Energy supply and demand balance regulation and control system suitable for electric ship

Abstract

The invention discloses an energy supply and demand balance regulation and control system suitable for an electric ship, and the system comprises a data collection and preprocessing module which is used for obtaining and preprocessing multi-source operation data, and forming a basic data set; the energy shape situation field construction module is used for constructing a topological variable energy shape situation field and solving a topological gradient to obtain an energy distribution trend; the ecological super-tensor modeling module is used for constructing a cross-domain energy ecological super-tensor and modeling an environment domain, a propulsion domain, a battery domain and a cabin load domain; the dynamic ecological topology decomposition module is used for generating a cross-domain energy ecological topology structure and forming an analysis result; the energy self-organizing evolution regulation and control module is used for generating an energy regulation and control output result; and the power execution and closed-loop updating module is used for issuing an instruction according to the energy regulation and control output result to realize dynamic updating. According to the invention, through potential field modeling and an ecological regulation and control mechanism, dynamic balance and intelligent allocation of energy supply and demand of the electric ship are realized.

Description

Technical Field

[0001] This invention relates to the field of energy management technology for electric ships, and in particular to an energy supply and demand balance control system suitable for electric ships. Background Technology

[0002] Electric ships, as an important direction for green shipping, typically consist of a propulsion system, onboard load system, and battery energy storage system. Current technologies for energy management of such ships largely rely on fixed-priority load allocation strategies, linear power regulation strategies, or energy control methods based on a single predictive model. While these methods can achieve basic energy allocation under stable operating conditions, their control capabilities are clearly insufficient under conditions of drastic environmental changes, significant load fluctuations, or complex battery states. Existing technologies generally lack the ability to characterize the coupling relationship between environmental factors, load demands, and battery states, making it difficult to accurately reflect the dynamic changes in energy supply and demand throughout the entire voyage of an electric ship.

[0003] With the expansion of electric ship applications, the challenges of wind and wave disturbances, propulsion load fluctuations, and sudden increases in cabin loads during navigation are becoming more common, leading to a strongly coupled, nonlinear, and time-varying energy supply and demand relationship. Traditional energy management methods often exhibit rigid control, delayed response, uneven energy supply, or excessive conservatism when dealing with these complex situations. For example, when multiple loads fluctuate simultaneously, existing methods cannot accurately identify the priority of energy flow, easily resulting in insufficient power supply to critical loads; when battery status changes drastically, traditional methods only adjust based on remaining capacity or fixed thresholds, failing to reflect battery health, temperature, and discharge risks in real time. Existing technologies lack holistic modeling and evolutionary analysis mechanisms for multi-domain energy behavior, making it difficult for the system to achieve globally optimal energy scheduling.

[0004] Therefore, how to provide an energy supply and demand balance control system suitable for electric ships is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] One objective of this invention is to propose an energy supply and demand balance control system suitable for electric ships. This invention fully utilizes topological variable energy form potential field, cross-domain energy ecological supertensor modeling, and energy self-organized evolution control technology. By modeling the multidimensional energy relationships in the environmental domain, propulsion domain, battery domain, and cabin load domain, it performs dynamic ecological topological decomposition and combines energy path evolution, load behavior stage migration, and niche competition regulation to achieve real-time, adaptive, and coordinated control of energy throughout the entire voyage of electric ships. This effectively solves the problem of dynamic imbalance between energy supply and demand under complex sea conditions, multi-source load coupling, and battery state fluctuations, and has the advantages of strong intelligence, high adaptability, and excellent energy allocation efficiency.

[0006] An energy supply and demand balance control system for electric ships according to an embodiment of the present invention includes the following modules: The data acquisition and preprocessing module is used to acquire multi-source operational data during the navigation process of electric ships and preprocess it to form a basic dataset. The energy form potential field construction module is used to generate energy form codes based on the basic dataset, construct a topologically variable energy form potential field and obtain the topological gradient to obtain the energy distribution trend. The Ecological Hypertensor Modeling Module is used to construct cross-domain energy ecological hypertensors based on basic datasets and energy form encoding, and to model the environmental domain, propulsion domain, battery domain and onboard load domain. The dynamic ecological topology decomposition module is used to perform dynamic ecological topology decomposition on the cross-domain energy-ecological hypertensor, generate the cross-domain energy-ecological topology structure, and form the analytical results. The energy self-organization evolution regulation module is used to perform energy path evolution, load behavior stage migration and niche competition regulation based on the analysis results and energy distribution trends, and generate energy regulation output results. The power execution and closed-loop update module is used to issue instructions based on the energy regulation output results and input the executed operating data back to the relevant modules to achieve dynamic updates.

[0007] Optionally, modules can be integrated using the following methods: Acquire multi-source operational data of electric ships during navigation, preprocess the multi-source operational data, and form a basic dataset; Based on the basic dataset, energy is morphologically encoded to generate energy morphological encoding results. A topologically variable energy morphological potential field is constructed, and the topological gradient of the topologically variable energy morphological potential field is obtained to generate energy distribution trends. Based on the basic dataset and energy form encoding results, a cross-domain energy ecology supertensor is constructed to model the environmental domain, propulsion domain, battery domain and onboard load domain, and load behavior stage information and energy form information are written into the tensor multidimensional structure. Dynamic ecological topology decomposition is performed on the cross-domain energy-ecology supertensor to generate the cross-domain energy-ecology topology structure, ecological energy perturbation sensitivity mapping is calculated, energy-ecology attractor regions are identified, and analytical results are generated. The energy supply and demand of electric ships are regulated by an energy self-organizing evolution regulator. The analysis results and energy distribution trends are processed, and energy path evolution, load behavior stage migration and niche competition regulation are performed to obtain the energy regulation output results. Based on the energy regulation results, power regulation commands are issued to adjust the power of the propulsion system, to limit, degrade or shut down the onboard load, to limit the discharge and regulate the thermal management system, and to input the executed operating data back into the energy potential field construction and ecological tensor modeling process for dynamic updates.

[0008] Optionally, the multi-source operating data includes ship speed data, heading data, attitude data, wind speed data, wave height data, tidal current data, battery state of charge data, battery health status data, battery temperature data, propulsion power data, and cabin load power data.

[0009] Optionally, the preprocessing of multi-source operating data includes performing noise reduction, time synchronization, outlier removal, and data format standardization on the collected multi-source operating data.

[0010] Optionally, the step of obtaining the topological gradient of the topologically variable energy potential field to generate the energy distribution trend includes: Read the basic dataset, establish a unified time axis and divide the prediction window according to a fixed step size, calculate the propulsion power change rate, battery state change rate and environmental disturbance change rate to form a morphological input set; Based on the morphological input set, the environmental risk score, load risk score, and battery risk score are calculated for each node, and the node risk score is formed according to the preset weighted combination rules. Based on the time series changes of the morphological input set and historical energy usage records, the propulsion energy demand, cabin load energy demand and battery available energy of the future flight segment are extrapolated and estimated by trend, generating energy prediction results for the future flight segment. Constructing a topologically variable energy potential field, specifically as follows: A two-layer topology network is established, with node risk scores used as node values ​​and risk differences and energy demand changes between adjacent nodes used as edge values. The two-layer topology network structure is dynamically adjusted based on node risk scores and energy prediction results. When a node risk score exceeds the high-risk threshold, the cost of the node-related edge is increased or the lateral connection is canceled. When the risk scores of consecutive nodes are lower than the low-risk threshold, the cost of adjacent edges is reduced or priority edges are inserted. When the energy prediction for future flight segments is lower than the budget threshold, temporally adjacent nodes are aggregated and the cost of the edges connected to the nodes is recalculated. In the deformed two-layer topology network, for each node, the adjacent direction with the largest descent is determined by finite neighborhood comparison and priority queue search as the local gradient direction. The energy distribution trend is obtained by sequentially differentiating the risk assignment of adjacent nodes along the time layer, thus forming the topological gradient field and energy distribution trend.

[0011] Optionally, the construction of the cross-domain energy ecology hypertensor based on the basic dataset and energy form encoding results includes: Establish the dimensional structure of the cross-domain energy ecosystem supertensor, and determine the domain dimension, behavioral stage dimension, energy form dimension, node dimension and time window dimension. The domain dimension includes the environmental domain, propulsion domain, battery domain and onboard load domain. The input records are integrated by node and time window. The environmental change characteristics, propulsion load characteristics, battery status characteristics and cabin load characteristics in the basic dataset are combined with node risk scores, energy prediction results and energy form encoding into structured fields, which serve as the basic data entries for constructing the hypertensor. For the environmental domain, propulsion domain, battery domain and onboard load domain, domain entries are generated respectively. Domain entries include changes in energy demand, changes in available energy, node risk score, energy form intensity and behavior phase indication, where the behavior phase indication is determined based on the time changes of energy form intensity and node risk score. Establish cross-domain relationships between domains. Generate cross-domain coupling entries for any two domains that have energy coupling or risk transmission relationships. Cross-domain coupling entries include the risk difference level, energy demand and supply difference level, and energy form compatibility level between the domains. Based on these three factors, cross-domain mutual conduction strength is formed. The entries within the domain and the entries coupled across the domain are written into the cross-domain energy ecology supertensor in a fixed order according to the domain dimension, behavior stage dimension, energy form dimension, node dimension and time window dimension, respectively, to generate the complete structure of the cross-domain energy ecology supertensor.

[0012] Optionally, forming the parsing result includes: Read the cross-domain energy ecology supertensor and the corresponding cross-domain coupling entries within the time window, and build an initial topology graph according to the time window from early to late and the node index from small to large. The initial topology graph is processed, and the node sequences within the same domain whose risk score fluctuations do not exceed the hysteresis threshold and whose energy demand and available energy difference are within the allowable range are temporally aggregated, and single-point nodes are replaced with stable segments within the domain. Cross-domain edges are sorted in order of mutual conductivity strength from high to low and morphological compatibility from high to low. Edges with mutual conductivity strength or morphological compatibility below the lower limit are pruned. Edges that satisfy both lower limits are retained and recorded as strongly connected edges, forming a cross-domain connected skeleton. Based on the cross-domain connectivity framework, the connectivity stability and energy budget deviation of each connected subgraph are calculated in each time window. Order-preserving matching is performed between adjacent time windows to maximize the subgraph overlap rate, resulting in a time-continuous set of cross-domain topological substructures, which serves as the cross-domain energy ecological topology. An ecological energy disturbance sensitivity mapping is generated on the cross-domain energy ecological topology, specifically as follows: For each node in a finite neighborhood, the risk score, energy requirement, and available energy values ​​of the neighboring nodes connected to the node through strongly connected edges are read. The differences between the current node and its neighboring nodes are accumulated according to preset weights to obtain the node's comprehensive sensitivity measure. Based on preset thresholds and minimum size constraints, connected regions with low sensitivity and high connectivity stability are identified as energy ecological attractor regions; The cross-domain energy-ecological topology, the ecological-energy disturbance sensitivity mapping, and the energy-ecological attractor region are output as analytical results.

[0013] Optionally, obtaining the energy regulation output result includes: Construct an energy self-organizing evolution regulator and set up a processing architecture with structural, behavioral, and ecological layers; In the structural layer, feasible connectivity relationships are selected based on topological gradients. Highly sensitive regions are shielded by combining cross-domain energy ecological topology and ecological energy disturbance sensitivity mapping. Path priority is set for energy ecological attractor regions to generate a set of candidate energy paths. Time continuity verification, power reachability verification, and battery safety boundary verification are performed on the candidate energy path set to output energy path schemes and structural layer constraint sets. In the behavior layer, each load is judged and migrated in stages based on the energy distribution trend and node risk score, forming the stage configuration of embryo, growth, maturity and decay. Hysteresis rules and minimum holding time constraints are applied to the stage configuration. Limiting and degradation are performed on overloaded loads based on the structural layer constraint set, and the stage configuration table and stage power range table are output. In the ecological layer, the load ecological niche priority is determined based on the mutual conduction intensity, morphological compatibility and energy ecological attractor region in the cross-domain energy ecological topology. The resource allocation weight is calculated by combining the stage configuration table and the stage power interval table. The minimum quota for critical loads and the upper limit quota for non-critical loads are set, and the ecological layer allocation constraints and ecological layer priority table are output. The energy path scheme is consistently integrated with the structural layer constraint set, the stage configuration table and the stage power range table, and the ecological layer allocation constraint and the ecological layer priority table. Conflict resolution is performed in the order of safety first, stability second, and efficiency third to generate energy regulation results.

[0014] Optionally, the step of feeding the executed runtime data back into the energy potential field construction and ecological tensor modeling process for dynamic updates includes: The system receives energy regulation results, which include propulsion system power adjustment commands, onboard load power adjustment commands, and battery management system discharge and thermal management adjustment commands, and issues them to the corresponding equipment in a predetermined execution order and effective time window. Execute instructions within the current time window, collect the execution data, and perform time alignment, noise reduction, and outlier removal to form a feedback record with timestamps and device indexes. Based on the equipment's rated constraints and issued instructions, generate expected response values. Compare the expected response values ​​with the feedback records item by item to obtain an execution deviation list. Then, merge the feedback records and execution deviations into the updated basic dataset. The updated base dataset and execution bias are fed back into the energy morphology potential field construction and cross-domain energy ecology supertensor modeling process to regenerate the energy morphology potential field, topological gradient and energy distribution trend, and update the cross-domain energy ecology topology, ecological energy disturbance sensitivity mapping and energy ecology attractor region.

[0015] The beneficial effects of this invention are: This invention constructs a topologically variable energy potential field, enabling real-time mapping of the dynamic changes in energy supply and demand in electric ships under complex navigation environments. Compared to traditional energy regulation methods that rely on single variables or linear rules, the potential field model of this invention can simultaneously reflect the coupled effects of environmental disturbances, load changes, and battery status, allowing energy distribution trends and supply and demand relationships to be presented in a more refined and continuous manner, providing stable and accurate basic data support for regulation decisions.

[0016] By constructing a cross-domain energy-ecological supertensor and performing dynamic ecological topology decomposition, this invention can establish a multi-dimensional energy relationship model among the environmental domain, propulsion domain, battery domain, and onboard load domain. It identifies key energy channels, sensitive areas, and energy-ecological attractors within the system, enabling the system to identify potential risks and stable regions in energy supply and demand under multi-domain interaction conditions. This cross-domain modeling and topology decomposition approach overcomes the limitations of existing technologies in handling multi-domain coupling relationships, shifting energy regulation from local optimization to overall coordination, and providing strong support for achieving stability and continuity in energy management throughout the entire flight.

[0017] By introducing a self-organizing evolutionary regulation mechanism for energy, this invention enables dynamic and coordinated adjustment in energy paths, load behavior stages, and resource competition among loads. This allows the system to automatically adjust its energy allocation strategy in situations of energy shortage, sudden load changes, or rapid changes in battery status, prioritizing critical loads and maintaining overall system stability. The closed-loop feedback mechanism of the execution results allows the regulation strategy to gradually converge and optimize over continuous voyages, forming an adaptive and self-evolving energy management capability. This invention surpasses existing technologies in terms of intelligence, stability, and energy efficiency, providing an effective technical means for electric ships to achieve highly reliable and precise energy supply and demand balance regulation in complex sea conditions. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the structure of an energy supply and demand balance control system for electric ships proposed in this invention; Figure 2 This is a schematic flowchart of an energy supply and demand balance control method for electric ships proposed in this invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0020] refer to Figure 1 An energy supply and demand balance control system suitable for electric ships includes the following modules: The data acquisition and preprocessing module is used to acquire multi-source operational data during the navigation process of electric ships and preprocess it to form a basic dataset. The energy form potential field construction module is used to generate energy form codes based on the basic dataset, construct a topologically variable energy form potential field and obtain the topological gradient to obtain the energy distribution trend. The Ecological Hypertensor Modeling Module is used to construct cross-domain energy ecological hypertensors based on basic datasets and energy form encoding, and to model the environmental domain, propulsion domain, battery domain and onboard load domain. The dynamic ecological topology decomposition module is used to perform dynamic ecological topology decomposition on the cross-domain energy-ecological hypertensor, generate the cross-domain energy-ecological topology structure, and form the analytical results. The energy self-organization evolution regulation module is used to perform energy path evolution, load behavior stage migration and niche competition regulation based on the analysis results and energy distribution trends, and generate energy regulation output results. The power execution and closed-loop update module is used to issue instructions based on the energy regulation output results and input the executed operating data back to the relevant modules to achieve dynamic updates.

[0021] refer to Figure 2 A method for regulating the energy supply and demand balance of electric ships, comprising: Acquire multi-source operational data of electric ships during navigation, preprocess the multi-source operational data, and form a basic dataset; Based on the basic dataset, energy is morphologically encoded to generate energy morphological encoding results. A topologically variable energy morphological potential field is constructed, and the topological gradient of the topologically variable energy morphological potential field is obtained to generate energy distribution trends. Based on the basic dataset and energy form encoding results, a cross-domain energy ecology supertensor is constructed to model the environmental domain, propulsion domain, battery domain and onboard load domain, and load behavior stage information and energy form information are written into the tensor multidimensional structure. Dynamic ecological topology decomposition is performed on the cross-domain energy-ecology supertensor to generate the cross-domain energy-ecology topology structure, ecological energy perturbation sensitivity mapping is calculated, energy-ecology attractor regions are identified, and analytical results are generated. The energy supply and demand of electric ships are regulated by an energy self-organizing evolution regulator. The analysis results and energy distribution trends are processed, and energy path evolution, load behavior stage migration and niche competition regulation are performed to obtain the energy regulation output results. Based on the energy regulation results, power regulation commands are issued to adjust the power of the propulsion system, to limit, degrade or shut down the onboard load, to limit the discharge and regulate the thermal management system, and to input the executed operating data back into the energy potential field construction and ecological tensor modeling process for dynamic updates.

[0022] In this embodiment, the multi-source operating data includes ship speed data, heading data, attitude data, wind speed data, wave height data, tidal current data, battery state of charge data, battery health status data, battery temperature data, propulsion power data, and cabin load power data.

[0023] In this embodiment, the preprocessing of multi-source operational data includes performing noise reduction, time synchronization, outlier removal, and data format standardization on the collected multi-source operational data.

[0024] In this embodiment, the step of obtaining the topological gradient of the topologically variable energy potential field and generating the energy distribution trend includes: Read the basic dataset, establish a unified time axis and divide the prediction window according to a fixed step size, calculate the propulsion power change rate, battery state change rate, and environmental disturbance change rate to form a morphological input set. Specifically, the calculation of the propulsion power change rate, battery state change rate, and environmental disturbance change rate involves: Within each prediction window, the initial and final values ​​of propulsion power are read, and the numerical difference between the two is calculated. The numerical difference is then proportionalized to the window duration, and the processing result is used as a representation of the rate of change of propulsion power within the window to obtain the propulsion power change rate. Within each prediction window, the initial and final values ​​of the battery state of charge and battery health are read, the changes in both within the window are calculated, and the two changes are merged according to a preset weight. The merged value is used as a representation of the rate of change of the overall battery state to obtain the battery state change rate. Within each prediction window, the initial and final values ​​of wind speed, wave height, and tidal current speed are read, and the change amplitude of the three types of disturbances is calculated respectively. The three types of change amplitudes are then processed in a comprehensive manner according to a fixed ratio, and the comprehensive result is used as the rate of change of environmental disturbance intensity to obtain the rate of change of environmental disturbance. Based on the morphological input set, an environmental risk score, a load risk score, and a battery risk score are calculated for each node. These are then combined according to a preset weighted combination rule to form the node risk score. Specifically, the calculation of the environmental risk score, load risk score, and battery risk score involves: Within the time window corresponding to each node, the changes in wind speed, wave height, and tidal current speed are read. The changes in the three types of environmental disturbances are combined according to a preset ratio. The larger the combined value, the stronger the environmental disturbance. The combined result is used as the environmental risk score. Within the time window corresponding to each node, the change in cabin load power, the frequency of load surges, and the degree to which the load approaches the rated power are read. The three indicators are combined with fixed weights. The higher the combined value, the more severe the load fluctuation. The combined result is used as the load risk score. Within the time window corresponding to each node, the changes in battery state of charge, battery health status, and battery temperature are read and weighted according to factors such as whether the battery temperature is close to the safety limit, whether the battery state of charge is in a low range, and whether the battery health has decreased. The overall result is used as the battery risk score. The preset weighted combination rules are as follows: The environmental risk score has a weight of 40%. The load risk score has a weight of 30%. The battery risk score has a weight of 30%. Based on the time-series changes of the morphological input set and historical energy usage records, trend extrapolation and interval estimation are performed on the propulsion energy demand, payload energy demand, and available battery energy for future flight segments to generate energy prediction results for future flight segments. Specifically, the trend extrapolation and interval estimation of propulsion energy demand, payload energy demand, and available battery energy for future flight segments are performed as follows: The direction, rate of change, and periodic fluctuation characteristics of propulsion power in a continuous time window are read and compared with the propulsion energy consumption records under the same speed and sea state in history. Based on the current trend, the propulsion power range for several future windows is extended, and the energy consumption range corresponding to the range is used as the range estimate of future propulsion energy demand. By reading the rising or falling trend of the onboard load power, the load start-stop mode and the phased periodic characteristics of the load in a continuous time window, and combining the historical load power records under similar task conditions, the direction of power change and the stable range of the onboard load in the future window are obtained. The energy consumption range corresponding to the power range is used as the range estimate of the energy demand of the onboard load. Read the trends of battery state of charge, battery health status, and battery temperature within a continuous window, and combine them with the energy output records of the battery under similar discharge rates and temperature conditions in historical flight segments to extend the range of energy that the battery can provide within the future window. Use the energy range as the interval estimation result of the battery's usable energy. Constructing a topologically variable energy potential field, specifically as follows: A two-layer topology network is established, with node risk scores used as node values ​​and risk differences and energy demand changes between adjacent nodes used as edge values. The two-layer topology network structure is dynamically adjusted based on node risk scores and energy prediction results. When a node risk score exceeds the high-risk threshold, the cost of the node-related edge is increased or the lateral connection is canceled. When the risk scores of consecutive nodes are lower than the low-risk threshold, the cost of adjacent edges is reduced or priority edges are inserted. When the energy prediction for future flight segments is lower than the budget threshold, temporally adjacent nodes are aggregated and the cost of the edges connected to the nodes is recalculated. In the deformed two-layer topology network, for each node, finite neighborhood comparison and priority queue search are used to determine the adjacent direction with the largest descent as the local gradient direction. The energy distribution trend is obtained by sequentially differencing the risk assignments of adjacent nodes along the time layer, forming a topological gradient field and energy distribution trend. Specifically, for each node, finite neighborhood comparison and priority queue search are used to determine the adjacent direction with the largest descent as the local gradient direction. In the modified two-layer topology network, a limited neighborhood range is defined for each node. The risk assignment difference of each neighboring node is read from the neighborhood. The decrease rate between the current node and each neighboring node is compared. The decrease rate is placed into a priority queue in descending order. The adjacent nodes with the largest decrease are sequentially taken from the priority queue, and the corresponding connection direction is determined as the local gradient direction of the node. The change of risk assignment is tracked along this direction in the continuous time layer.

[0025] The topological variable energy form potential field is a representation of the energy supply and demand state and risk state during the navigation of electric ships. It discretizes nodes on a unified time axis using a fixed window. Each node is assigned a scalar value similar to potential energy height, based on a node risk score formed by environmental disturbances, cabin load fluctuations, and battery status. The risk difference and energy demand change between adjacent nodes serve as the cost or coupling strength of edges, forming a dynamic graph whose connectivity, edge weights, and node aggregation methods change according to risk thresholds and energy prediction results. The mathematical core uses finite neighborhood comparisons on the dynamic graph to obtain the adjacent direction pointing to the largest risk reduction as a local topological gradient. Sequential differences are then applied to the risk assignments of adjacent nodes along the time layer to form the energy distribution trend. Its engineering counterpart includes multi-source data synchronization and denoising standardization using a sliding window, real-time calculation of risk scores and energy predictions, online reconstruction of a two-layer topology network, and gradient direction determination based on a neighborhood priority queue.

[0026] In this embodiment, the construction of the cross-domain energy ecology hypertensor based on the basic dataset and energy form encoding results includes: Establish the dimensional structure of the cross-domain energy ecosystem supertensor, and determine the domain dimension, behavioral stage dimension, energy form dimension, node dimension and time window dimension. The domain dimension includes the environmental domain, propulsion domain, battery domain and onboard load domain. The input records are integrated by node and time window. The environmental change characteristics, propulsion load characteristics, battery status characteristics and cabin load characteristics in the basic dataset are combined with node risk scores, energy prediction results and energy form encoding into structured fields, which serve as the basic data entries for constructing the hypertensor. For the environmental domain, propulsion domain, battery domain and onboard load domain, domain entries are generated respectively. Domain entries include changes in energy demand, changes in available energy, node risk score, energy form intensity and behavior phase indication, where the behavior phase indication is determined based on the time changes of energy form intensity and node risk score. Establish cross-domain relationships between domains. Generate cross-domain coupling entries for any two domains that have energy coupling or risk transmission relationships. Cross-domain coupling entries include the risk difference level, energy demand and supply difference level, and energy form compatibility level between the domains. Based on these three factors, cross-domain mutual conduction strength is formed. The entries within the domain and the entries coupled across the domain are written into the cross-domain energy ecology supertensor in a fixed order according to the domain dimension, behavior stage dimension, energy form dimension, node dimension and time window dimension, respectively, to generate the complete structure of the cross-domain energy ecology supertensor.

[0027] Energy state encoding is a data transformation method used to convert raw multi-source operational data in a basic dataset into high-dimensional feature vectors that can describe changes in energy state.

[0028] The cross-domain energy ecology supertensor is a tensor model that unifies the environmental domain, propulsion domain, battery domain, and cabin load domain of electric ships in the same high-dimensional structure. It uses a domain dimension, behavior stage dimension, energy form dimension, node dimension, and time window dimension to form an index framework. The information of each node in the domain of each time window is written into the tensor. Cross-domain coupling entries are written for two domains that have coupling or risk transmission. The mutual conduction strength is synthesized by risk difference level, supply and demand difference level, and form compatibility level. In this way, the energy ecology relationship of multi-domain coupling and time evolution is compressed into a computable unified representation.

[0029] In this embodiment, forming the analysis result includes: Read the cross-domain energy ecology supertensor and the corresponding cross-domain coupling entries within the time window, and build an initial topology graph according to the time window from early to late and the node index from small to large. The initial topology graph is processed, and the node sequences within the same domain whose risk score fluctuations do not exceed the hysteresis threshold and whose energy demand and available energy difference are within the allowable range are temporally aggregated, and single-point nodes are replaced with stable segments within the domain. Cross-domain edges are sorted in order of mutual conductivity strength from high to low and morphological compatibility from high to low. Edges with mutual conductivity strength or morphological compatibility below the lower limit are pruned. Edges that satisfy both lower limits are retained and recorded as strongly connected edges, forming a cross-domain connected skeleton. Based on the cross-domain connectivity framework, the connectivity stability and energy budget deviation of each connected subgraph are calculated within each time window. Order-preserving matching is performed between adjacent time windows to maximize the subgraph overlap rate, resulting in a time-continuous set of cross-domain topological substructures, which serve as the cross-domain energy-ecological topology. Based on the cross-domain connectivity framework, the connectivity stability and energy budget deviation of each connected subgraph are calculated within each time window, specifically as follows: Within a specified time window, multiple connected subgraphs are divided based on the edge set of the cross-domain connected skeleton. The connectivity between nodes in each subgraph is read, and the proportion of edges that remain unbroken in adjacent time windows is counted. This proportion is used as the connection stability. The energy demand and available energy of nodes within each subgraph are read, and the cumulative difference between the two is calculated. This cumulative difference is used as the energy income and expenditure deviation. To maximize subgraph overlap, order-preserving matching is performed between adjacent time windows, specifically as follows: In two adjacent time windows, a set of connected subgraphs is extracted. The intersection of the node sets of each pair of subgraphs is compared. The subgraphs with the highest node overlap are selected for matching. During matching, the continuity of subgraph numbers within the time window is maintained. Unmatched subgraphs are supplemented or marked as new subgraphs based on their similarity, forming a set of cross-domain topological substructures with the least change in node composition between consecutive time windows. An ecological energy disturbance sensitivity mapping is generated on the cross-domain energy ecological topology, specifically as follows: For each node in a finite neighborhood, the risk score, energy requirement, and available energy values ​​of the neighboring nodes connected to the node through strongly connected edges are read. The differences between the current node and its neighboring nodes are accumulated according to preset weights to obtain the node's comprehensive sensitivity measure. Based on preset thresholds and minimum size constraints, connected regions with low sensitivity and high connectivity stability are identified as energy ecological attractor regions. The minimum size constraint requires that the number of nodes in the candidate connected region must not be less than the preset minimum number of nodes, and that there must be continuous connectivity between the nodes in the region. When the number of nodes or the connectivity continuity of the connected region does not meet the requirements, the region is excluded from the energy ecological attractor region. The cross-domain energy-ecological topology, the ecological-energy disturbance sensitivity mapping, and the energy-ecological attractor region are output as analytical results.

[0030] An energy ecological attractor refers to a connected region in a cross-domain energy ecological topology that consists of several nodes and simultaneously satisfies low disturbance sensitivity and high connection stability. Sensitivity is obtained by weighted summation of the differences between a node and its strongly connected neighbors in terms of risk score, energy demand, and available energy. Connection stability reflects the degree to which the internal edges of the region remain unbroken in adjacent time windows. Furthermore, minimum scale and continuous connectivity constraints are imposed, thus the region corresponds to a set of energy distributions and coupling relationships that are more stable in time evolution.

[0031] In this embodiment, obtaining the energy regulation output result includes: Construct an energy self-organizing evolution regulator and set up a processing architecture with structural, behavioral, and ecological layers; In the structural layer, feasible connectivity relationships are selected based on topological gradients. Highly sensitive regions are shielded by combining cross-domain energy ecological topology and ecological energy disturbance sensitivity mapping. Path priority is set for energy ecological attractor regions to generate a set of candidate energy paths. Time continuity verification, power reachability verification, and battery safety boundary verification are performed on the candidate energy path set to output energy path schemes and structural layer constraint sets. In the behavioral layer, based on energy distribution trends and node risk scores, each load is classified into stages and migrated accordingly, forming stage configurations of embryonic, growth, maturity, and decay. Hysteresis rules and minimum hold duration constraints are applied to the stage configurations. Overloaded loads are limited and degraded based on the structural layer constraint set. A stage configuration table and a stage power range table are output. Specifically, the stage classification and migration based on energy distribution trends and node risk scores are as follows: In each control cycle, the risk score of each load node and the future energy availability change direction reflected in the energy distribution trend are read. When the energy distribution trend is rising or the risk score is low, the load is judged to be able to enter the growth or maturity stage. When the energy distribution trend is falling or the risk score is rising, the load is judged to be in the decline or low load embryonic stage. After completing the stage determination, based on the stage status of the load in the current control cycle and the stage status of the previous cycle, compare the direction of change in the energy distribution trend with the magnitude of change in the risk score. When the energy support capacity of the node where the load is located is enhanced, it triggers migration to a higher stage; when the energy support capacity is weakened, it triggers migration to a lower stage. During the migration, the corresponding stage power range is adjusted synchronously. Hysteresis rules and minimum hold duration constraints require that each load must maintain a duration of no less than the preset minimum duration after entering a phase. Before the duration is reached, even if the energy distribution trend or risk score deteriorates or improves in a short period of time, the phase change will not be triggered. Only when the minimum hold duration is reached will the decision on whether to perform phase migration be made based on the latest energy distribution trend and risk score. Within the ecological layer, load niche priorities are determined based on the mutual conduction strength, morphological compatibility, and energy ecological attractor regions in the cross-domain energy ecological topology. Resource allocation weights are calculated using the stage configuration table and stage power interval table. Minimum allowances for critical loads and maximum allowances for non-critical loads are set, outputting ecological layer allocation constraints and an ecological layer priority table. Specifically, determining load niche priorities involves: Read the mutual conduction strength and morphological compatibility of the node where the load is located in the cross-domain energy ecological topology. Loads with high mutual conduction strength or morphological compatibility that are closer to the stable energy form are assigned a higher ecological niche level. Analyze the correlation between the node to which the load belongs and the energy ecological attractor region. When a node is in the attractor region for a long time or enters the attractor region multiple times in a continuous time window, it indicates that the load has high energy stability and resource utilization efficiency, thus increasing the ecological niche priority. For loads that are far from the attractor region or frequently leave the attractor region, the ecological niche priority is reduced. Based on the connectivity of the load in the cross-domain topology and its criticality in the structure, if the node where the load is located is on a high connectivity path, maintains a strong and stable connectivity relationship with multiple stable nodes, or plays an important role in energy flow, then the niche priority is increased; if the load node has weak connectivity, contributes little to the structure, or has a limited impact on energy flow, then the niche priority is decreased. The resource allocation weights are calculated by combining the phase configuration table and the phase power range table, specifically as follows: Read the behavior stage of each load from the stage configuration table, set a basic weight for each load according to the power demand level corresponding to the behavior stage, and directly determine the size of the basic weight based on the stage level. Read the power range of each load from the stage power range table, adjust the basic weight according to the upper limit position and range width of the power range, and increase or decrease the weight according to the range value. The basic weights obtained based on the behavior stage are merged with the weights adjusted based on the power range, and the final resource allocation weights are output according to the preset weight superposition order. The energy path scheme is consistently integrated with the structural layer constraint set, the stage configuration table and the stage power range table, and the ecological layer allocation constraint and the ecological layer priority table. Conflict resolution is performed in the order of safety first, stability second, and efficiency third to generate energy regulation results.

[0032] Niche competition regulation refers to treating each load as an individual competing for limited energy resources under the multi-domain coupling relationship characterized by the cross-domain energy ecological topology. The priority of the niche is determined by the mutual conduction strength, morphological compatibility and correlation stability with the energy ecological attractor region in the topology. Then, the power demand corresponding to the load behavior stage is combined to form the resource allocation weight, and the minimum quota for critical loads and the upper limit quota for non-critical loads are set to reflect the priority and inhibition relationship between different loads under the energy supply constraint.

[0033] In this embodiment, the step of inputting the executed running data back into the energy potential field construction and ecological tensor modeling process for dynamic updating includes: The system receives energy regulation results, which include propulsion system power adjustment commands, onboard load power adjustment commands, and battery management system discharge and thermal management adjustment commands, and issues them to the corresponding equipment in a predetermined execution order and effective time window. Execute instructions within the current time window, collect the execution data, and perform time alignment, noise reduction, and outlier removal to form a feedback record with timestamps and device indexes. Based on the equipment's rated constraints and issued instructions, generate expected response values. Compare the expected response values ​​with the feedback records item by item to obtain an execution deviation list. Then, merge the feedback records and execution deviations into the updated basic dataset. The updated base dataset and execution bias are fed back into the energy morphology potential field construction and cross-domain energy ecology supertensor modeling process to regenerate the energy morphology potential field, topological gradient and energy distribution trend, and update the cross-domain energy ecology topology, ecological energy disturbance sensitivity mapping and energy ecology attractor region.

[0034] Example 1:

[0035] To verify the feasibility of this invention in practice, it was applied to an electric marine monitoring vessel with a rated power of 320kW and a capacity of 720kWh lithium battery pack. On September 12, 2025, the vessel conducted an 8-hour observation mission in a certain sea area, covering three typical environmental types: calm sea, slightly turbulent sea, and moderately disturbed sea. During this process, propulsion load, battery temperature, and the power of onboard scientific instruments all fluctuated to varying degrees. Particularly during periods of sudden sea state changes, problems such as uneven energy distribution, increased battery discharge pressure, and untimely reduction of onboard load power arose.

[0036] After the voyage begins, the system first collects multi-source data including ship speed, wind speed, wave height, propulsion power, battery state of charge, battery temperature, and onboard load power. This data is then processed through time alignment, noise reduction, and format standardization to form a basic dataset. During the 09:00 segment, sea conditions were relatively stable, with a wind speed of approximately 5.2 m / s, a wave height of approximately 0.7 meters, and propulsion power maintained between 145 and 160 kW. In this scenario, the system performs morphological encoding of energy, constructs a topologically variable energy morphological potential field, and calculates the topological gradient. It identifies a stable overall energy consumption trend, indicating no triggering of behavioral phase migration or ecological competition regulation. A cross-domain energy ecology supertensor is constructed, incorporating energy behavior characteristics from the propulsion domain, battery domain, environmental domain, and onboard load domain into the tensor structure, providing a data foundation for topological decomposition.

[0037] At 10:20 AM, the vessel entered a slightly turbulent sea area, with wind speeds increasing to 7.4 m / s and wave heights reaching 1.2 meters. Propulsion power surged from 158 kW to 182 kW within a short period. The system performed dynamic ecological topology decomposition, identifying sensitive nodes in the propulsion and battery domains from the hypertensor, calculating ecological energy disturbance sensitivity, and locating energy-ecological attractor regions. Due to intermittent high-power sampling by onboard scientific instruments in this section, onboard load demand increased by approximately 12.6% within 15 minutes. Using traditional fixed-priority control methods, propulsion power would be subject to significant fluctuations due to disturbances. This invention identifies the degree of impact of load changes based on sensitivity mapping and automatically migrates onboard load priorities and behavioral stages through S5's energy self-organizing evolution control. This allows two non-critical loads to transition from the mature stage to the growth or buffer stage, thereby reducing low-priority power and maintaining stable propulsion energy consumption.

[0038] At 13:30, the vessel entered a slightly disturbed sea area with wind speeds reaching 9.1 m / s and wave heights reaching 1.6 meters. During this period, propulsion power showed a significant upward trend, and battery temperature rose from 32.8℃ to 35.4℃. The system automatically adjusted the energy distribution between propulsion and onboard loads through energy path evolution, reducing the power of non-critical onboard loads by approximately 6%–10% without affecting the normal sampling of critical acoustic monitoring equipment. During this phase, the system fine-tuned the resource allocation between loads through niche competition suppression, thereby suppressing the battery discharge rate in high-energy-consumption voyages and keeping the battery temperature within a controllable range. Traditional methods would result in propulsion power fluctuations exceeding 10% under the same conditions, while the method of this invention reduces this fluctuation to approximately 5%.

[0039] Throughout the mission, the system continuously feeds back the execution results, making the potential field structure and hypertensor modeling of the next control cycle more closely match the actual operating state, thereby achieving adaptive optimization of the energy allocation strategy.

[0040] Table 1 Comparison data between the method of the present invention and traditional energy regulation methods

[0041] Table 1 shows that this invention improves the stability of propulsion system power under various sea states. In the stable sea state at 09:00, this invention reduces propulsion power fluctuations to about half that of traditional methods. As sea states intensify, such as during periods of increased wind and waves at 10:30 and 13:30, the fluctuations of traditional methods increase significantly, while this invention maintains them within a controllable range. This indicates that this invention, through the synergistic characterization of the system state by the topologically variable energy form potential field and the cross-domain energy ecological supertensor, can adjust the power supply strategy in a timely manner when external disturbances intensify, thus maintaining higher stability of the propulsion system.

[0042] The comparison of battery temperatures further demonstrates the advantages of this invention in balancing energy supply and demand. During the periods of 13:30 and 15:00, when wind and wave disturbances are strong, traditional methods cause a significant increase in battery temperature, while this invention effectively suppresses the rate of temperature rise, maintaining the temperature within a safer operating range. This invention identifies energy stress sources through dynamic ecological topology decomposition and redistributes load behavior and energy flow paths during the regulation phase, reducing the instantaneous discharge load of the battery system and improving operational safety.

[0043] The number of payload reduction events best reflects the supply-demand matching capability. During three typical disturbances, traditional methods generated multiple reduction events, while the present invention did not experience any payload power loss throughout the entire voyage. This demonstrates that the present invention can coordinate the energy demands of the propulsion domain, battery domain, and payload domain, enabling limited energy to continuously and reliably cover the mission payload even under complex sea conditions. Overall, the present invention outperforms traditional solutions in terms of propulsion stability, battery safety, and payload continuity, achieving a higher level of energy supply-demand balance control.

[0044] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An energy supply and demand balance control system suitable for electric ships, characterized in that, Includes the following modules: The data acquisition and preprocessing module is used to acquire multi-source operational data during the navigation process of electric ships and preprocess it to form a basic dataset. The energy form potential field construction module is used to generate energy form codes based on the basic dataset, construct a topologically variable energy form potential field and obtain the topological gradient to obtain the energy distribution trend. The Ecological Hypertensor Modeling Module is used to construct cross-domain energy ecological hypertensors based on basic datasets and energy form encoding, and to model the environmental domain, propulsion domain, battery domain and onboard load domain. The dynamic ecological topology decomposition module is used to perform dynamic ecological topology decomposition on the cross-domain energy-ecological hypertensor, generate the cross-domain energy-ecological topology structure, and form the analytical results. The energy self-organization evolution regulation module is used to perform energy path evolution, load behavior stage migration and niche competition regulation based on the analysis results and energy distribution trends, and generate energy regulation output results. The power execution and closed-loop update module is used to issue instructions based on the energy regulation output results and input the executed operating data back to the relevant modules to achieve dynamic updates.

2. A method for energy supply and demand balance regulation suitable for electric ships, applied to the energy supply and demand balance regulation system for electric ships as described in claim 1, characterized in that, include: Acquire multi-source operational data of electric ships during navigation, preprocess the multi-source operational data, and form a basic dataset; Based on the basic dataset, energy is morphologically encoded to generate energy morphological encoding results. A topologically variable energy morphological potential field is constructed, and the topological gradient of the topologically variable energy morphological potential field is obtained to generate energy distribution trends. Based on the basic dataset and energy form encoding results, a cross-domain energy ecology supertensor is constructed to model the environmental domain, propulsion domain, battery domain and onboard load domain, and load behavior stage information and energy form information are written into the tensor multidimensional structure. Dynamic ecological topology decomposition is performed on the cross-domain energy-ecology supertensor to generate the cross-domain energy-ecology topology structure, ecological energy perturbation sensitivity mapping is calculated, energy-ecology attractor regions are identified, and analytical results are generated. The energy supply and demand of electric ships are regulated by an energy self-organizing evolution regulator. The analysis results and energy distribution trends are processed, and energy path evolution, load behavior stage migration and niche competition regulation are performed to obtain the energy regulation output results. Based on the energy regulation results, power regulation commands are issued to adjust the power of the propulsion system, to limit, degrade or shut down the onboard load, to limit the discharge and regulate the thermal management system, and to input the executed operating data back into the energy potential field construction and ecological tensor modeling process for dynamic updates.

3. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The multi-source operational data includes ship speed data, heading data, attitude data, wind speed data, wave height data, tidal current data, battery state of charge data, battery health status data, battery temperature data, propulsion power data, and cabin load power data.

4. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The preprocessing of multi-source operational data includes performing noise reduction, time synchronization, outlier removal, and data format standardization on the collected multi-source operational data.

5. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The process of obtaining the topological gradient of the topologically variable energy potential field and generating the energy distribution trend includes: Read the basic dataset, establish a unified time axis and divide the prediction window according to a fixed step size, calculate the propulsion power change rate, battery state change rate and environmental disturbance change rate to form a morphological input set; Based on the morphological input set, the environmental risk score, load risk score, and battery risk score are calculated for each node, and the node risk score is formed according to the preset weighted combination rules. Based on the time series changes of the morphological input set and historical energy usage records, the propulsion energy demand, cabin load energy demand and battery available energy of the future flight segment are extrapolated and estimated by trend, generating energy prediction results for the future flight segment. Constructing a topologically variable energy potential field, specifically as follows: A two-layer topology network is established, with node risk scores used as node values ​​and risk differences and energy demand changes between adjacent nodes used as edge values. The two-layer topology network structure is dynamically adjusted based on node risk scores and energy prediction results. When a node risk score exceeds the high-risk threshold, the cost of the node-related edge is increased or the lateral connection is canceled. When the risk scores of consecutive nodes are lower than the low-risk threshold, the cost of adjacent edges is reduced or priority edges are inserted. When the energy prediction for future flight segments is lower than the budget threshold, temporally adjacent nodes are aggregated and the cost of the edges connected to the nodes is recalculated. In the deformed two-layer topology network, for each node, the adjacent direction with the largest descent is determined by finite neighborhood comparison and priority queue search as the local gradient direction. The energy distribution trend is obtained by sequentially differentiating the risk assignment of adjacent nodes along the time layer, thus forming the topological gradient field and energy distribution trend.

6. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The construction of the cross-domain energy ecology hypertensor based on the basic dataset and energy form encoding results includes: Establish the dimensional structure of the cross-domain energy ecosystem supertensor, and determine the domain dimension, behavioral stage dimension, energy form dimension, node dimension and time window dimension. The domain dimension includes the environmental domain, propulsion domain, battery domain and onboard load domain. The input records are integrated by node and time window. The environmental change characteristics, propulsion load characteristics, battery status characteristics and cabin load characteristics in the basic dataset are combined with node risk scores, energy prediction results and energy form encoding into structured fields, which serve as the basic data entries for constructing the hypertensor. For the environmental domain, propulsion domain, battery domain and onboard load domain, domain entries are generated respectively. Domain entries include changes in energy demand, changes in available energy, node risk score, energy form intensity and behavior phase indication, where the behavior phase indication is determined based on the time changes of energy form intensity and node risk score. Establish cross-domain relationships between domains. Generate cross-domain coupling entries for any two domains that have energy coupling or risk transmission relationships. Cross-domain coupling entries include the risk difference level, energy demand and supply difference level, and energy form compatibility level between the domains. Based on these three factors, cross-domain mutual conduction strength is formed. The entries within the domain and the entries coupled across the domain are written into the cross-domain energy ecology supertensor in a fixed order according to the domain dimension, behavior stage dimension, energy form dimension, node dimension and time window dimension, respectively, to generate the complete structure of the cross-domain energy ecology supertensor.

7. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The formation of the analysis result includes: Read the cross-domain energy ecology supertensor and the corresponding cross-domain coupling entries within the time window, and build an initial topology graph according to the time window from early to late and the node index from small to large. The initial topology graph is processed, and the node sequences within the same domain whose risk score fluctuations do not exceed the hysteresis threshold and whose energy demand and available energy difference are within the allowable range are temporally aggregated, and single-point nodes are replaced with stable segments within the domain. Cross-domain edges are sorted in order of mutual conductivity strength from high to low and morphological compatibility from high to low. Edges with mutual conductivity strength or morphological compatibility below the lower limit are pruned. Edges that satisfy both lower limits are retained and recorded as strongly connected edges, forming a cross-domain connected skeleton. Based on the cross-domain connectivity framework, the connectivity stability and energy budget deviation of each connected subgraph are calculated in each time window. Order-preserving matching is performed between adjacent time windows to maximize the subgraph overlap rate, resulting in a time-continuous set of cross-domain topological substructures, which serves as the cross-domain energy ecological topology. An ecological energy disturbance sensitivity mapping is generated on the cross-domain energy ecological topology, specifically as follows: For each node in a finite neighborhood, the risk score, energy requirement, and available energy values ​​of the neighboring nodes connected to the node through strongly connected edges are read. The differences between the current node and its neighboring nodes are accumulated according to preset weights to obtain the node's comprehensive sensitivity measure. Based on preset thresholds and minimum size constraints, connected regions with low sensitivity and high connectivity stability are identified as energy ecological attractor regions; The cross-domain energy-ecological topology, the ecological-energy disturbance sensitivity mapping, and the energy-ecological attractor region are output as analytical results.

8. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The obtained energy regulation output results include: Construct an energy self-organizing evolution regulator and set up a processing architecture with structural, behavioral, and ecological layers; In the structural layer, feasible connectivity relationships are selected based on topological gradients. Highly sensitive regions are shielded by combining cross-domain energy ecological topology and ecological energy disturbance sensitivity mapping. Path priority is set for energy ecological attractor regions to generate a set of candidate energy paths. Time continuity verification, power reachability verification, and battery safety boundary verification are performed on the candidate energy path set to output energy path schemes and structural layer constraint sets. In the behavior layer, each load is judged and migrated in stages based on the energy distribution trend and node risk score, forming the stage configuration of embryo, growth, maturity and decay. Hysteresis rules and minimum holding time constraints are applied to the stage configuration. Limiting and degradation are performed on overloaded loads based on the structural layer constraint set, and the stage configuration table and stage power range table are output. In the ecological layer, the load ecological niche priority is determined based on the mutual conduction intensity, morphological compatibility and energy ecological attractor region in the cross-domain energy ecological topology. The resource allocation weight is calculated by combining the stage configuration table and the stage power interval table. The minimum quota for critical loads and the upper limit quota for non-critical loads are set, and the ecological layer allocation constraints and ecological layer priority table are output. The energy path scheme is consistently integrated with the structural layer constraint set, the stage configuration table and the stage power range table, and the ecological layer allocation constraint and the ecological layer priority table. Conflict resolution is performed in the order of safety first, stability second, and efficiency third to generate energy regulation results.

9. The energy supply and demand balance control method applicable to electric ships according to claim 2, characterized in that, The step of feeding the executed operational data back into the energy potential field construction and ecological tensor modeling process for dynamic updates includes: The system receives energy regulation results, which include propulsion system power adjustment commands, onboard load power adjustment commands, and battery management system discharge and thermal management adjustment commands, and issues them to the corresponding equipment in a predetermined execution order and effective time window. Execute instructions within the current time window, collect the execution data, and perform time alignment, noise reduction, and outlier removal to form a feedback record with timestamps and device indexes. Based on the equipment's rated constraints and issued instructions, generate expected response values. Compare the expected response values ​​with the feedback records item by item to obtain an execution deviation list. Then, merge the feedback records and execution deviations into the updated basic dataset. The updated base dataset and execution bias are fed back into the energy morphology potential field construction and cross-domain energy ecology supertensor modeling process to regenerate the energy morphology potential field, topological gradient and energy distribution trend, and update the cross-domain energy ecology topology, ecological energy disturbance sensitivity mapping and energy ecology attractor region.

Images