161 results about "Curve analysis" patented technology

The invention provides an oil seal press-fitting abnormity real-time detection method based on pressure displacement curve analysis, and relates to the technical field of oil seal assembly control. Pressure displacement data acquisition is executed and a real-time pressure displacement curve is fitted when press-fitting equipment is started; based on displacement progress threshold triggering stage switching of the time sequence displacement data, an association stage characteristic curve is called from the oil seal press-fitting reference template; comparing the characteristic curve in the association stage with the real-time pressure displacement curve to solve a real-time curve deviation vector; and according to the real-time curve deviation vector mapping physical failure abnormal mode, self-adaptive adjustment of oil seal press-fitting process parameters is triggered. The technical problems that in the prior art, a continuous press-fitting process is adopted for oil seal assembly, rubber stress concentration is caused, hydraulic locking is formed due to lubricant retention, and the press-fitting depth is out of control are solved. Physical failure prevention in the oil seal assembly process is achieved, it is guaranteed that the press-fitting depth meets the technological specification tolerance in the whole process, oil seal assembly defects are restrained, and the sealing reliability of an engine is improved.

The invention belongs to the technical field of transformer operation loss monitoring, and particularly relates to a transformer operation loss condition optimization monitoring method, which comprises the following steps of: calculating an anomaly degree based on oil temperature and vibration characteristics, generating a real-time working condition in combination with load intensity, and associating accumulated operation duration to analyze a real-time operation loss anomaly index; the method comprises the following steps: realizing dynamic correlation between loss and the actual state of equipment, analyzing and extracting fluctuation parameters through a load characteristic change curve, calculating a comprehensive load fluctuation degree, correcting an operation loss abnormal index to eliminate load fluctuation interference, finally determining a dynamic reference threshold according to a historical loss curve, judging continuous overrun through double rules, and triggering early warning. Meanwhile, guide factors such as overload and internal heating abnormity are accurately recognized, corresponding adjustment instructions are output, loss monitoring accuracy and scene adaptability are improved, operation and maintenance closed loop of early warning, guide factors and disposal is achieved, and the method is suitable for transformer full-life-cycle loss monitoring.

The invention provides an electromagnetic safety test system and method for switch integrated equipment, and relates to the technical field of power system switch equipment. The system is used for solving the problems that multi-parameter dynamic coupling analysis of a traditional test is insufficient, transient and gradual interference is difficult to capture by a fixed window, and comprehensive quantitative evaluation is lacked, and comprises a test module, an interference analysis module and a test evaluation module, calculating a total coupling interference variable; the interference analysis module draws a change curve based on the total coupling interference variable, analyzes a threshold over-area, a threshold duration and a peak value, and generates an interference risk index; and the test evaluation module calculates a total evaluation value and a threshold super-window proportion by combining the risk indexes of the windows, and divides safety levels. According to the invention, multi-parameter dynamic coupling simulation is realized, different types of interferences are accurately captured, a full-period dynamic evaluation system is constructed, and the accuracy of electromagnetic safety performance judgment of equipment is improved.

The invention relates to the technical field of oil and gas field development, in particular to a fracture parameter inversion method based on a Bayesian neural network, which comprises the following steps: establishing a bottom hole net pressure conversion model based on an actual construction curve, and drawing a bottom hole net pressure curve; calculating a bottom hole net pressure index sequence and a corresponding time sequence; establishing a shaft bottom crack extension mode judgment criterion on the basis of a classic double logarithmic curve analysis method; taking the net pressure index sequence and the time sequence obtained in the previous step as input data, combining actual physical parameter constraints, and establishing an inversion fracture parameter model based on a Bayesian neural network; and inputting the pressure index sequence to be inverted into the Bayesian neural network model to obtain a specific fracture parameter inversion result. According to the technical scheme, the confidence interval of the prediction result can be given, the prediction result and uncertainty quantification capability can be synchronously provided, and the reliability and decision support value of the inversion result are greatly improved.

The invention relates to the technical field of motion planning, and discloses a robot path fairing and synchronizing method based on an Airthoid curve, comprising the following steps: determining a tool nose fairing curve transition length according to a tool nose approximation error and a parameter synchronization strategy; determining a preliminary cutter shaft fairing curve; calculating the transition length of the cutter shaft fairing curve according to the preliminary cutter shaft fairing curve and a parameter synchronization strategy, and determining a final cutter shaft fairing curve; according to the maximum curvature of the final cutter shaft fairing curve and the maximum curvature of the cutter tip fairing curve, curvature parameter synchronization is achieved; and spline parameter synchronization is realized according to the B spline curve corresponding to the remaining straight line segment of the cutter shaft fairing curve and the cutter tip fairing curve. According to the method, the approximation error constraint of the cutter shaft and the cutter tip is realized by utilizing Airtoid curve analysis, and the calculation efficiency is high; parameter synchronization of the fairing curve and the remaining straight line segment is achieved through the remaining straight line segment B spline curve, and it is guaranteed that C3 at the connecting point of the curve segment and the straight line segment is continuous.

The invention discloses a factory equipment operation data analysis and anomaly detection system based on machine learning, and relates to the technical field of anomaly detection, comprising a first feature acquisition module for acquiring a first anomaly detection feature set; the second feature acquisition module is used for acquiring a second anomaly detection feature set; combining the first anomaly detection feature set and the second anomaly detection feature set, and constructing a plurality of groups of different to-be-screened anomaly detection feature sets; the anomaly detection and evaluation module is used for acquiring model generalization data and detection effect data of anomaly detection and constructing an anomaly detection and evaluation model based on a bp neural network; a feature set-anomaly detection effect display model is constructed, curve analysis is carried out to obtain a screened anomaly detection feature set, the screened anomaly detection feature set is applied to current anomaly detection, and the problems that when anomaly detection is carried out in different scenes, a feature extraction strategy is often difficult to flexibly adjust, key anomaly modes cannot be fully captured under specific requirements, and the detection efficiency is poor are solved. And the detection accuracy and adaptability are influenced.

The invention discloses a BIM-based municipal building energy efficiency intelligent analysis method, and particularly relates to the field of BIM, and the method comprises the steps: S1, building energy efficiency model construction, S2, energy consumption monitoring, S3, human traffic analysis, S4, space aggregation analysis, S5, energy consumption analysis, S6, anomaly extraction, and S7, energy efficiency optimization early warning. According to the method, digital mirroring of geometric features and equipment parameters of municipal buildings is realized through a BIM model, a building energy efficiency model is constructed according to a model deviation condition, an abnormal energy efficiency timestamp is extracted through correlation curve analysis, and early warning is performed when risks exist according to the abnormal energy efficiency timestamp and a correlation analysis result. The method provides guidance for taking targeted countermeasures for different levels of early warning, reduces manual intervention frequency, predicts to reduce operation and maintenance labor cost, provides cost optimization and management support of a whole life cycle for an operation and maintenance party, guarantees continuity of municipal building energy efficiency analysis to a certain extent, and achieves the goals of improving quality, being convenient and economical, and being green and environment-friendly.

The invention discloses a cooling effect stability evaluation method for an aluminum alloy wheel hub low-pressure casting cooling system, and relates to the technical field of automobile wheel hub low-pressure casting. Comprises: selecting a mold object; arranging a thermocouple, and collecting temperature data; changing the initial temperature of the temperature measuring point of the thermocouple, and further acquiring real-time temperature data; feature temperature data are extracted, and linear regression of initial temperature-feature temperature values is carried out; and carrying out temperature data fluctuation analysis, and quantitatively measuring the cooling effect stability of the cooling system according to the maximum value and minimum value indexes of the discrete point deviation fitting curve. The stability of the cooling effect of the cooling system can be visually reflected, the problem that testing cannot be achieved under the condition that the initial temperature of the mold is consistent is solved, and by analyzing the relation between the initial mold temperature and the characteristic temperature value and selecting the temperature value on the temperature curve before mold opening as the characteristic temperature value, the testing accuracy is improved. The complexity of curve analysis is avoided, and the method is worthy of industrial popularization and application.

The invention discloses a method for predicting the social distraction risk of a colorectal cancer chemotherapy patient. The method comprises the following steps: firstly, acquiring an initial sample data set of colorectal cancer chemotherapy patients, carrying out potential profile analysis and subject working characteristic curve analysis on the patients, dividing a colorectal cancer chemotherapy patient group into a group without social distraction and a group with social distraction, and defining a social distraction threshold according to group division; on the basis, performing variable screening on an initial influence factor in the initial sample data set, determining main influence factor data and a corresponding data set, and then training a constructed initial prediction model of the social distraction risk based on the main influence factor data and the corresponding data set to obtain an initial prediction model of the social distraction risk; according to the method, a fully trained social distraction risk prediction model is obtained, and the social distraction risk of the patient is accurately predicted based on the fully trained social distraction risk prediction model, so that the risk that the colorectal cancer chemotherapy patient encounters the social distraction in the treatment process can be effectively prevented in advance.

The invention discloses a construction management system for a foundation pit tube well, relates to the technical field of foundation pit construction management, and solves the technical problems that a systematic analysis method is lacked, specific reasons of progress deviation cannot be deeply analyzed, and timely and effective adjustment cannot be performed. The construction progress is comprehensively analyzed by combining a crosspiece chart or an S curve analysis method, so that the construction progress condition can be accurately judged, the progress deviation can be intuitively displayed, the accuracy and the visualization degree of progress analysis are improved, and the construction progress analysis efficiency is improved. An abnormal condition analysis module and a comprehensive analysis module respectively perform deep analysis on lead and lag conditions of an abnormal progress, and perform flexible adjustment on adjustable information in combination with a residual construction period and a resource condition; for unadjustable information, the construction sequence is optimized through reasonable arrangement of procedures, adverse effects on the project are reduced to the maximum extent, and smooth proceeding of the project is guaranteed.

The invention discloses an AI-based press-fitting curve intelligent analysis and standard curve generation method and system, and relates to the field of press-fitting curve analysis, and the method comprises the steps: constructing an AI classification model and a training set; training an AI classification model by using the training set to obtain a trained AI classification model; performing quality classification on the training set by using the trained AI classification model to obtain a quality classification result; performing standard curve fitting on the press-fitting curve samples with normal quality classification results to obtain a standard curve; and performing quality classification on the to-be-analyzed press fitting curve by using the trained AI classification model to obtain a quality classification result of the to-be-analyzed press fitting curve. The AI intelligent classification diagnosis of the press-fitting curve to be analyzed is automatically realized, manual intervention is reduced, the judgment consistency is improved, and the classification accuracy is improved.

The invention relates to the field of rail transit, and discloses an online monitoring system adapted to a SiC power device recycling scene, and the system comprises the steps: collecting key operation parameters of a SiC power device in real time, and when the collected parameters exceed a preset warning threshold value, judging whether the subsequent failure trend prediction needs to be carried out or not; constructing a device performance change curve based on the collected operation parameters, analyzing the curve type and the fluctuation trend, judging whether the device has a potential failure risk or not, and judging whether key monitoring or load regulation and control need to be performed on a specific module or not by combining the parallel operation state of the device and a historical abnormal fluctuation event; the influence degree of the potential failure trend is determined by analyzing the relevance among on-resistance sudden change, junction temperature fluctuation, welding layer cavity expansion and bonding interface aging, and according to the influence degree of the potential failure trend and a warning threshold value, an online early warning signal is output and a regulation and control instruction is executed. The method has the advantage of finding the potential failure trend in advance.

The invention relates to the technical field of earth science, in particular to a multi-plate paleolatitude evolution curve analysis method and system based on Monte Carlo simulation, and the method comprises the following steps: 1) building a selective plate division scheme of a research area; construction areas are firstly divided, then an optimal scheme is automatically matched, and finally plates are automatically and accurately positioned; 2) capturing and screening corresponding plate data; 3) simulating a paleo-latitude evolution curve; 4) the user obtains a paleo-latitude value of a target age; and 5) analyzing and visually expressing the palaeolatitude motion characteristics of the plate. According to the method, multi-plate paleolatitude evolution curve analysis can be well carried out.

The invention discloses a numerical control machine tool positioning error dynamic real-time compensation system which is specially used for aluminum shaft machining and comprises a machining parameter acquisition component, a precision curve analysis component and a cutting control component. The machining parameter obtaining component inputs a parameter a needing to be machined, the precision curve analysis component obtains cutting parameters of rough machining and finish machining according to the first machining compensation model and the second machining compensation curve, and the cutting control component conducts rough machining and finish machining according to the parameters. In the rough machining stage, the system completes 80% of machining amount in a uniform cutting mode, stress data in the cutting process are obtained through a stress sensor, and a first machining compensation model is constructed and used for dynamically adjusting follow-up cutting parameters. And in the finish machining stage, the system ensures the stability of the cutting process by using the second machining compensation curve. According to the system, through the dynamic real-time compensation technology, the machining precision and efficiency are remarkably improved, cutting force fluctuation is reduced, and the system is particularly suitable for machining of materials such as aluminum shafts.

The invention relates to the technical field of pure adhesive film quality evaluation, and discloses a pure adhesive film quality evaluation method based on big data. The method comprises the following steps: acquiring a spectrum sequence, a surface texture image and environment temperature and humidity data acquired by a multi-source sensor in a production process, aligning different sampling frequency data sources through a multi-dimensional data fusion field, and performing abnormal marking on original feature data; the abnormal marking comprises adhesive film reference quality curve analysis, single-source data abnormal identification and composite verification, and the composite verification comprises time sequence fluctuation verification and spatial distribution verification; inputting the feature data into a feature quantization module to extract depth features, binding timestamps, inputting the depth features into a multi-stage evaluation transition module, performing iterative evaluation through a hierarchical analysis strategy, and outputting a result; and the dynamic adjusting and optimizing unit receives the result, adjusts and analyzes the step length parameter in real time, and judges whether the adhesive film slitting interval and the main production line need physical isolation or not.

The invention discloses a power distribution box internal component aging diagnosis maintenance decision method and system, and relates to the technical field of component aging diagnosis maintenance, and the method comprises the steps: carrying out the random disturbance of an initial component decomposition value based on a computer random number, and obtaining a plurality of different component decomposition values; decomposing the operation data by applying a plurality of different component decomposition values based on EMD, and further performing element aging diagnosis; acquiring equipment aging identification data and anomaly detection sensitive data in an abnormal data detection process based on EMD; constructing a component decomposition value evaluation model based on a bp neural network; and establishing a component decomposition value-effect curve by using the output of the component decomposition value evaluation model and the corresponding component decomposition value, and performing curve analysis to obtain an improved component decomposition value to be applied to element aging diagnosis. The problem that setting of component decomposition values lacks a unified standard, and accuracy of aging diagnosis and maintenance decision is affected is solved.

The invention relates to the technical field of intelligent paying-off, in particular to an intelligent high-precision paying-off power construction method, which comprises the following steps of: acquiring multi-source observation data and analyzing disturbance distribution, generating an ultralimit section identifier, calculating a three-dimensional coordinate and safety margin to generate a position optimization scheme, and monitoring tension and track deviation for adjustment. According to the method, the wind speed and the vehicle wake flow in different height intervals are collected, vector analysis is carried out, disturbance distribution can be rapidly constructed, the risk area is locked, dynamic calculation is carried out in combination with the space position and the safety boundary distance, the anchoring position is adjusted in time, interference is avoided, and the method has the advantages of being high in practicability and easy to popularize. Real-time tension difference value monitoring is used for achieving automatic adjustment of the inhaul cable, real-time traction force correction is executed in cooperation with three-dimensional track continuous comparison, deviation is positioned in sag curve analysis, targeted adjustment is carried out, and a wire keeps track stability and spatial form consistency under multi-source disturbance.

The invention discloses an artificial intelligence prediction system for oil and gas saturation based on frequency dispersion attributes. The artificial intelligence prediction system comprises a data acquisition module for acquiring required data and generating a saturation curve belonging to depth domain data according to the acquired data; the time-depth calibration module obtains a synthetic seismic record according to the logging data and the post-stack seismic data, compares the synthetic seismic record with the actual seismic data to obtain a time-depth relationship, and generates a saturation curve belonging to the time domain data in combination with the saturation curve belonging to the depth domain data; a frequency dispersion attribute optimization module obtains frequency dispersion attributes of different angles and different frequencies of sound waves according to the pre-stack gather data, obtains contribution degrees of the frequency dispersion attributes to saturation prediction based on an SHAP method and saturation curve analysis of a time domain, and screens out optimal frequency dispersion attributes according to the contribution degrees; and the saturation prediction module inputs the preferable frequency dispersion attribute to a neural network model for saturation prediction. The method can solve the problems of geological hypothesis simplification, insufficient data representativeness, neglect of reservoir complexity and low calculation efficiency.

The invention discloses a frozen fresh-keeping production line detection system and method based on the Internet of Things, and relates to the technical field of the Internet of Things. Production line environment monitoring equipment is networked and connected in series through an Internet of Things platform; setting a monitoring window, and collecting monitoring data; performing overall planning processing on data acquired by each device through data cleaning, and constructing a curve model for the acquired data by using a curve simulation model; mapping each data curve model into the same coordinate system, and constructing a contrast curve analysis model; analyzing the trend difference degree of each time point of the contrast curve, and performing difference fusion analysis processing on each curve to obtain a production environment comprehensive curve model; performing abnormal correlation analysis based on the comprehensive curve model in combination with historical production environment data, and regulating and controlling the production line environment data according to the analysis data; and displaying the monitoring data of each environment device through a visual port, and regulating and controlling the real-time production line environment data based on the abnormal correlation analysis data of the comprehensive curve model.

The invention relates to the technical field of intelligent pipe networks, in particular to a regional pipe network safety assessment method and system for an intelligent pipe network, and the method comprises the steps: 1, collecting the monitoring data of pipe network monitoring points in a to-be-assessed region according to a preset collection frequency, and obtaining the geographic position data and pipeline burial depth data of the monitoring points from a pipe network database; step 2, judging whether the to-be-evaluated area comprises various pipe networks or not, if so, entering step 4, and otherwise, entering step 3; 3, constructing a safety evaluation network of the single pipe network in the area, and evaluating the safety level of the single pipe network; step 4, constructing a multi-pipe network three-dimensional fusion curved surface under the condition that the to-be-evaluated area contains various pipe networks, and performing safety evaluation on the pipe networks in the area in combination with a projection analysis algorithm; and 5, constructing a fault curve corresponding to each pipe network, analyzing the characteristics of the fault curves, and carrying out safety grade division on the corresponding pipe networks.

An information processing apparatus according to one embodiment comprises an acquisition unit for acquiring a first S-P table and a second S-P table that indicate respondents which are arranged in a prescribed order of scores and questions which are arranged in a prescribed order of the number of correct answers, a conversion processing unit for reordering the respondents in the second S-P table in accordance with the arrangement of the respondents in the first S-P table, drawing a first S curve and a first P curve of the first S-P table, and converting the first S curve and the first P curve according to the horizontal width of the second S-P table, an analysis processing unit for evaluating learning of the respondents and the questions in the second S-P tables on the basis of the converted first S curve and first P curve and the reordered second S-P table, and an output control unit for outputting the result of the evaluation.

The invention relates to the technical field of modular design, in particular to a parameter matching and debugging method and system for modular transformation of a large-scale wire barrow, and the method comprises the following steps: collecting operation data of the wire barrow, generating a real-time data set, calculating the offset of a part, analyzing a residual error, generating a comparison table, and calculating response time and inertia characteristics; the method comprises the following steps of: acquiring and synchronously storing data in real time, analyzing correction quantity and sequencing through an optimization algorithm, correcting parts and generating a response curve, analyzing errors and generating a feedback graph, calculating offset, and re-correcting if the error exceeds a threshold value. In the method, through real-time data acquisition and synchronous storage, monitoring the running state of the equipment, capturing the dynamic change of the parts, calculating the offset and the deviation, and evaluating the stability of the equipment; based on an offset value comparison table and a particle swarm optimization algorithm, components are corrected, the priority is reasonably sorted and adjusted, the correction time and cost are reduced, optimization is continued until an expected state is reached, the debugging precision and efficiency are improved, and long-term stable operation of equipment is ensured.

The invention discloses an electric reactor multi-field coupling vibration fatigue simulation method and system based on an S-N curve, and belongs to the technical field of electric power equipment fatigue simulation, and the method comprises the steps: obtaining structure parameters and material parameters of a target electric reactor, and building a parameterized physical model comprising an iron core, a winding and an insulation assembly; based on the parameterized physical model of the target reactor, simulating a time-varying mechanical force in the target reactor under the excitation of rated power frequency current; according to the time-varying mechanical force, calculating the vibration response of each part of the target reactor, and extracting the alternating stress distribution of the target reactor under the steady-state operation condition; and based on the material S-N curve of each component of the target reactor, analyzing the alternating stress distribution by adopting a linear cumulative damage theory, and calculating the failure cycle index and the fatigue life of each component under the preset cycle load. According to the method, through multi-field coupling simulation and S-N curve analysis, the fatigue life of the electric reactor is predicted, and a basis is provided for preventive maintenance.

The present invention discloses a plant equipment operation data analysis and anomaly detection system based on machine learning, which relates to the technical field of anomaly detection. It includes a first feature acquisition module for acquiring a first set of anomaly detection features; a second feature acquisition module for a second set of anomaly detection features; merging the first set of anomaly detection features and the second set of anomaly detection features to construct several different sets of anomaly detection features to be screened; an anomaly detection evaluation module for obtaining the model generalization data and detection effect data of anomaly detection, and constructing an anomaly detection evaluation model based on a BP neural network; constructing a feature set - anomaly detection effect display model, and performing curve analysis to obtain the screened anomaly detection features applied to the current anomaly detection, solving the problem that when performing anomaly detection in different scenarios, it is often difficult to flexibly adjust the feature extraction strategy, resulting in the inability to fully capture key anomaly patterns under specific requirements, affecting the accuracy and adaptability of detection.

The application belongs to the technical field of image processing, and particularly relates to a seedling root system disc identification method and system based on image processing, which comprises the following steps: collecting an original image and preprocessing, calculating an optimal segmentation threshold, extracting a root system region of interest, obtaining a root system skeleton through thinning processing, and calculating a support radius of each point of the skeleton and a discrete curvature of each skeleton point to obtain an average root system curvature by averaging; calculating a root system fractal dimension; constructing a seedling root system sample library, normalizing the average root system curvature and the fractal dimension, calculating two types of feature weight coefficients based on information entropy theory, and obtaining a disc root degree evaluation index through weighted summation; determining mild and severe disc root threshold values by combining sample library statistical characteristics and ROC curve analysis, mapping a classification judgment function to a grading signal, introducing a confidence degree checking mechanism to process boundary samples, and finally completing seedling automatic sorting and manual review triggering. The application improves the accuracy of seedling root system disc identification.

The invention relates to the technical field of intelligent security check, and discloses a video analysis and prediction system applied to intelligent scheduling management of a security check channel, and the system comprises a data processing module which obtains flight information from a flight information system of an airport, fuses real-time passenger flow data collected by video monitoring, and constructs a multi-dimensional data set; the multi-curve analysis module is used for generating a full-load curve, an actual attendance curve, a historical prediction curve and a crew curve based on the multi-dimensional data set; and the intelligent scheduling recommendation module is used for calculating the number of security check channels required at the current time point according to the real-time number of people obtained by superposing the full load curve, the actual attendance curve, the historical prediction curve and the crew curve in combination with a preset security check speed. By means of the scheme, minute-level data processing, multi-dimensional prediction analysis and intelligent scheduling can be achieved, so that the security check efficiency and the resource utilization rate are improved, and efficient and safe operation of air transportation is guaranteed.

The invention discloses a tight gas well unstable open-flow capacity evaluation method based on production dynamic data, and belongs to the technical field of gas reservoir engineering. The method comprises the following steps: firstly, collecting tight gas well production dynamic data and basic parameters and preprocessing; then, a deconvolution algorithm and a Duhamel principle are introduced, and variable-yield variable-pressure data are standardized into constant-pressure variable-yield data; then, in combination with a gas reservoir seepage differential equation considering natural gas compressibility, a'bottom hole flowing pressure-yield-time 'relation model is established; a yield-time function is reversely deduced, the bottom hole flowing pressure is made to be 0, and the unstable open-flow capacity is calculated; and finally, drawing a change curve, analyzing a dynamic rule and a decline mechanism, and providing reference for development countermeasures. The method realizes efficient utilization of production dynamic data, accurately captures an open-flow capacity decline rule, improves evaluation precision and timeliness, reduces cost, is suitable for unconventional gas reservoirs such as tight gas reservoirs, and provides support for efficient development of the gas reservoirs.

The present application belongs to the technical field of transformer operation loss monitoring, and in particular relates to a transformer operation loss condition optimization monitoring method. The present application calculates abnormality degree based on oil temperature and vibration characteristics, generates real-time working conditions in combination with load intensity, analyzes real-time operation loss abnormality index in association with cumulative operation duration, realizes dynamic association of loss and actual state of equipment, extracts fluctuation parameters through load characteristic change curve analysis, calculates comprehensive load fluctuation degree, corrects operation loss abnormality index to eliminate load fluctuation interference, finally determines dynamic reference threshold value according to historical loss curve, determines sustained over-limiting through double rules and triggers early warning, simultaneously accurately identifies causes such as load overload and internal heating abnormality, and outputs corresponding adjustment instructions, which improves loss monitoring accuracy and scene adaptability, realizes operation and maintenance closed loop of early warning, cause and disposal, and is suitable for transformer full life cycle loss monitoring.

The invention discloses an industrial edge rolling technology unknown fault early warning method and system based on unsupervised learning. The method comprises the steps that multi-dimensional time sequence track data in the operation process of an edge rolling robot is acquired; dividing the data into independent track segments; extracting a group of statistical features covering a centralized trend, a dispersion degree and a distribution pattern from each track segment to construct a high-dimensional feature vector; a sorting method based on feature importance is adopted, a key feature subset with the most discriminative power is screened out from all the features, and therefore optimal feature representation is constructed; training a deep auto-encoder model only by using sample features in a normal state, so that the deep auto-encoder model learns the accurate reconstruction capability of normal feature vectors; during state judgment, a reconstruction error between a new sample feature vector and a model reconstruction result is calculated to serve as an abnormal score; and a decision threshold is scientifically determined through receiver operation characteristic curve analysis and a Youden index maximization criterion, and the abnormal score is compared with the threshold, so that real-time and automatic judgment of the process state is realized.

The invention discloses a dynamic power control method and system for an electric vehicle charging pile, and relates to the technical field of dynamic power control of charging piles, including primary regulation and secondary regulation of the charging piles; obtaining a plurality of groups of different regulation time scale arrays; applying the plurality of groups of different regulation time scale arrays to the primary regulation and the secondary regulation, respectively, and obtaining oscillation evaluation data of parameter changes and multi-pile energy distribution uniformity data during the primary regulation and the secondary regulation; and constructing a regulation time scale evaluation model based on SVR; and constructing a three-dimensional change curve of the time scale-power regulation effect, and performing curve analysis to obtain an improved regulation time scale array, which is applied to the power control of the charging piles; and solving the problem that the difference between the time scales of the two is not fully considered, resulting in a system response hysteresis and difficulty in quickly adapting to load fluctuations or other dynamic changes.