207 results about "Discharge data" patented technology

The invention discloses an energy storage battery health feature extraction and state evaluation method based on transfer learning. The method comprises the following steps: S1, constructing a source domain health feature library and pre-training a model; according to the method, dependence on complete cyclic data is broken through, high precision and robustness are still achieved under the conditions of data sparsity and working condition difference, and the method is suitable for intelligent operation and maintenance and predictive maintenance of an energy storage power station. And meanwhile, common incomplete and partial charge and discharge data fragments under actual working conditions can be directly utilized for feature extraction and state evaluation, dependence on complete charge and discharge cycles is avoided, and the application scene of the data driving method is greatly widened.

The invention discloses an air pollution intelligent early warning method and system based on machine learning, and the method comprises the steps: obtaining the multi-source data of environment, wind direction, geography, traffic, emission, population and the like, carrying out the multi-scale decomposition of the environment data, extracting the short-term fluctuation and long-term trend, and achieving the time series prediction in combination with a historical mode; constructing a regional association graph by using a graph convolutional network, identifying cross-regional diffusion features, and fusing wind direction and space information to determine a high-risk region and a potential diffusion path; carrying out weighted calculation in combination with traffic and emission data to obtain a comprehensive risk score, and completing region division and early warning level determination; and finally, generating a regional visual early warning report. According to the invention, pollution trend accurate prediction and diffusion path identification can be realized.

The invention relates to a 10kV switch cabinet intelligent automatic operation monitoring and early warning method and system and a medium, and the method specifically comprises the following steps: 1, collecting the partial discharge data information of each key part, including a sound wave signal, an instantaneous pulse current signal and a high-frequency electromagnetic wave signal; 2, carrying out data processing, and obtaining a partial discharge abnormal point data value of each key part; 3, realizing partial discharge early warning of the switch cabinet by adopting a fault early warning method based on current and sound wave abnormal point data of the switch cabinet; 4, collecting the temperature of each key part and the environment temperature; 5, carrying out data processing, and obtaining a change relation function of the temperature rise of each key part along with the current; step 6, establishing a switch cabinet dynamic early warning model to realize dynamic temperature rise early warning; the method has the advantages that the operation and maintenance cost is reduced, the equipment operation efficiency is improved, potential safety hazards are found in time, and remote monitoring and early warning are achieved.

The invention discloses a data verification method, device and equipment, a medium and a product. The method comprises the following steps: acquiring a data resource set and an equipment running state; on the basis of the pollution source pollution discharge detection data and the pollution source pollution discharge permission data, carrying out standard exceeding judgment on the discharge port concentration of the equipment to obtain a judgment result; if the judgment result indicates that the discharge port concentration of the equipment is normal, establishing a correlation matrix based on the pollution source power utilization detection data and the pollution source pollution discharge detection data, and determining a correlation coefficient corresponding to the correlation matrix; determining a response level corresponding to the correlation coefficient; and determining a data counterfeiting result according to the response level. By calculating the correlation coefficient between the pollution source electricity utilization detection data and the pollution source pollution discharge detection data, the pollution discharge data counterfeiting phenomenon of the industrial enterprise is effectively identified, and the coverage rate of illegal clues of the industrial enterprise is further improved.

The invention relates to the technical field of pollution source supervision, and discloses a pollution source off-site supervision method and system based on big data, and the method comprises the steps: building a dynamic verification rule library through coupling the operation beat data of enterprise key equipment and the total power consumption fingerprint data, and comparing the matching degree between the actual energy consumption and the expected mode in real time and generating a process credibility index. According to the invention, conversion from traditional pollution discharge data verification to production process stability verification is realized, and an unforgeable supervision trust mechanism is constructed through the physical relevance between the equipment start-stop signal and the power utilization characteristic; and meanwhile, the judgment accuracy under the extreme working condition is ensured by utilizing environmental parameter dynamic compensation and micro-differential pressure bypass verification, and finally a quantifiable accumulation enterprise environmental protection credit evaluation system is formed.

The invention relates to an environmental factor deduction carbon emission transaction method based on privacy protection, and belongs to the technical field of carbon emission data management. Comprising the steps of calculating and determining an effective result of the day based on observation data of each accounting subject, performing threshold signature and updating a global model; calculating the current-day voucher anchor hash; calculating a document deduction amount, a total voucher deduction amount and a final discharge amount by using the global model, generating a remote report and a corresponding standard hash, and outputting a three-item binding hash; constructing an aggregable zero-knowledge proof circuit, and jointly verifying a deduction equation, a membership relationship, uniqueness constraint and binding consistency in one proof; and verifying the threshold signature, the standard hash corresponding to the remote report, the three-item binding hash and the zero-knowledge proof in a single transaction to complete the transaction. According to the method, automation, credibility and traceability of the whole environmental factor deduction carbon emission transaction process based on privacy protection are realized, and end-to-end verifiable, low-cost and high-throughput carbon emission settlement is realized.

The invention discloses an improvement potential quantification method for carbon emission in a mining life cycle process, and relates to the technical field of mine carbon emission management, and the method comprises the steps: collecting real-time sensor data, equipment operation logs, process parameters and unstructured documents of each stage in the full life cycle of a gold mine, and constructing a dynamically updated multi-source data pool; based on the multi-source data pool, carbon emission entities and a carbon emission entity semantic relationship are extracted, and a carbon footprint knowledge graph is constructed; performing multi-objective optimization on a collaborative decision result, solving a non-dominated solution set, screening an optimal equilibrium solution, and generating an optimization strategy sequence and corresponding expected carbon emission; and inputting the optimization strategy sequence into a mine production scheduling process for execution, monitoring actual carbon emission data in real time, comparing the actual carbon emission data with expected carbon emission, and correcting the dynamic carbon emission reference model through a comparison result. According to the method, the carbon footprint knowledge graph is constructed, so that the semantic understanding and correlation analysis of the whole-process carbon emission of the mine from exploration to pit closing are realized.

The invention provides a power plant carbon emission data full life cycle management method and device, and relates to the technical field of data processing, and the method comprises the steps: collecting production, fuel and coal quality test system multi-source data in real time through a standard interface, and forming a dynamic data flow; predicting the carbon content of the fire coal element based on a BP neural network, and calculating the carbon emission by combining an emission factor method; carrying out back calculation on the discharge amount by using a mass balance method, analyzing and verifying the relevance between data and an accounting result through multiple regression, and triggering correction when an error exceeds a threshold value; a double-stage attention deep learning network is adopted to predict the carbon price in a rolling mode, a carbon asset dynamic analysis result is generated in combination with a quota estimation model, real-time emission data, quota surplus and shortage and the carbon price trend are displayed through a visual interface, and graded early warning is conducted on abnormal data. According to the method, full-life-cycle automatic management is realized, the data acquisition real-time performance and accounting precision are improved, the data authenticity and traceability are guaranteed, and the carbon asset management decision support capability is optimized.

The invention discloses a water pollution traceability analysis and ecological restoration decision support system. The system comprises a data acquisition layer, a data management and fusion layer, a pollution traceability analysis layer, an ecological restoration decision layer and a user interaction and visualization layer. The system collects multi-source water quality, hydrology and discharge data by deploying a distributed water quality sensing network, remote sensing and unmanned aerial vehicle monitoring equipment and a pollution discharge source online access module; through data cleaning, fusion and knowledge modeling, a pollution propagation model and a traceability inversion algorithm are constructed, and a pollution source and a contribution rate are identified; and simulating a multi-class repair measure response process in the digital twin water environment, and automatically recommending an optimal repair scheme in combination with a multi-objective optimization algorithm. The method has the advantages of accurate traceability, scientific decision, eco-friendliness, dynamic optimization and the like, and is suitable for pollution treatment and ecological restoration scenes of various water bodies such as drainage basins, lakes, urban rivers and the like.

The invention discloses a multi-parameter linkage water quality sampling method and system, and relates to the field of water quality sampling, and the method comprises the steps: obtaining hydrological data, main transportation path data and pollutant discharge data; setting water area sampling points of the target monitoring water area according to the hydrological data, the pollutant emission data and the main transportation path data; water quality sampling is conducted on the water area surface sampling points and the water area vertical sampling points; verifying the validity of the sampling point data, and determining the validity of the water area surface sampling points and the water area vertical sampling points according to the validity verification result; acquiring water quality monitoring data of the water area surface sampling points and the water area vertical sampling points; calculating a pollution source linkage index according to the water quality monitoring data, and determining a pollution emission event type; determining a time sequence phase definition for the pollution event based on the pollution emission event type and the water quality monitoring data; and determining the corresponding water quality sampling frequency and water quality sampling volume by combining with the time sequence stage definition. The accuracy of water quality monitoring can be effectively improved.

The invention discloses a mine ecological restoration effect evaluation method based on a life cycle evaluation model, and relates to the technical field of ecological restoration effect evaluation, and the method comprises the steps: defining a mine restoration effect evaluation target, an evaluation range and a function unit, and carrying out the evaluation of the mine ecological restoration effect based on the evaluation range and the function unit; collecting resource consumption data, environmental emission data and ecological state data in the restoration process, and constructing a dynamic life cycle list database; and carrying out characterization calculation on data in the dynamic life cycle list database to obtain a preliminary environmental influence index value, establishing a recognition rule of a life cycle evaluation model based on sensitivity analysis of the soil microbial community in the mine restoration area to heavy metal stress, and recognizing the ecological toxicity potential of fresh water from the preliminary environmental influence index value. According to the method, the decision support report is generated visually, and multi-dimensional, dynamic and accurate evaluation of the mine restoration effect is realized.

The invention relates to the technical field of power equipment, in particular to a high-voltage cable fault diagnosis method and system. The method comprises the following steps: collecting discharge data in a discharge process of a to-be-detected high-voltage cable, and analyzing fault trend data of the high-voltage cable in a target interval to obtain a trend prediction waveform; the method comprises the following steps: constructing a reference fault-sample waveform signal mapping library through historical discharge data of a to-be-detected high-voltage cable, selecting at least one reference fault from the reference fault-sample waveform signal mapping library to form a fault combination, and superposing sample waveform signals corresponding to the reference faults in the fault combination to obtain a sample waveform signal; obtaining a sample comprehensive signal corresponding to the fault combination; and predicting a potential fault and a possible occurrence time point of the high-voltage cable based on the sample comprehensive signal and the corresponding fault combination. According to the invention, before the fault occurs, the possible fault can be checked in time, and targeted advanced processing is carried out.

The invention discloses a carbon emission prediction method and device based on iTransform, and the method comprises the steps: constructing an improved data preprocessing network based on data features, and carrying out the preprocessing of carbon emission data through the data preprocessing network; on the basis of the data preprocessing network, in combination with multivariable and non-stationary data features of the data preprocessing network, selecting an adaptive converter architecture, integrating a wavelet convolution module for capturing multi-scale features of data, and introducing a frequency enhancement channel attention mechanism to serve as an improved iTransform prediction network; inputting the preprocessed carbon emission data into an improved iTransform prediction network, and outputting a carbon emission prediction value at a future moment; pollution prevention and control are assisted based on the carbon emission prediction value, and environmental pollution is reduced. The device comprises a processor and a memory. According to the method, the accuracy of carbon emission content prediction in the future and the precise CO2 emission prediction capability are guaranteed, early warning and regulation prompts can be provided for related industries, emission reduction of each industry is supported, pollution prevention and control are assisted, then environment improvement is promoted, and the pressure of carbon emission on an ecological system is relieved.

The invention discloses a super capacitor monomer health state estimation method based on charging and discharging data, which belongs to the technical field of measurement, and comprises the following steps: carrying out de-noising processing on circulating charging and discharging data of a tested super capacitor monomer, calculating instantaneous frequency and instantaneous energy of the charging and discharging data after de-noising; energy feature distribution of time domain and frequency domain decoupling of the de-noised charging and discharging data is obtained, an energy feature distribution decomposition result is synchronously input into the health state estimation model based on the convolutional neural network, and the real-time capacity and the maximum capacity of the measured super capacitor monomer are fed back to the health state estimation model. And obtaining a predicted value of the capacity of the tested super capacitor monomer after training, and estimating the health degree of the super capacitor monomer. According to the method, collaborative analysis is carried out by fusing the time-domain dynamic characteristics and the frequency-domain steady-state characteristics of the charging and discharging data of the supercapacitor monomer, the limitation of traditional single-dimensional data representation can be broken through, and the detection sensitivity of the degradation phenomenon of the lithium ion supercapacitor is improved.

The invention relates to a marine ship carbon emission accounting method and device based on interpretable residual learning, and a medium, and the method comprises the steps: complementing the missing of static parameters through ship IMO numbers, carrying out the unified coding and unit standardization of marine log dynamic data, combining the host power and oil consumption dual-path verification of the data validity, and carrying out the calculation of the marine ship carbon emission. After abnormity is cleaned, dynamic, static and meteorological data are fused to construct a carbon emission database; calculating the comprehensive action of wind resistance, still water resistance and wave resistance by applying a fluid mechanics law, generating a physically-driven carbon emission reference value, and synchronously extracting key characteristics such as navigational speed, wave height and ship length; with reference values and features as samples, uncovered dynamic features are learned through a multi-layer residual block chain structure, and a carbon emission accounting model is trained in combination with a cross validation and regularization mechanism; and finally realizing high-precision carbon emission prediction based on the training model. Compared with the prior art, the method has the advantages that the residual network is comprehensively used, interpretability analysis is achieved, and high-precision prediction of ship carbon emission is achieved.

The invention relates to the technical field of park carbon monitoring, and discloses a park carbon emission real-time intelligent monitoring method and system based on big data acquisition. The method comprises the following steps: continuously capturing dynamic emission data streams of various carbon emission sources in a park through a distributed sensing node group, and synchronously acquiring production activity and time sequence data and environment state time sequence data; performing instantaneous feature enhancement processing on the dynamic emission data stream, extracting a multi-dimensional carbon emission intensity sequence containing time sequence relevance, and generating a park carbon emission dynamic portrait; based on the sequence and production activity time sequence data, constructing a multi-level carbon emission quantification framework containing different carbon emission polymerization granularity levels, and calculating an emission abnormal concentration index of each level; and generating a current carbon emission state summary, mapping the current carbon emission state summary to a corresponding level, calculating distribution similarity to obtain a real-time carbon emission online deviation index, generating a graded early warning signal and a regulation and control strategy sequence according to the real-time carbon emission online deviation index, and realizing real-time accurate monitoring and effective regulation and control of the park carbon emission.

The invention discloses a carbon emission data restoration method and system, and relates to the technical field of carbon emission data restoration, and the method comprises the steps: obtaining to-be-restored carbon emission data, carrying out the data preprocessing of the to-be-restored carbon emission data, obtaining corresponding original carbon emission data, carrying out the data restoration of the original carbon emission data, obtaining corresponding initial correction data, and carrying out the restoration of the initial correction data; and performing correction processing on the initial correction data to obtain corresponding correction data, and optimizing the correction data according to a preset target function and the initial correction data to obtain corresponding target carbon emission data. The technical problems that according to a traditional carbon emission data restoration method, the influence of factors such as energy structure adjustment, policy change and extreme weather on carbon emission data is ignored, distribution has strong time-varying characteristics, the deviation between the restored carbon emission data and the actual situation is obvious, and the accuracy of carbon emission level evaluation is reduced are solved.

The invention discloses a carbon emission monitoring management method and device, and relates to the technical field of carbon emission monitoring management, and the method comprises the steps: multi-source data acquisition, adaptive calibration analysis, carbon emission deviation positioning management and early warning prompt, based on the acquired multi-source data, a corresponding theoretical carbon emission value and an actually measured emission value are generated, and the carbon emission value and the actually measured emission value are calculated. According to the method, whether a self-adaptive calibration mechanism generated by corresponding carbon emission prediction of the industrial park is triggered or not is predicted, a corresponding adjustment state transition matrix is obtained through analysis, monitoring and management of carbon emission data are achieved, intelligent and efficient carbon emission is achieved, and meanwhile, according to a multi-source data dynamic calibration mechanism, a multi-source data dynamic calibration mechanism is obtained. According to the method, the consistency and reliability of carbon emission data can be remarkably improved, system errors of traditional single-source monitoring are effectively reduced, automatic identification and positioning of error sources are efficiently achieved, the robustness of the system is improved, and intelligent and efficient carbon emission monitoring management is achieved.

The invention relates to the field of electrical equipment detection, in particular to a transformer partial discharge detection method and system based on optical signal characteristics. The method comprises the following steps: acquiring an optical signal generated by partial discharge in the transformer and converting the optical signal into a voltage pulse signal; synchronizing the signal with the power frequency voltage of the transformer, and recording the peak amplitude of each pulse signal and the corresponding power frequency voltage phase angle to form a discharge data point set; carrying out statistical distribution analysis on the phase angles of the data points in a power frequency period of 0-360 degrees, and extracting a characteristic phase interval representing a discharge phase aggregation rule; and matching the characteristic phase interval with a predefined discharge type discrimination rule, and determining the partial discharge type. The method solves a problem that a conventional optical detection method cannot accurately recognize the partial discharge type of the transformer, is high in electromagnetic interference resistance, and improves the detection precision and engineering practicality.

The invention relates to an intelligent pollution discharge management and control system for an environment-friendly enterprise based on big data of the Internet of Things, and belongs to the technical field of environmental protection and automatic control. The system comprises a data acquisition module for acquiring pollution discharge data in real time through a multi-parameter sensor network; the data transmission module is used for safely uploading the data to a cloud end by using a wireless technology; the big data processing and analysis platform is deployed at a cloud end, is internally provided with data models of pollution discharge trend prediction, water quality abnormity diagnosis, energy efficiency optimization and the like, and is used for deeply analyzing and mining mass data; and the intelligent decision-making and control module automatically generates and issues a control instruction to the field execution equipment according to the analysis result. By constructing a closed-loop control system of perception-analysis-decision-execution, real-time monitoring, intelligent early warning, accurate control, cost reduction and benefit increase of the enterprise pollution discharge process are realized, and the environmental protection management level and compliance capability of the enterprise are remarkably improved.

The application discloses a shale oil and gas casing deformation early warning method and device. The method comprises the following steps: determining the microseismic event point risk level of the real-time fracturing well section of the target fracturing operation well according to the microseismic event point data of the target fracturing operation well and the pre-determined microseismic event point risk level standard; determining the construction pressure risk level of the real-time fracturing well section according to the construction pressure data and the construction discharge data of the target fracturing operation well and the pre-determined construction pressure drop early warning standard; determining the natural fracture / fault slip probability according to the risk level of the microseismic event point, the spatial distribution characteristics of the microseismic event point and the risk level of the construction pressure of the real-time fracturing well section of the target fracturing operation well; and issuing the early warning reminder of the casing deformation according to the natural fracture / fault slip probability and the position relationship between the microseismic event point and the wellbore of the real-time fracturing well section. The casing deformation can be effectively early warned to reduce the probability of the casing deformation, and the accuracy and reliability are higher.

The application discloses a carbon metering method and device based on measured data and a medium, comprising the following steps: obtaining emission data through a pre-set sensing network, determining energy consumption data of an enterprise, and obtaining environmental data through a pre-set integrated sensor; fusing the emission data, the energy consumption data and the environmental data to integrate data of different sources into a pre-set time coordinate system, and spatially aggregating the fused data according to production units to determine a comparison model; determining direct emission according to the emission data, and determining indirect emission according to the energy consumption data; comparing the direct emission and the indirect emission through the comparison model to determine emission deviation, comparing the emission deviation with a pre-set deviation threshold; if the emission deviation is greater than the deviation threshold, determining deviation reasons according to equipment operation logs and the environmental data, determining review data points according to the deviation reasons, and performing review metering according to the review data points.

The application relates to a battery state of charge prediction method, device, equipment, medium and program product. The method comprises the following steps: acquiring discharge sequence data of a target battery, wherein the discharge sequence data comprises real-time discharge data and historical discharge data before the real-time discharge data; inputting the historical discharge data into an encoder of a target state of charge prediction model to extract discharge time sequence features of the historical discharge data; inputting the discharge time sequence features and the real-time discharge data into a decoder of the target state of charge prediction model together, and outputting a state of charge prediction result of the target battery under the guidance of the real-time discharge data. The method can improve the accuracy of the state of charge prediction of the battery.

A charging device includes: a communication unit; and a processor, configured to obtain charge-and-discharge data of a battery through the communication unit, where the charge-and-discharge data includes: the number of cycles of the battery and / or a cycling time of the battery. The processor is configured to perform the following operations: reducing a charge rate of the battery in response to a condition that the number of cycles of the battery exceeds a first threshold within a preset temperature range and / or in response to a condition that the cycling time of the battery exceeds a second threshold within a preset temperature range.

The present application relates to a kind of seawater fuel cell discharge service performance test method.The present application includes the following steps, first step to determine the variable that needs to be simulated and construct the test device that can reflect above-mentioned variable, second step is to place seawater fuel cell in test device, adjust water hole parameter to meet the requirement and make seawater fuel cell enter discharge working state, third step is to collect the electric parameter of seawater fuel cell under discharge service condition and evaluate its discharge performance.In the present application, first construct the working condition environment consistent with real application scene, then place fuel cell in above-mentioned working condition environment and make it discharge, collect the discharge information of seawater fuel cell, i.e.can collect its relevant discharge data when used in underwater submersible, to ensure the environmental adaptability of seawater fuel cell under the joint action of multiple factors, and verify the correctness of theory and technical feasibility of seawater fuel cell used in underwater equipment.

The present application relates to the field of smart city environment management, and more particularly to a carbon emission information data analysis method and system based on CIM, which overcomes the drawback of ambiguous spatial location of emission data in traditional methods by implementing precise coding matching and binding of CIM spatial units and carbon emission data. The above spatial characteristics and real-time dynamic data are input into a strength analysis model based on machine learning. Personalized intelligent evaluation is achieved. After identifying high-emission units, the method can use the established precise spatial mapping for reverse intelligent tracing, quickly locating and quantifying specific activities and sources causing high emissions, and achieving precise diagnosis. Based on the clear diagnosis conclusion, further combined with the knowledge base and optimization algorithm, customized and quantifiable emission reduction strategies are automatically generated. Finally, differentiated results are pushed to different roles.

The invention discloses an enterprise emission monitoring data flow abnormity intelligent identification method comprising the following steps: collecting multi-source emission data, executing preprocessing, generating a continuous data flow, and dividing time sequence segments; mapping data according to phase state categories, constructing sub-manifolds, and establishing a reversible relationship to form a multiphase transformation manifold; a working condition neighborhood relation is constructed, a topological structure is modeled, and folding is executed to form a cross-scale topological structure; dividing an emission threshold domain interval, generating a plurality of types of energy basins, and coupling with a working condition state to form a threshold domain discriminator; three modeling results are fused, and joint anomaly discrimination is executed on each piece of data to generate score output; and identifying abnormity based on the scoring result, and outputting an abnormity type, an abnormity time period and an associated monitoring index. According to the method, accurate identification and early warning of abnormal emission behaviors under complex emission working conditions are realized by performing multiphase mechanism modeling, working condition structure expression and dynamic threshold domain discrimination on enterprise multi-source emission monitoring data streams.

The application discloses a kind of carbon emission improvement potential method of mine exploitation life cycle process, it is related to mine carbon emission management technical field, including, collecting real-time sensor data of each stage in the whole life cycle of gold mine, equipment operation log, process parameters and unstructured document, construct dynamic updating multi-source data pool;Based on multi-source data pool, extract carbon emission entity and carbon emission entity semantic relationship, construct carbon footprint knowledge graph;Multi-objective optimization is carried out to the collaborative decision result, and the non-dominated solution set is solved and the optimal balanced solution is screened, and the optimization strategy sequence and the corresponding expected carbon emission are generated;The optimization strategy sequence is input into mine production scheduling process execution, real-time monitoring actual carbon emission data and comparing with expected carbon emission, and the dynamic carbon emission benchmark model is revised by comparison result.The application realizes semantic understanding and correlation analysis to mine carbon emission from exploration to whole process of closing pit by constructing carbon footprint knowledge graph.

The present application belongs to the technical field of data processing, and particularly relates to a method and system for dynamic monitoring of energy carbon based on multi-source data fusion, which comprises the following steps: collecting high-frequency energy data and low-frequency emission data of a target area, performing frequency-raising processing on the low-frequency emission data by using an interpolation algorithm, and performing standardization processing on the two types of data to obtain energy sequences and emission sequences with unified time resolution; calculating the causal fluctuation confidence of the energy sequences at each time, and constructing a cumulative distance matrix in combination with the causal fluctuation confidence; searching for an optimal alignment path of the energy sequences and the emission sequences by using a dynamic time warping algorithm, and calculating a local slope index at each path point; dynamically adjusting the observation noise covariance of Kalman filtering according to the local slope index, performing smoothing correction on the aligned emission data, and outputting the energy carbon monitoring result. The present application solves the problem of distortion in calculation of carbon emission intensity caused by phase misplacement of multi-source data.

In some examples, a system may determine route information for a route to be traversed by a vehicle from a start location to a destination location, the vehicle including a battery having a state of health. The system receives discharge data corresponding to traversal of the vehicle along the route, the discharge data indicative of a rate of discharge of the battery at a plurality of locations along the route. Based on the received discharge data, the system at least one of trains or updates a first machine learning model configured for predicting the state of health of the battery. The vehicle may determine an estimated battery state of health based at least on the first machine learning model, and may receive control information while traversing the route based on the estimated battery state of health to at least partially minimize battery degradation during traversal of the route.