621 results about "Transient response" patented technology

The invention discloses an intelligent operation and maintenance system and method for a high-voltage switch cabinet based on digital twinning, relates to the technical field of intelligent power grids, and solves the problems of nonlinear effect modeling distortion, cross-spatio-temporal scale coupling deviation accumulation, time sequence real-time contradiction and insufficient extreme working condition adaptation in the prior art. Hysteresis parameters of the ferromagnetic material are dynamically calibrated through a quantum annealing optimization algorithm, and electromagnetic-thermal field strong coupling synchronous calculation is realized by combining multi-scale mesh generation and an implicit thermal field iterative algorithm; constructing an incremental transfer learning framework to fuse aging features and real-time data, and correcting boundary conditions of the model by adopting four-dimensional variational assimilation; establishing a hybrid verification platform to dynamically feed back extreme working condition parameters, and generating a credible operation and maintenance instruction in combination with a block chain; according to the method, the contact temperature rise prediction precision, the residual life evaluation reliability and the circuit breaker transient response real-time performance are remarkably improved, and active immune type intelligent operation and maintenance of the high-voltage switch cabinet under the extreme working condition are achieved.

The invention discloses a starting control system and method for a main and auxiliary power hybrid ship, and relates to the field of ship power control, and the system comprises a multi-source data collection module which is used for obtaining a multi-source state data set of a ship power system; the fusion processing module is used for performing feature alignment processing on the multi-source state data set based on a multi-modal sensing data fusion algorithm to generate a hybrid power starting parameter set; the dynamic weight distribution module is used for constructing a dynamic weight distribution matrix according to the hybrid power starting parameter set; the cooperative control module is used for generating a main and auxiliary engine combined starting instruction set based on an improved cooperative control algorithm; and the execution and monitoring module is used for executing the main and auxiliary engine combined starting instruction set in a preset time window and monitoring the transient response characteristic of the propulsion system. According to the starting control system and method for the main and auxiliary power hybrid ship, the starting stability and the energy utilization efficiency of the hybrid ship are improved.

The invention discloses a real-time computing system resource coordination and decision engine system based on a large language model. According to the system, a large language model is innovatively used as a central strategic decision engine, and a'decision-coordination-execution 'three-layer architecture is constructed. According to the system, macroscopic strategy generation and microscopic real-time control are decoupled by introducing a hierarchical decision-making mechanism (a strategic layer, a tactical layer and an execution layer), so that the core contradiction between LLM high reasoning delay and the microsecond / millisecond-level real-time requirement of the system is effectively solved, and the method is suitable for local computing equipment and a cloud data center. The system comprises a predictive strategy preloading system, and transient response can be achieved. Meanwhile, the system adopts an asynchronous event-driven decision-making mechanism for continuous intelligent optimization. According to the method, the top-down, semantic understanding-based and global collaborative intelligent management of the computing resources is realized, and the resource utilization efficiency, the system automation degree and the overall energy efficiency in a complex and dynamic computing environment are remarkably improved.

The invention relates to the technical field of power systems, in particular to a multi-dimensional network performance index evaluation system and method based on measured data of a new energy station. Comprising a data acquisition preprocessing unit; a multi-dimensional index construction unit; the weight adjustment unit dynamically adjusts the index weight according to the operation condition based on a fusion algorithm of an improved analytic hierarchy process and an entropy weight method; the performance evaluation model unit adopts a comprehensive evaluation model fused with a random forest algorithm to generate multi-dimensional performance scores and risk early warning; and the visual interaction terminal unit is used for displaying the evaluation result in real time and generating a trend analysis report. According to the invention, the three-dimensional evaluation system covering the static electrical parameters, the dynamic stability indexes and the harmonic pollution indexes is constructed, so that the limitation of single-dimensional evaluation in the prior art is broken through, and the extension of the network-related performance of the new energy station from local detection to full-dimensional coverage is realized; and the multi-scene evaluation requirements of the power grid on the steady-state operation, the transient response and the electric energy quality of the station are accurately matched.

The invention provides a photovoltaic power station modeling simulation analysis system, and relates to the field of power station modeling simulation. Comprising the steps that a parameter acquisition module collects key data such as voltage, current, temperature and illumination intensity in real time through a sensor and detects the integrity of the data in combination with historical data; the modeling module adopts a segmented modeling mode according to the collected data and the physical topology of the power station, and the transient processes of different stages are respectively ensured, so that the simulation precision is ensured; the simulation module inputs different disturbance parameters based on the model, simulates load and power grid fluctuation through a disturbance generator, analyzes transient response of the system and corrects the model in real time; the multi-scene generation unit is used for automatically generating a simulation scene in an extreme environment by using historical weather and power grid data, and testing the stability of a power station; and the evaluation module generates an evaluation report by analyzing the simulation result, identifies equipment faults, proposes specific suggestions for optimizing operation, and ensures the performance and stability of the power station under complex operation conditions.

The invention discloses an energy storage system state evolution trend prediction method based on multi-source data fusion. The method comprises the steps of terminal voltage, current and temperature time sequence data acquisition, time sequence segmentation normalization, multi-physics field coupling feature construction, trend prediction model construction and training and energy storage system state evolution trend prediction. According to the method, the distinguishing capacity of the model for charging and discharging physical characteristics is improved, meanwhile, the voltage change rate, the multi-dimensional feature vector of the differential internal resistance and the thermal-electric coupling effect and the explicit encoding electric-thermal-resistance coupling relation are constructed, the transient response and the temperature hysteresis effect can be effectively captured, and then the model can be used for analyzing the charging and discharging physical characteristics. A degradation-aware cross-cycle feature extraction and gating mechanism is adopted, short-term fluctuation and long-term trend are adaptively balanced in multi-scale prediction, the prediction conflict problem is relieved, finally, physical constraints based on the electrochemical law and the internal resistance temperature characteristic are embedded in a loss function, it is ensured that the prediction result is accurate in numerical value and conforms to the physical law, and the prediction accuracy is improved. And generation of physically impossible solutions is avoided.

The invention relates to a signal emission control method and device based on optical communication and a signal emission laser. The method comprises the following steps: acquiring output light intensity data of the vertical cavity surface emitting laser in a preset temperature range and a bias current range, and determining a bistable working characteristic curve; acquiring a vibration resonance gain signal; gaussian distribution transformation and target phase sampling are performed on the vibration resonance gain signal, and a signal nonlinear compensation model is established; applying a driving current pulse synchronous with communication data, symmetrically sampling the output signal subjected to transient response control at a rising edge and a falling edge, and constructing a signal quality evaluation vector; and inputting a support vector regression model, establishing a mapping relation between the light intensity characteristic parameter and the driving control parameter, and generating a real-time control signal of the laser. According to the invention, the self-adaptive accurate control of the vertical cavity surface emitting laser is realized, and the stable operation of the signal emitting laser under the temperature and bias current fluctuation condition is ensured.

The invention discloses a switching method and device of a primary and secondary fusion switch, a medium and a program product, and relates to the field of electric power intelligent dispatching, and the method comprises the steps: obtaining real-time electrical parameters and historical electrical parameters of an operation power supply; determining a system disturbance level based on the change rate of the real-time electrical parameters relative to the historical electrical parameters, and determining a system sampling period according to the system disturbance level; collecting a target electrical parameter of a power supply to be switched in a system sampling period, and calculating a switching evaluation index according to the real-time electrical parameter and the target electrical parameter; according to the switching evaluation index and the system disturbance level, determining a dynamic control parameter of the power supply to be switched in; when the adjustment rate of the dynamic control parameter is greater than a first preset threshold value and the switching evaluation index is smaller than a second preset threshold value, generating a switching enable signal; and controlling the primary and secondary fusion switch to execute switching operation according to the switching enable signal. According to the invention, the transient response performance in the power supply switching process can be improved.

The invention relates to the technical field of data processing, and discloses a large steel structure hoisting precision real-time monitoring method and system. The method comprises the following steps: scanning and collecting the edge of a steel structure through a distributed sensor network, and obtaining structure data to obtain an original monitoring data set; extracting and analyzing structure transient response features based on the original monitoring data set, and generating a state feature matrix; analyzing a coupling relationship between local deformation and overall torsion to obtain a dynamic response prediction matrix; performing space-time registration and data fusion on the monitoring data, and constructing a state evaluation data set; carrying out optimization calculation on the hoisting parameters to generate a hoisting control data packet; and performing intelligent reasoning analysis in combination with historical case data, and outputting a control instruction sequence. Real-time monitoring of the coupling relation between local deformation and overall torsional deformation in the hoisting process of the large steel structure is achieved, and the technical defect that in the prior art, the structural deformation development trend cannot be accurately recognized and predicted is overcome.

The invention provides a mixed precursor chemical gas identification method based on feature construction, and relates to the technical field of mixed gas identification, and the method comprises the steps: determining a division criterion of a transient response stage and a steady-state response stage through a working mode of fixed temperature modulation as an inverted step wave; and taking the average value in the steady-state response stage as the characteristic of the structure, establishing a sample library of the structure characteristic and the mixed gas concentration, and fitting the relationship between the structure characteristic and the mixed gas concentration by using a regression model with a simple structure, so that the function of identifying the mixed precursor chemical gas by using the simple model is realized. The current situation that the model for identifying the mixed precursor chemical gas is more complex and the algorithm scale is more huge is changed, and the working efficiency of mixed gas detection is improved.

The invention discloses a method and system for testing the electrical performance of a semiconductor chip, and the method comprises the steps: building a mapping relation between dielectric loss and loss frequency through obtaining the dielectric loss data, internal material characteristics and dynamic load power factors of the chip under each frequency band, recognizing a loss abnormal region, and analyzing the influence of the loss abnormal region on the performance of the chip; correcting test data by adopting a minimum mean square error algorithm, and optimizing the impedance of the power supply network and the dynamic switching characteristic of a functional module; calculating the breakdown voltage distribution of each layer in the chip, identifying a high-voltage risk region, optimizing the voltage distribution by using a gradient descent algorithm, and dynamically adjusting the working voltage and current; power factor fluctuation is analyzed through a sliding window algorithm, an abnormal area is identified, a power factor is optimized through a PID algorithm, and the working state and transient response of a chip in a dynamic load are improved. According to the method, the problems of nonlinearity of high-frequency dielectric loss, non-uniform breakdown voltage distribution and unstable power factor response under a dynamic load are solved, and comprehensive and accurate evaluation and optimization of chip performance are realized.

The invention provides a robust high-precision tracking control method based on a dynamic sliding mode surface. The method comprises the following steps: 1, constructing a driver model; 2, introducing a nonlinear dynamic function term, a high-order error term and an integral term of a displacement error into the dynamic sliding mode surface; and 3, constructing a robust high-precision tracking sliding mode controller based on the dynamic sliding mode surface and the driver model obtained in the step 2. 4, inputting the displacement error of the driver into the sliding mode controller; and the sliding mode controller outputs a voltage signal to the piezoelectric actuator. The nonlinear dynamic function term, the high-order error term and the integral term of the displacement error are introduced into the dynamic sliding mode surface, the transient response speed of the sliding mode surface is improved through the integral term, and the steady-state error of the sliding mode surface is reduced; the nonlinear dynamic function improves the response speed and robustness of the sliding mode surface; the high-order error term enables the sliding mode surface to be more sensitive to error changes.

The invention relates to the technical field of petroleum drilling engineering, and particularly discloses an analysis and prediction system based on petroleum drilling and production data, which realizes virtual reconstruction of a physical drilling system by constructing a digital mapping body, and adopts a multi-source sensor network to collect drill string vortex frequency and wellbore temperature field gradient data. Forming a time sequence feature set through time sequence alignment and fusion processing; performing differential transformation and modal decomposition on the feature set to extract transient response feature vectors and generate a risk distribution diagram; establishing a parameter-risk correlation model, and decoupling mechanical vibration and thermal stress interference by solving a physical equation to obtain a dynamic risk index; the risk indexes and the process parameters are mapped to a three-dimensional grid to construct a multi-dimensional feature space, a feature fusion grid is adopted to achieve information aggregation and reconstruction, and comprehensive risk assessment indexes are generated; a dynamic early warning threshold value is established according to the evaluation indexes, and drilling parameters are optimized in real time through closed-loop control.

The invention discloses a fast transient response Buck converter based on AOT control, which is compatible with the advantages of current mode control and AOT control, and is characterized in that an error amplifier is added in a negative feedback control loop, so that the precision of output voltage is improved, a current loop is stable, and the series resistance value of an output capacitor is not specifically limited; the change rate of the switching frequency in heavy load is reduced by using the improved self-adaptive turn-on time module, and the change of the switching frequency under different load current conditions is ensured to be small. The transient enhancement module is used for judging the output voltage, it is ensured that when the load current suddenly changes, the output voltage is fast and stable, overshoot and undershoot are reduced, and therefore the overall stability and reliability are improved. The sleep detection module Sleep is used for detecting a switching tube driving signal VCOMP so as to control the Buck converter to enter a dormant state, and therefore the purposes of reducing power consumption and improving efficiency are achieved.

The invention discloses a method and system for measuring inertia and damping parameters of a grid-forming type wind turbine generator, and relates to the technical field of grid-connected converter detection, and the method comprises the steps: connecting a test device to a to-be-tested grid-forming type wind turbine generator operating in an off-grid mode, enabling the test device to apply frequency disturbance and small-phase disturbance to the to-be-tested wind turbine generator at an initial moment, acquiring voltage and current of a unit port until the virtual synchronous generator reaches a stable working state, and respectively acquiring a steady-state value of the active power variation of the unit and a transient response curve of the active power variation in the whole response process; on the basis of the steady-state value of the active power variable quantity, the applied frequency disturbance is combined, and a damping coefficient is calculated according to a first relation; and on the basis of the transient response curve, an envelope line of the curve is calculated and obtained, then a product value of the natural oscillation angular frequency and the damping ratio is obtained, the reference angular frequency and the damping coefficient are combined, and the virtual inertia is calculated according to a second relation. According to the invention, accurate and efficient measurement of inertia and damping parameters can be realized.

The embodiment of the invention discloses a load transient enhancement circuit of a voltage mode DCDC converter, which is characterized in that a traditional 3-type compensation network is improved, a zero-point adjustable compensation network is generated by a compensation transconductance amplifier to carry out loop compensation, and then zero-point adjustment is carried out by detecting output voltage. When the output voltage exceeds the threshold due to the overshoot current, the transconductance can be increased through the change of the tail current, so that the zero point moves to the low frequency, the bandwidth can be increased, and the transient response capability is improved. According to the embodiment of the invention, a zero point adjustable compensation network is generated by adding a feedback path to compensate the loop, and the adjustable zero point is adjusted by detecting the output voltage. And when the load jumps, the bandwidth is increased to low frequency through zero point shift, so that the load response capability is enhanced, and the overshoot of the output voltage is reduced.

The invention discloses a coal mine dust concentration real-time monitoring method based on big data analysis, and relates to the technical field of coal mine dust concentration monitoring, and the method comprises the following steps: obtaining various parameter information generated in the data fusion process of a big data system in real time, including transient response information and time synchronization information, and analyzing and processing the obtained transient response information and the time synchronization information. According to the method, the transient deviation index and the phase error index are introduced, the data fusion process is intelligently evaluated in combination with the machine learning model, the error problem caused by the sensor response time difference is solved, and multi-level intelligent sensing is achieved. The system accurately judges the risk level, reduces false alarm and missing alarm, takes maintenance, adjustment or emergency intervention measures corresponding to different risks through a grading early warning mechanism, remarkably improves the coal mine monitoring safety, and reduces the accident risk.

The invention discloses a power module seamless switching control method and system, and the method comprises the steps: obtaining the current output voltage of a power module in a low-voltage power distribution network, and carrying out the voltage distortion monitoring of the power module; if the power module generates voltage distortion and the distortion degree exceeds a preset threshold value, determining that the power module is a module to be switched; obtaining a current load parameter, and determining a switched target module from the standby power modules according to the load parameter; based on a preset control parameter library, control parameters of the target module are determined, and the target module is pre-adjusted; switching the to-be-switched module into an offline state, and switching the target module into a running state; and obtaining the current output voltage of the target module, performing transient response evaluation on the target module, and correcting the control parameters of the target module according to an evaluation result. According to the invention, intelligent seamless switching of the power module when voltage distortion occurs can be realized, and the power supply reliability of the low-voltage power distribution network is improved.

The invention discloses a gas meter power supply abnormity monitoring system and method based on a signal acquisition technology, and relates to the technical field of gas meter power supplies, the system comprises a multi-dimensional sensing module, a dynamic adaptive module, a multi-dimensional fusion analysis module, an intelligent judgment module and a verification execution module; according to the invention, the multi-dimensional fusion analysis module accurately calculates key parameters such as a power supply ripple coefficient and load transient response time in a corresponding mode, the electromagnetic interference intensity is calculated through professional algorithms such as discrete Fourier transform in combination with an environment electric field and a magnetic field sequence, a threshold value is dynamically adjusted based on a working mode, and a comprehensive score is calculated through multi-parameter deviation weighted fusion. The power supply state is deeply and comprehensively evaluated; the intelligent judgment module scientifically divides anomaly levels according to comprehensive scores, quantifies and distinguishes real anomaly and environmental interference by means of an environmental interference contribution degree formula, solves the problem that traditional monitoring cannot accurately recognize power anomaly under complex working conditions, and improves anomaly judgment accuracy.

The invention discloses a heat flow dynamic tracking AI prediction type server temperature control system and device. The system comprises an AI prediction model module, a dynamic temperature control execution module and a data acquisition and feedback module. The AI prediction model module is connected with the output end of the data acquisition and feedback module through a data bus, and the input end of the dynamic temperature control execution module is connected with the output end of the AI prediction model module; the input end of the data acquisition and feedback module is connected with a temperature sensor group deployed in a server, the input data of the AI prediction model module comprises a server real-time load, an environment temperature and a historical temperature time sequence, and the output data of the AI prediction model module is a temperature prediction value in the future 5 minutes. According to the method, the time advantage is established through AI prediction, space buffering is established through PCM cold storage, a three-dimensional regulation and control mechanism which is precisely executed is achieved through pump valve cooperation, optimization is achieved in the heat dissipation stability, the energy efficiency, the transient response, the self-adaptive capacity and the operation and maintenance mode, and intelligent temperature control is provided for the high-density computing power center.

The invention discloses a nonlinear correction method, device and equipment for excitation characteristics of a mutual inductor, and particularly relates to the technical field of mutual inductors. The method comprises the following steps: acquiring original data corresponding to a mutual inductor, wherein the original data comprises an original current signal, historical magnetic flux and distributed temperature data; performing noise reduction processing on the original current signal to obtain a purified current signal, and determining a magnetic conductivity correction coefficient and a temperature gradient according to the distributed temperature data; inputting the purification current signal and the magnetic conductivity correction coefficient into a hysteresis model to obtain a real-time magnetic flux, and inputting the purification current signal, the distributed temperature data, the temperature gradient and the historical magnetic flux into a magnetic flux correction network to obtain a magnetic flux correction value; and compensating the real-time magnetic flux according to the magnetic flux correction value to obtain a corrected magnetic flux result of the mutual inductor. According to the invention, multi-physical field fusion input can be realized, transient response can be improved, harmonic distortion compensation can be carried out, and high precision of data can be maintained.

The invention relates to the field of PWM control, in particular to a high-precision PWM control method, device and equipment and a storage medium. The method comprises the following steps: collecting real-time charging monitoring parameters of a vehicle battery, carrying out dynamic load trend evolution, and constructing a time sequence load trend evolution curve; identifying a real-time PWM signal of the charging gun; carrying out nonlinear compensation adjustment based on the time sequence load trend evolution curve to obtain a nonlinear frequency compensation strategy; performing transient load fluctuation situation mining and transient load response matching adjustment according to the time sequence load trend evolution curve, and constructing a transient response frequency adjustment curve; and calculating the maximum safe charging power according to the real-time charging monitoring parameters, performing intelligent PWM waveform fine tuning on a transient response frequency adjustment curve, and constructing an optimal duty ratio fine tuning strategy. According to the invention, through dynamic and high-precision PWM control, the charging efficiency, safety and stability of the new energy vehicle are improved.

The invention provides a line loss reason analysis and positioning method and system based on artificial intelligence and a medium, and relates to the technical field of power system management.The method comprises the steps that when the bus loss rate of a target station area exceeds a preset line loss threshold value, an excitation sequence is issued to a general meter, a circuit breaker and a user electricity meter of the target station area, respectively obtaining corresponding transient response data, and further constructing a dynamic topological graph of the target transformer area; and a topological structure is optimized by combining a graph neural network and an attention mechanism, and abnormal connection points are detected. And determining downstream associated user electric meters, and calculating the power utilization abnormity score based on the load data. And injecting a steep front traveling wave signal into the electric meter with the highest score to obtain a reflection characteristic parameter. And performing iterative analysis by using a Double DQN reinforcement learning algorithm, determining a fault point, a line loss reason type and confidence, and realizing efficient and accurate line loss reason positioning and analysis. According to the invention, efficient and accurate troubleshooting of line loss reasons is realized.

The invention discloses a simulation IC test method and system based on load transient response, and aims to improve the performance evaluation precision of a simulation IC under a complex load condition. The method comprises the following steps: firstly, acquiring transient response waveforms and response delay characteristics of an analog IC (Integrated Circuit) in various load jump modes, and constructing a mapping matrix between a load current step change rate and response delay time; in the real-time test, response delay time in a future preset time period is predicted according to the step change rate of the current load current. When the steady-state time window exists, a fixed sampling time sequence is set based on the window to generate a first test waveform; and if the steady-state window does not exist, correcting the transient waveform, and generating a compensated second test waveform. And then time sequence decomposition is carried out on the test waveform, and a load transient stability evaluation matrix is constructed, so that the dynamic performance of the simulation IC is accurately evaluated, and a detailed performance test report is output.

The invention provides an SVG controller modeling analysis method based on field experiment data, and relates to the technical field of electric power grids. According to the method, the high-low penetration fault test is carried out on the SVG, the field experiment data is obtained, the calculation data of the transient model of the SVG under different model parameters are combined, the least square method is adopted to carry out parameter identification, and the optimal model parameters of the transient model of the SVG controller are obtained. The process avoids modeling of complex topology in the SVG and laboratory communication protocol debugging, is suitable for testing of SVGs of various models, and is high in field test result accuracy, simple in wiring and extremely low in risk. A field test result can truly reflect transient response characteristics of station-level current and reactive power, and a modeling path based on the field test is opened. The SVG control characteristic extraction efficiency is improved, the model precision and modeling efficiency of the complex topology SVG are improved, and therefore the analysis efficiency of the SVG analysis process is improved.

The invention relates to a test method and device for examining broadband measurement performance of a capacitor voltage transformer. The method comprises the following steps: frequency response test: acquiring sine wave voltage signals and first output voltage signals output by the capacitor voltage transformer to be tested for multiple times; processing the signals at different frequency points to obtain gain, phase characteristics and steady-state response capability at each frequency point; a square wave signal transmission test: collecting a square wave voltage signal and a second output voltage signal output by the capacitor voltage transformer to be tested; carrying out signal time-frequency domain analysis to obtain rapid voltage change transient response capability and broadband measurement accuracy; an operation / lightning impulse transmission test: collecting a transient operation / lightning impulse wave voltage signal and a third output voltage signal output by the capacitor voltage transformer to be tested; signal processing is carried out, and the response capability and high-frequency measurement accuracy of the transient high-frequency voltage signal are obtained; and the broadband measurement performance is obtained. Compared with the prior art, the method has the advantages of comprehensively examining the broadband measurement performance of the capacitor voltage transformer and the like.

The invention discloses a power distribution network fault detection method based on multi-source information fusion and related equipment, and the method comprises the steps: extracting transient high-frequency voltage information from original data collected by power distribution automation equipment, and capturing a transient response when a fault occurs; fault detection data which is not influenced by a control strategy and a conversion resistor of the converter are provided for subsequent power distribution network fault detection; according to the method, manufacturer equipment data and equipment operation data are utilized and analyzed, and equipment fault degree information can be obtained through analysis, so that the reason and possibility of fault occurrence can be understood from the equipment level; fault detection can be carried out through analysis of fusion information, fault elements and fault probability in the power distribution network can be accurately positioned, and effective fault positioning can be realized even under the condition that fault features are not obvious.

The invention discloses an equipment health state operation and maintenance method based on multi-channel perception, and relates to the technical field of equipment health monitoring, and the method comprises the steps: collecting the perception data of equipment in a steady-state operation stage and a dynamic working condition transition stage; according to the method, time domain segmentation marking is carried out on sensing data of a dynamic working condition transition section, transient response parameters and parameter change rate characteristics of all segments are extracted, a dynamic working condition characteristic library is established, early degradation signs are found by identifying the consistent trend of parameter slow change characteristics and interval shrinkage, and self-adaptive operation and maintenance of the equipment health state are achieved.

The invention discloses an unattended power distribution room intensive power distribution system and method based on machine learning, and the method comprises the steps: collecting the operation data of an unattended power distribution room, executing the preprocessing, and generating a standardized input data set; constructing a topology, working condition and equipment feature association relationship to form a three-dimensional coupling feature structure; inputting a hierarchical graph time sequence decision network, and fusing topology and time sequence characteristics to output candidate control actions; constructing a dynamic feasible region according to the joint features, and screening a safety action set meeting constraints; inputting into a topological reversible neural differential control module, predicting transient response and optimizing an execution action; and issuing a control instruction, collecting feedback, and triggering model increment update to realize closed-loop optimization. According to the invention, by introducing a hierarchical graph time sequence decision network and a topological reversible neural differential control mechanism, intelligent intensive scheduling and self-learning closed-loop optimization control of the unattended power distribution room are realized.

The invention discloses a physical constraint multi-scale deep learning battery SOC estimation method and device and a medium. The method comprises the following steps: performing multi-scale time sequence feature extraction by using multi-branch separable convolution, and modeling fast transient response and long-term trend dynamics; a cross-scale feature enhancement strategy is configured, the problems of feature redundancy and network degradation in the multi-branch fusion process are solved, and cross-scale integration is achieved in combination with a dimension self-adaptive residual path and a channel attention mechanism; a physical information constraint module is embedded, soft charging and discharging constraint, zero-current constraint and hard SOC boundary constraint are fused, dynamic consistency and physical reasonability are ensured, multi-scale feature representation and physical information constraint are combined, the model generalization ability and field consistency are remarkably enhanced, SOC estimation precision is improved, and the method is suitable for large-scale SOC estimation. And a reliable solution is provided for a battery management system.