trend fuzzy characteristic data; The fuzzy trend feature data of the operating load transformation is subjected to fuzzy rule membership mapping through a preset operating load transformation fuzzy logic rule library to generate operating load transformation trend fuzzy feature data of the membership mapping, and the fuzzy feature data of the operating load transformation trend of the membership mapping is trained and optimized through a back propagation algorithm to generate optimized operating load transformation trend fuzzy feature data.

10. The deep learning-based intelligent analysis method for deep-sea aquaculture power energy load according to claim 1 is characterized in that: The step S4 of collecting the real-time load data of multiple energy sources by optimizing the operation of the solar-diesel-storage power energy integration model includes: Based on the optimized operation of the solar-diesel-storage power energy integration model, the power multi-energy operation optimization operation is performed on the deep-sea aquaculture power energy system; the power multi-energy real-time operation load data is collected according to the power multi-energy operation optimization operation.

Images