Intelligent charging and discharging strategy optimization method for ship battery

Abstract

The invention discloses an intelligent charging and discharging strategy optimization method for a ship battery, and particularly relates to the technical field of battery management. Propulsion power data, auxiliary equipment power data and environment disturbance parameters generated in the ship operation process are collected, a high-frequency time sequence working condition input sequence is built, a load nonlinear evolution model is built through a multi-scale dynamic modeling algorithm, and a disturbance perception prediction sequence is generated; a heterogeneous state characteristic matrix is constructed by combining historical charging and discharging state records of a target battery and current sensing data, voltage sensing data and temperature sensing data, and then the state of charge and the state of health of the battery are evaluated; generating a multi-dimensional strategy candidate set by constructing an energy scheduling constraint map according to the evaluation results of the state of charge and the state of health of the battery; and selecting an optimal strategy path from the multi-dimensional strategy candidate set, outputting a target charging and discharging control instruction, and realizing comprehensive optimization control of dynamic sensing, state driving and strategy self-adaption of the ship battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and more specifically, to an intelligent charging and discharging strategy optimization method for ship batteries. Background Art

[0002] Existing ship battery energy management methods mostly rely on static load prediction and fixed threshold control strategies, usually assuming that the navigation conditions change slowly and external disturbances can be ignored.

[0003] In actual operation, the propulsion load, auxiliary equipment load, and environmental factors show high-frequency and non-linear fluctuations, resulting in the difficulty of existing ship battery energy management methods to respond to power mutations in a timely manner and accurately reflect the remaining energy and reliability of the battery, and unable to maintain accuracy in complex navigation scenarios, reducing the efficiency of the power system and accelerating battery degradation.

[0004] To solve the above problems, a technical solution is provided now. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent charging and discharging strategy optimization method for ship batteries to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions: An intelligent charging and discharging strategy optimization method for ship batteries, comprising the following steps: S1: Collect real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship, and construct a high-frequency time-series working condition input sequence; S2: Based on the high-frequency time-series working condition input sequence, use a multi-scale dynamic modeling algorithm to establish a load non-linear evolution model, and output a disturbance perception prediction sequence; S3: Obtain the historical charging and discharging state records of the target battery, and combine the current current, voltage, and temperature signals of the target battery during navigation to construct a heterogeneous state feature matrix; S4: Based on the heterogeneous state feature matrix, evaluate the current state of charge and health state of the target battery; S5: According to the disturbance perception prediction sequence, the evaluation results of the state of charge and health state of the battery, construct an energy scheduling constraint map, and generate a multi-dimensional strategy candidate set; S6: Select the optimal strategy path from the multi-dimensional strategy candidate set and output the target charging and discharging control instruction.

[0007] In a preferred embodiment, S1 is specifically: Synchronously collect real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship at the same sampling frequency; Perform data alignment processing on the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data; Integrate the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data after data alignment processing in sequence according to the same time stamp to construct a high-frequency time-series operating condition input sequence.

[0008] In a preferred embodiment, S2 is specifically: Perform time-domain decomposition on the high-frequency time-series operating condition input sequence according to three preset time windows of short time scale, medium time scale, and long time scale to obtain three subsequences; For each subsequence, establish a corresponding non-linear dynamic sub-model, fit the load evolution curve through regression analysis method, and output the predicted increment of load change at each time scale; Perform weighted synthesis on the predicted increments at the three time scales according to the predetermined fusion weights to obtain a disturbance-aware prediction sequence.

[0009] In a preferred embodiment, S3 is specifically: Obtain the historical charge and discharge state records of the target battery during multiple historical navigation cycles; Collect the real-time current signal, voltage signal, and temperature signal of the target battery during the current navigation process; Perform joint mapping processing on the historical charge and discharge state records, real-time current signal, voltage signal, and temperature signal according to the time consistency standard, and extract the operating feature vectors at multiple time segments; Combine all the operating feature vectors to form a heterogeneous state feature matrix.

[0010] In a preferred embodiment, S4 is specifically: For the operating feature vectors at each time segment, extract the current feature components, voltage feature components, and temperature feature components during the current navigation cycle; Perform joint processing on the current feature components, voltage feature components, and temperature feature components according to the preset feature calculation rules, and generate the estimated results of the state of charge and the estimated results of the state of health of the battery at the corresponding time segments through a multi-parameter combination calculation method; Perform sequence integration on the estimated results of the state of charge and the estimated results of the state of health of the battery at all time segments to generate the evaluation output sequence of the state of charge and the evaluation output sequence of the state of health of the current target battery.

[0011] In a preferred embodiment, S5 is specifically: Based on the battery state of charge evaluation output sequence and the battery health state evaluation output sequence, combined with the disturbance perception prediction sequence, define the energy scheduling constraint nodes; In the order of the time series, connect each energy scheduling constraint node with directed edges in sequence to form an energy scheduling constraint graph; For the energy scheduling constraint graph, by traversing the node paths, list the control paths that meet the propulsion power demand and the critical load power demand, and record the corresponding charge and discharge schemes for each control path as policy candidates; Summarize all policy candidates to form a multi-dimensional policy candidate set.

[0012] In a preferred embodiment, S6 is specifically: For each control path in the multi-dimensional policy candidate set, perform path feasibility judgment according to the propulsion power demand and the critical load power demand in the current navigation cycle, and exclude the control paths with power default nodes or thermal safety boundary violations in the energy scheduling constraint graph; For the remaining feasible control paths, perform path scoring based on a preset comprehensive evaluation index; Select the control path with the highest path score, and determine the charge and discharge strategy corresponding to the control path as the optimal strategy path; Generate a target charge and discharge control instruction according to the optimal strategy path.

[0013] The technical effects and advantages of the intelligent charge and discharge strategy optimization method for ship batteries of the present invention: By collecting the propulsion power data, auxiliary equipment power data and environmental disturbance parameters during the operation of the ship, it can comprehensively reflect the actual navigation conditions; by constructing a high-frequency time-series working condition input sequence and adopting a multi-scale dynamic modeling algorithm, it can achieve accurate prediction of non-linear load changes; combined with historical charge and discharge state records and current current, voltage and temperature sensing information, construct a heterogeneous state feature matrix to enhance the dimension and depth of battery state evaluation; use the heterogeneous state feature matrix to evaluate the battery state of charge and health state, and improve the judgment accuracy of the current available capacity and reliability of the battery; based on the disturbance perception prediction sequence and the evaluation results of the battery state of charge and health state, construct an energy scheduling constraint graph, which can flexibly adjust the strategy search range according to actual changes and enhance the adaptability of strategy generation; finally, select the optimal path from multi-dimensional candidate strategies to output control instructions to ensure the extension of battery service life while meeting the power demand, and improve the intelligent level and operation efficiency of the ship energy management system. Brief Description of the Drawings

[0014] Figure 1 It is a schematic diagram of an intelligent charge and discharge strategy optimization method for ship batteries of the present invention. Detailed Embodiments

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0016] Embodiment

[0017] Figure 1 A method for optimizing the intelligent charging and discharging strategy of a ship battery according to the present invention is given, which includes the following steps: S1: Collect the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship, and construct a high-frequency time-series operating condition input sequence; S2: Based on the high-frequency time-series operating condition input sequence, use a multi-scale dynamic modeling algorithm to establish a load nonlinear evolution model and output a disturbance perception prediction sequence; S3: Obtain the historical charging and discharging state records of the target battery, and combine the current current, voltage, and temperature signals of the target battery during navigation to construct a heterogeneous state feature matrix; S4: Based on the heterogeneous state feature matrix, evaluate the current state of charge and health state of the target battery; S5: According to the disturbance perception prediction sequence, the evaluation results of the state of charge and health state of the battery, construct an energy scheduling constraint map, and generate a multi-dimensional strategy candidate set; S6: Select the optimal strategy path from the multi-dimensional strategy candidate set and output the target charging and discharging control instruction.

[0018] S1: Collect the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship, and construct a high-frequency time-series operating condition input sequence, including: Synchronously collect the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship at the same sampling frequency; During the actual navigation of a ship, multiple types of sensors are used to synchronously collect real-time power data of the ship's propulsion system, real-time power data of auxiliary equipment, and environmental disturbance parameter data at the same sampling frequency. Specifically, a power measurement sensor installed on the ship's propulsion shaft is used to collect the power consumption value of the ship's propulsion system during navigation in real time; dedicated power measurement sensors distributed on various main auxiliary equipment of the ship, such as radar, communication equipment, navigation instruments, etc., are used to synchronously collect the real-time power consumption data of the auxiliary equipment during navigation in real time; environmental parameter monitoring sensors installed on the hull or deck are used to collect external environmental disturbance parameter data such as wind speed, wind direction, wave height, current speed, and environmental temperature in real time. The sensors perform data synchronization sampling at the same sampling frequency to ensure that each data point can be aligned on a unified time axis.

[0019] Perform data alignment processing on the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data; Adopt an alignment processing method based on timestamp matching to establish a unified timestamp, and match the collected ship propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data with the timestamp respectively; if a sensor data point fails to accurately match the corresponding timestamp, it is filled through data interpolation processing of adjacent time points. The data interpolation processing method adopted is the linear interpolation method, that is, the values of the two nearest collected data points before and after the missing data point are used for linear interpolation to ensure the integrity and continuity of the data.

[0020] Integrate the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data after data alignment processing in sequence according to the same timestamp to construct a high-frequency time-series working condition input sequence; The data corresponding to each unified timestamp constitutes a unified working condition data point. Each working condition data point includes the propulsion power data value, auxiliary equipment power data value, and various environmental disturbance parameter values. All working condition data points are arranged in a unified time order, and finally a high-frequency time-series working condition input sequence is constructed. For example, if the power value of the ship's propulsion system, the overall power value of the auxiliary equipment, and the corresponding environmental disturbance data such as wind speed, wind direction, and wave height are obtained at a certain moment, these data are simultaneously recorded as a complete working condition data point at that moment; and the subsequent moment data is continuously processed in the same way to gradually form a complete high-frequency time-series working condition input sequence.

[0021] S2: Based on the high-frequency time-series working condition input sequence, use a multi-scale dynamic modeling algorithm to establish a load nonlinear evolution model and output a disturbance perception prediction sequence, including: Decompose the high-frequency time-series working condition input sequence in the time domain according to three preset time windows of short time scale, medium time scale, and long time scale to obtain three subsequences; The short - time scale is used to capture the subtle changes in the ship's propulsion power, auxiliary equipment power, and environmental disturbance parameters under short - term mutation states. The medium - time scale is used to analyze the periodic fluctuation characteristics during the ship's operation. The long - time scale is used to reflect the long - term operation trend and load change law of the ship. Specifically, it is gradually divided on the high - frequency time - series working condition input sequence by means of a sliding window. For example, the short - time scale window corresponds to the ship's navigation data from several seconds to several minutes, the medium - time scale window corresponds to the navigation data from dozens of minutes to several hours, and the long - time scale window corresponds to the navigation data from several hours to dozens of hours. The data within each scale window is completely extracted to form the corresponding subsequence.

[0022] For each subsequence, a corresponding non - linear dynamic sub - model is established respectively, and the load evolution curve is fitted by the regression analysis method to output the predicted increment of load change at each time scale; A non - linear dynamic sub - model is established for each subsequence obtained by time - domain decomposition to predict the load change trend. The non - linear dynamic sub - model adopted is a multiple non - linear regression model, where the independent variables of the model are the ship's propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data within the corresponding subsequence, and the dependent variable is the predicted power change increment at a future moment in the load evolution curve. For each subsequence, data normalization pre - processing is first carried out to eliminate the dimensional differences between variables. The multiple non - linear regression analysis algorithm is applied to train the pre - processed subsequence data, and the non - linear load evolution curve is obtained by least - squares fitting. The calculation method of the predicted increment of the load evolution curve is: based on the known historical data and the non - linear regression model, predict the power value difference between a specific future moment and the current moment to form the short - time scale prediction increment, medium - time scale prediction increment, and long - time scale prediction increment.

[0023] The predicted increments at the three time scales are weighted and synthesized according to the predetermined fusion weights to obtain the disturbance - aware prediction sequence; The predicted load change increments obtained under the above three time scales are weighted and synthesized according to pre-determined fusion weights to obtain a disturbance perception prediction sequence. Specifically, the pre-set fusion weights are determined based on statistical analysis of historical navigation data and error assessment. Among them, the short-term scale prediction increment has a relatively high weight because it is more sensitive to immediate changes. The medium-term scale prediction increment is given the second-highest weight because it takes into account both transient and trend change characteristics. The long-term scale prediction increment is set to a relatively low weight because it reflects the long-term trend change characteristics. The specific calculation method is as follows: multiply the short-term scale prediction increment by the corresponding larger weight, multiply the medium-term scale prediction increment by the second-largest weight, multiply the long-term scale prediction increment by the smaller weight, and then add the above three weighted results to form a disturbance perception prediction sequence. For example, when the navigation state changes suddenly, the contribution of the short-term scale prediction increment is greater, so that the disturbance perception prediction sequence can quickly respond to the load change; while during long-distance stable navigation, the contribution of the long-term scale prediction increment is relatively obvious, making the disturbance perception prediction sequence smoothly reflect the long-term trend. It can take into account the prediction characteristics under different time scales, enhance the flexibility and adaptability of the prediction, and overall improve the accuracy and reliability of the disturbance perception prediction, providing a prediction basis and decision-making support for optimizing the ship battery charging and discharging strategy.

[0024] S3: Obtain the historical charge and discharge state records of the target battery, and combine the current, voltage, and temperature signals of the target battery during the current navigation process to construct a heterogeneous state feature matrix, including: Obtain the historical charge and discharge state records of the target battery in multiple historical navigation cycles; The historical charge and discharge state records are the complete historical data obtained by real-time monitoring of the charging and discharging processes of the target battery during multiple previous voyages of the ship. Through a dedicated historical data recording module installed in the battery pack management system, during each historical navigation cycle, the charging current, discharging current, voltage level of the target battery, and the temperature change data of the battery cells and the overall battery pack are continuously recorded at a unified sampling frequency. The historical charge and discharge state records also include long-term tracking parameters such as the battery capacity decay curve, the battery state health index, and the number of charge and discharge cycles during the historical navigation cycle.

[0025] Collect the real-time current signal, voltage signal, and temperature signal of the target battery during the current navigation process; During the current ship navigation, the current signal, voltage signal, and temperature signal of the target battery are collected in real time. The charging or discharging current of the target battery during the current navigation cycle is measured in real time by a high-precision current sensor installed inside the battery management system; the current voltage values of the battery pack and each battery cell are measured in real time by a voltage sensor; the real-time temperature distribution inside the battery cell and the overall battery pack is monitored in real time by a temperature sensor array. The measurement accuracy and sampling frequency of the sensor are maintained consistent with the historical data acquisition method to ensure the comparability and consistency of the data. By adopting a high-precision and multi-point measurement method, the real-time state inside the battery can be comprehensively reflected, effectively improving the accuracy and reliability of battery state estimation.

[0026] The historical charge-discharge state records are jointly mapped and processed with the real-time current signal, voltage signal, and temperature signal according to the time consistency standard to extract the operation feature vectors under multiple time segments; The historical charge-discharge state records within multiple historical navigation cycles obtained are jointly mapped and processed with the current signal, voltage signal, and temperature signal collected in real time during the current navigation process. First, according to the unified time consistency standard, the historical navigation data and the current navigation data are aligned to the same time scale. A synchronous mapping method based on a unified timestamp is adopted, that is, first determine the timestamps of each sampling point of the real-time collected signal during the current navigation cycle, and then find the corresponding or approximate historical charge-discharge data points in the historical data according to the timestamps for matching. If there is no exactly the same timestamp in the historical data, the nearest neighbor interpolation method is used for data filling, that is, select the two historical data points closest to the current sampling point in time, and calculate the historical value corresponding to the real-time sampling point through linear interpolation to complete the time matching of the historical data and the real-time data.

[0027] According to the data after the joint mapping process, the operation feature vectors under multiple time segments are extracted. Each time segment is defined as a window of a certain time length, such as several minutes or dozens of minutes as a time segment; within each time segment, the mean value of the current signal, the maximum and minimum values of the current signal, the mean value and fluctuation amplitude of the voltage signal, the mean value and rising and falling trend of the temperature signal and other characteristic parameters are extracted respectively. The fluctuation amplitude of the voltage signal is the difference between the maximum value and the minimum value of the voltage within the time segment; the rising and falling trend of the temperature signal is obtained by comparing the starting temperature and the ending temperature values of the time segment. The extracted characteristic parameters are combined to form an operation feature vector corresponding to the time segment.

[0028] All the operation feature vectors are combined to form a heterogeneous state feature matrix; All the operation feature vectors obtained in different time segments are sequentially combined and arranged to form a heterogeneous state feature matrix. Each row of the heterogeneous state feature matrix represents a specific time segment, and each column represents the specific feature parameters uniformly extracted in each time segment.

[0029] S4: Based on the heterogeneous state feature matrix, evaluate the current state of charge and health state of the target battery, including: For the operation feature vectors in each time segment, extract the current characteristic components of current, voltage, and temperature within the current navigation cycle; The current characteristic components include the average value of current, the maximum value of current, and the minimum value of current; the voltage characteristic components include the average value of voltage and the voltage fluctuation amplitude; the temperature characteristic components include the average value of temperature and the temperature change trend.

[0030] Jointly process the current characteristic components, voltage characteristic components, and temperature characteristic components according to the preset feature calculation rules, and generate the estimated results of the state of charge and the estimated results of the health state of the battery corresponding to each time segment respectively through the multi-parameter combination calculation method; The feature calculation rules are determined by analyzing, summarizing, and verifying a large amount of historical data, and selecting the mathematical relationships that comprehensively consider the internal physical and chemical characteristics of the battery. Specifically, the calculation method of the estimated result of the state of charge of the battery is as follows: First, jointly process the average values of the current characteristic components and the voltage characteristic components, and estimate the initial value of the state of charge of the current battery through the voltage integration method; then, correct and adjust the initial value of the state of charge of the current battery through the temperature characteristic components. The specific correction and adjustment method is: when the average value of the battery temperature is higher than the predetermined standard temperature, the initial value of the state of charge needs to be reduced for compensation; when the average temperature is lower than the standard temperature, the initial value of the state of charge needs to be appropriately increased to ensure the accuracy of the state of charge estimation. The calculation method of the estimated result of the health state of the battery is: Use the voltage fluctuation amplitude of the voltage characteristic components, the maximum value of the current characteristic components, and the temperature change trend of the temperature characteristic components to generate a preliminary estimated result of the health state through a non-linear combination calculation method; then, dynamically correct the preliminary estimated result of the health state according to the health degradation characteristic curve of the battery in the historical navigation cycle to obtain an accurate estimated result of the health state.

[0031] The dynamic correction is specifically as follows: Compare the preliminary estimated result of the health state with the theoretical health state value corresponding to the health degradation characteristic curve of the battery in the historical navigation cycle. If the preliminary estimated result of the health state is significantly higher than the theoretical health state value, it means that the preliminary estimated result of the health state needs to be reduced; if the preliminary estimated result of the health state is significantly lower than the theoretical health state value, it means that the preliminary estimated result of the health state needs to be appropriately increased.

[0032] A correction proportionality coefficient is calculated based on the difference between the theoretical health state value and the preliminary health state estimation result; then, the preliminary health state estimation result is multiplicatively adjusted using the proportionality coefficient. For example, if the preliminary health state estimation result is significantly higher than the theoretical health state value, the preliminary health state estimation result is multiplied by a correction proportionality coefficient less than one; if the preliminary health state estimation result is significantly lower than the theoretical health state value, the preliminary health state estimation result is multiplied by a correction proportionality coefficient greater than one to obtain the final health state estimation result.

[0033] The multi-parameter combination calculation can comprehensively evaluate the real-time state and long-term health condition of the battery, significantly improving the comprehensiveness and reliability of state evaluation.

[0034] The estimation results of the state of charge and the health state of the battery for all time segments are serially integrated to generate the state of charge evaluation output sequence and the health state evaluation output sequence of the current target battery. The estimation results of the state of charge of the battery obtained by calculation for each time segment are arranged in the original time order to form the state of charge evaluation output sequence; similarly, the estimation results of the health state of the battery obtained by calculation for each time segment are arranged in the original time order to form the health state evaluation output sequence. Serial integration can clearly show the real-time change trend and long-term operation trend of the battery state during the navigation process, facilitating effective decision-making based on real-time and accurate data in the process of optimizing the battery charge and discharge strategy.

[0035] S5: According to the disturbance perception prediction sequence, and the evaluation results of the state of charge and health state of the battery, construct an energy scheduling constraint graph and generate a multi-dimensional strategy candidate set, including: Define energy scheduling constraint nodes based on the state of charge evaluation output sequence and the health state evaluation output sequence of the battery, in combination with the disturbance perception prediction sequence. Each specific time point is regarded as an energy scheduling constraint node, and each energy scheduling constraint node includes three types of constraint conditions: the state of charge constraint condition of the battery, the health state constraint condition of the battery, and the disturbance perception prediction constraint condition. The definition method of the state of charge constraint condition of the battery is: determine the allowable charge and discharge current range of the battery according to the value of the state of charge evaluation output sequence at the moment where the energy scheduling constraint node is located; the definition method of the health state constraint condition of the battery is: determine the maximum allowable charge and discharge power limit of the battery according to the value of the health state evaluation output sequence corresponding to the moment of the energy scheduling constraint node; the definition method of the disturbance perception prediction constraint condition is: determine the power response condition that must be satisfied according to the predicted load change trend and amplitude at the corresponding moment in the disturbance perception prediction sequence. The three types of constraint conditions comprehensively reflect the real-time state, long-term health, and load change requirements of the battery, ensuring the comprehensiveness and applicability of the definition of energy scheduling nodes.

[0036] In the order of time series, each energy scheduling constraint node is connected in sequence through directed edges to form an energy scheduling constraint graph; Each node is only connected to the node at the next moment in a one-way manner, forming a graph structure arranged in the order of time. Each directed edge represents the transfer path of energy from the current node to the next node. The constraints represented by each node in the graph determine the transfer method and transfer range of the node to the next node.

[0037] For the energy scheduling constraint graph, by traversing the node paths, the control paths that meet the propulsion power requirements and the critical load power requirements are listed, and the charging and discharging schemes corresponding to each control path are recorded as strategy candidates; All node paths in the energy scheduling constraint graph are traversed by depth-first search or breadth-first search methods. The feasibility judgment criteria for each path are defined as follows: among all nodes on the path, the battery's allowable charging and discharging power range must meet the corresponding moment's propulsion power requirements and critical load power requirements, and must not exceed the battery's allowable power limit or violate the state of charge safety boundary. For example, if there is a node in a certain path that predicts a sharp increase in load power, if the current battery health status does not allow this power level to be provided, the path is abandoned and the next path is traversed. In the end, all the charging and discharging control schemes corresponding to the paths that meet the above judgment criteria are retained and become valid strategy candidates.

[0038] Aggregate all strategy candidates to form a multi-dimensional strategy candidate set; The control information such as the charge and discharge power level, charge and discharge timing, charge and discharge duration, and corresponding energy transfer direction in each effective control path are recorded one by one in a unified data structure. For example, a specific strategy candidate includes: the current intensity of charging at a certain moment, the power level and specific duration of discharge at the next moment, and how the battery adjusts the power output in the next few moments to smoothly meet the load demand. The multi-dimensional strategy candidate set provides a complete set of solutions that can comprehensively reflect multiple possible scenarios, so as to accurately select the optimal strategy.

[0039] S6: Select the optimal strategy path from the multi-dimensional strategy candidate set and output the target charge and discharge control instructions, including: For each control path in the multi-dimensional strategy candidate set, the feasibility of the path is judged according to the propulsion power demand and key load power demand in the current navigation cycle, and the control paths with power default nodes or thermal safety boundary violations in the energy scheduling constraint map are excluded; For each control path, according to the ship propulsion power demand and critical load power demand within the current navigation cycle, a node-by-node review and verification are carried out. The criteria for judging the path feasibility include: for each node in the path, check whether the charging and discharging power provided by the battery at this node exceeds the pre-defined allowable power limit of the node. If the situation of exceeding the allowable power limit occurs at any node, it is defined as a power default node; according to the battery temperature status parameter corresponding to the node, check whether the battery may break through the pre-determined battery thermal safety boundary condition under the action of the charging and discharging power at the node. If the battery temperature may exceed the upper limit temperature value allowed by safety or be lower than the allowed lower limit temperature value at any node, it is defined as a thermal safety boundary violation node. For the path with a power default node or a thermal safety boundary violation node, it is excluded.

[0040] For the remaining feasible control paths, path scoring is performed based on a preset comprehensive evaluation index; The comprehensive evaluation index includes the battery life loss degree index, the battery energy utilization efficiency index, the load response sensitivity index, and the charge and discharge control stability index. The specific calculation methods are as follows: The battery life loss degree index is determined by estimating the cycle life loss caused by power fluctuations during the execution of the path by the battery. The path with less loss has a higher score; the battery energy utilization efficiency index is clearly determined by calculating the effective energy utilization rate during the execution of the path. The path with less energy loss has a higher score; the load response sensitivity index is clearly determined by the timeliness and accuracy of the path's response to disturbance prediction. The path with a more sensitive and accurate response has a higher score; the charge and discharge control stability index is specifically determined according to the smoothness of the change in the charge and discharge power between adjacent nodes on the path. The path with a more stable change has a higher score. The above four evaluation indexes are respectively assigned corresponding scores according to the specific performance of the path. By weighting and summing the four scores according to the weights, the path score value is obtained. The setting of the weights is determined by comprehensively analyzing historical data and expert experience to ensure the comprehensiveness and accuracy of the scoring results.

[0041] The method for determining the battery life loss degree index is as follows: The battery life loss degree index is determined by analyzing the cycle life loss caused by the charging and discharging power fluctuations during the implementation of a certain control path by the target battery: Based on the battery cycle life characteristic curve provided by the battery manufacturer or established through long-term historical test data, the calculation benchmark for battery life loss is determined. The battery cycle life characteristic curve describes the clear mathematical relationship between the battery capacity attenuation degree, the charge and discharge depth, and the charge and discharge current intensity.

[0042] According to the change of charge and discharge power at each node at each moment in each specific control path, the corresponding charge and discharge depth and charge and discharge current intensity are determined respectively, and the specific capacity loss of the battery life caused by a single charge and discharge action at each node is calculated based on the battery cycle life characteristic curve.

[0043] The capacity depreciation caused by all nodes on the path is accumulated and summed up in sequence to obtain the total battery life loss degree under the control path.

[0044] In order to unify the scoring standards, the life loss degree of all paths is processed according to a unified normalization method. That is, the total battery life loss of each path is compared with the maximum value of the total loss of all paths, and the life loss degree score of the path is calculated based on this. The smaller the life loss degree of the path, the higher its score.

[0045] The battery energy utilization efficiency index is the ratio between the effective energy actually output by the battery and the total energy released by the battery during the implementation of the control path: Determine the total amount of energy delivered by the battery during all discharge phases during the path implementation.

[0046] Based on the effective power demand of the actual load (propulsion system and key equipment) during the implementation of the route, the total amount of energy actually effectively utilized by the load is calculated.

[0047] The energy utilization efficiency is obtained by dividing the total amount of energy effectively utilized by the actual load by the total amount of energy output by the battery.

[0048] The score of the energy utilization efficiency index is determined in a normalized manner, that is, the energy utilization efficiency values ​​calculated for all paths are compared. The higher the efficiency, the higher the score, and the lower the efficiency, the lower the score.

[0049] The load response sensitivity index indicates the matching accuracy and response timeliness between the control action of the path and the load change demand in the disturbance perception prediction sequence: The difference between the power demand given by the disturbance-aware prediction sequence and the actual power provided by the control path is explicitly calculated node by node, and the smaller the difference, the more accurate the response.

[0050] Determine the timeliness of each node's response, that is, the delay between the power adjustment action provided in the path and the disturbance prediction change moment. The shorter the delay, the more timely the response.

[0051] The power demand difference and delay time of each node are comprehensively weighted and calculated with appropriate weights to obtain the response sensitivity score of each node. The weight is determined by historical statistical analysis, that is, the more drastic the load changes in history, the higher the weight of response timeliness, and the more stable the load changes, the more significant the weight of accuracy.

[0052] Calculate the arithmetic mean of the response sensitivity scores of all nodes on the path to obtain the overall load response sensitivity index score of the path. The higher the score, the more sensitive and accurate the path's response to load changes.

[0053] The charge and discharge control stability index represents the smoothness of the charge and discharge power changes between adjacent nodes on the path: For each pair of adjacent nodes on the path, explicitly calculate the absolute value of the difference between the charge and discharge power values. The smaller the value, the smoother the power change between adjacent nodes.

[0054] Accumulate the absolute values of the power change differences of all adjacent nodes on the path to obtain the total power change amplitude.

[0055] Process the total power change amplitudes of all paths in a normalized manner, that is, compare the total power change amplitude of each path with the largest value among all paths. The smaller the change amplitude of the path, the higher the normalized score.

[0056] Use the normalized score as the final score of the charge and discharge control stability index of the path.

[0057] Select the control path with the highest path score and determine the charge and discharge strategy corresponding to the control path as the optimal strategy path; Sort the scoring results of all feasible paths in descending order; select the path with the highest score as the optimal strategy path. The reason for choosing the path with the highest score is that this path achieves the best balance in comprehensively considering aspects such as battery life protection, energy efficient utilization, load response accuracy, and charge and discharge stability, and can meet the needs of the ship during actual operation and the battery state optimization goal to the greatest extent.

[0058] Generate the target charge and discharge control instructions according to the optimal strategy path; For each specific time node on the optimal strategy path, set parameters such as the corresponding battery charge or discharge power value, the specific start and stop time points, the charge and discharge duration, and the auxiliary control strategy (such as starting the cooling system, gradually increasing the charging current, etc.). The target charge and discharge control instructions can accurately and comprehensively guide the real-time charge and discharge actions during the actual operation of the ship's battery.

[0059] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters and threshold selections in the formulas are set by those skilled in the art according to the actual situation.

[0060] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0061] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0062] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0063] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there can be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or modules can be in an electrical, mechanical, or other form.

[0064] The module described as a separation component may or may not be physically separated. The component shown as a module may or may not be a physical module. It may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0065] In addition, in each embodiment of this application, each functional module can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0066] If the described function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0067] The above description is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

[0068] Finally: The above description is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An optimization method for the intelligent charging and discharging strategy of ship batteries, characterized in that, It includes the following steps: S1: Collect the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship, and construct a high-frequency time-series operating condition input sequence; S2: Based on the high-frequency time-series operating condition input sequence, use a multi-scale dynamic modeling algorithm to establish a load nonlinear evolution model, and output a disturbance perception prediction sequence; S3: Obtain the historical charge and discharge state records of the target battery, and combine the current, voltage, and temperature signals of the target battery during the current voyage to construct a heterogeneous state feature matrix; S4: Based on the heterogeneous state feature matrix, evaluate the current state of charge and health state of the target battery; S5: According to the disturbance perception prediction sequence, the evaluation results of the state of charge and health state of the battery, construct an energy scheduling constraint map, and generate a multi-dimensional strategy candidate set; S6: Select the optimal strategy path from the multi-dimensional strategy candidate set, and output the target charge and discharge control instruction.

2. The intelligent charge and discharge strategy optimization method for ship batteries according to claim 1, characterized in that S1 is specifically as follows: Synchronously collect the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data generated during the operation of the ship at the same sampling frequency; Perform data alignment processing on the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data; Integrate the real-time propulsion power data, auxiliary equipment power data, and environmental disturbance parameter data after data alignment processing in sequence according to the same time stamp to construct a high-frequency time-series operating condition input sequence.

3. The intelligent charging and discharging strategy optimization method for ship batteries according to claim 2, characterized in that, S2 is specifically as follows: Perform time-domain decomposition on the high-frequency time-series operating condition input sequence according to three preset time windows of short time scale, medium time scale, and long time scale to obtain three sub-sequences; For each sub-sequence, establish a corresponding nonlinear dynamic sub-model, fit the load evolution curve through regression analysis method, and output the load change prediction increment at each time scale; Perform weighted synthesis on the prediction increments at the three time scales according to the predetermined fusion weights to obtain a disturbance perception prediction sequence.

4. The intelligent charge and discharge strategy optimization method for ship batteries according to claim 3, characterized in that S3 is specifically as follows: Obtain the historical charge and discharge state records of the target battery in multiple historical voyage cycles; Collect the real-time current signal, voltage signal, and temperature signal of the target battery during the current voyage; Perform joint mapping processing on the historical charge and discharge state records, real-time current signal, voltage signal, and temperature signal according to the time consistency standard, and extract the operation feature vectors under multiple time segments; Combine all the operation feature vectors to form a heterogeneous state feature matrix.

5. An optimization method for intelligent charging and discharging strategies of ship batteries according to claim 4, characterized in that, S4 is specifically as follows: For the operation feature vectors under each time segment, extract the current feature component, voltage feature component, and temperature feature component within the current voyage cycle; Perform joint processing on the current feature component, voltage feature component, and temperature feature component according to the preset feature calculation rules, and generate the state of charge estimation result and battery health state estimation result corresponding to each time segment through a multi-parameter combination calculation method; Perform sequence integration on the state of charge estimation results and battery health state estimation results under all time segments to generate the state of charge evaluation output sequence and battery health state evaluation output sequence of the current target battery.

6. The intelligent charge and discharge strategy optimization method for ship batteries according to claim 5, characterized in that S5 is specifically as follows: Based on the battery state of charge evaluation output sequence and the battery health status evaluation output sequence, combined with the disturbance perception prediction sequence, define the energy scheduling constraint node; In the order of time series, each energy scheduling constraint node is connected in sequence through directed edges to form an energy scheduling constraint graph; For the energy scheduling constraint graph, by traversing the node paths, the control paths that meet the propulsion power requirements and the critical load power requirements are listed, and the charging and discharging schemes corresponding to each control path are recorded as strategy candidates; All strategy candidates are aggregated to form a multi-dimensional strategy candidate set.

7. The method for optimizing the intelligent charging and discharging strategy of a ship battery according to claim 6, wherein, S6, specifically: For each control path in the multi-dimensional strategy candidate set, the feasibility of the path is judged according to the propulsion power demand and key load power demand in the current navigation cycle, and the control paths with power default nodes or thermal safety boundary violations in the energy scheduling constraint map are excluded; For the remaining feasible control paths, path scoring is performed based on the preset comprehensive evaluation indicators; Select the control path with the highest path score, and determine the charging and discharging strategy corresponding to the control path as the optimal strategy path; Generate target charge and discharge control instructions based on the optimal strategy path.

Images