50 results about "Time series reconstruction" patented technology

The invention discloses a lithium battery health state estimation method based on transfer learning, and the method comprises the steps: firstly carrying out the normalization and time sequence reconstruction of a battery charging voltage-capacity curve, and constructing a unified input sequence; extracting long-time-sequence degradation characteristics based on a Mama network, and completing SOH regression prediction through a two-stage full connection layer; in the cross-domain adaptation stage, in combination with an alignment strategy of dynamic time warping and weighted maximum mean difference, time sequence matching and feature distribution alignment in the degradation stage are realized; meanwhile, on the basis of a sample weighting mechanism of a Wasserstein distance, the effectiveness of migrating source domain knowledge to a target domain is improved; through a two-stage strategy of source domain pre-training and source-target joint training, a relatively low prediction error and a relatively high fitting degree can be kept under the condition that a target domain is not labeled; the method shows good generalization ability and robustness under different battery types and different working conditions, and can provide reference for health management of the electric vehicle.

The invention relates to the technical field of safety and stability control of a power distribution network, in particular to a regional power distribution network collection point voltage reactive coupling characteristic evaluation method. The method comprises the following steps: deploying a synchronous phasor measurement unit for a plurality of transformer area convergent points, collecting electrical parameters and generating original state data of the convergent points; performing quality verification, feature extraction and time sequence reconstruction on the original state data to obtain a convergence point key feature parameter set; based on the key characteristic parameter set, an inverter group initial model is constructed through inverter operation characteristic clustering and control link characteristic analysis; performing electromagnetic transient-steady state hybrid simulation on the inverter group based on the model to generate a transient-steady state hybrid simulation scheme; on the basis of the scheme, through multi-working-condition interactive simulation and interactive gain calculation, a key coupling path of a collection point is mined. According to the method, the dynamic interaction influence between the inverter group and the power grid can be accurately captured, and the resonance path and the reactive coupling path can be effectively identified.

The invention provides a multi-task learning optimized corpus annotation automation system, and aims to solve the problems of low efficiency and poor accuracy of a traditional corpus annotation method. The system comprises a shared encoder, a multi-task decoder, a dynamic adaptive module and an online incremental learning module. The shared encoder adopts a Transform architecture to carry out word segmentation, sliding window segmentation and time embedding processing on an input corpus to generate global semantic representation; the multi-task decoder completes tasks such as event boundary detection, time sequence reconstruction and causal relationship extraction in parallel; the dynamic adaptive module adjusts the loss weight in real time according to the consistency score output by each task, and enhances the adaptive ability of the model; and the online incremental learning module realizes continuous optimization of the model through expert auditing and model parameter updating. According to the system, through multi-task learning and self-adaptive optimization, the efficiency and accuracy of corpus tagging are remarkably improved, and the system is suitable for multiple fields such as news reports, social media analysis and legal document processing.

The invention discloses a workflow analysis linkage verification method and device based on log analysis, equipment and a medium, and relates to the technical field of workflow verification, and the method comprises the following steps: collecting operation log data, and carrying out time sequence reconstruction and event extraction to obtain a structured log record; performing semantic analysis on the structured log record, extracting a log semantic feature vector of each event, and constructing a semantic feature matrix; based on the semantic feature matrix, performing semantic modeling on nodes in a preset process model to obtain a process node semantic vector; matching the log semantic feature vector with the flow node semantic vector, and constructing and optimizing an initial mapping relation matrix according to a matching result; executing flow rule compliance and rationality joint verification according to an optimization result; according to the method, deep coupling of the log, the process model and the business rule is realized, so that the problems that log analysis and rule coupling are insufficient and process compliance and rationality verification are difficult to accurately carry out are solved.

The invention provides a spatial propagation source identification method and system based on meteorological extreme events, and the method comprises the steps: carrying out the time sequence reconstruction of historical meteorological observation data of a research region, and generating a meteorological data sequence set containing multi-element meteorological indexes; performing event extraction processing on the meteorological data sequence set based on a preset extreme event judgment strategy to obtain a meteorological extreme event sequence set; performing space-time correlation intensity calculation on the spatial position events, generating a space-time correlation intensity matrix describing correlation degrees among the different spatial position events, performing significance test on the space-time correlation intensity matrix, screening correlation intensity elements, and constructing a significant correlation set; and carrying out propagation path reverse tracking analysis based on the spatial propagation source identification result, determining a propagation source position of the meteorological extreme event in the spatial dimension and a corresponding propagation influence parameter, and generating a spatial propagation source identification result. According to the method, the propagation source identification efficiency and the positioning accuracy can be improved, and excessive dependence of a traditional model driving method on initial conditions is avoided.

The present application relates to geothermal resource development and evaluation technical field, especially to the balanced water and heat type geothermal resource dynamic evaluation method, including: collecting the flow, temperature and pressure data of the produced fluid and the recharging fluid, and carrying out time synchronization, outlier elimination and time series reconstruction processing on the data to form a continuous operation data set; jointly analyzing the time series change of the operation data set, extracting flow response characteristic quantity and heat exchange response characteristic quantity, and dividing the underground heat reservoir into flow and heat storage units according to the similarity of the characteristic quantity; on the scale of the flow and heat storage unit, respectively establishing a flow structure model and a heat exchange structure model, and coupling them, through a unified state variable to describe the corresponding relationship between the fluid migration behavior and the energy transfer behavior, a flow and heat storage dynamic evaluation model is constructed; based on the model, the change of the state variable of each flow and heat storage unit with time during the recharging operation process is calculated, and the dynamic evaluation data is obtained.

The disclosure discloses a multi-modal based reflectivity remote sensing time sequence reconstruction method, device, medium and equipment, the method comprises: acquiring original multispectral remote sensing image and corresponding cloud mask and original multi-modal remote sensing image, and carrying out shadow detection based on the cloud mask to obtain a cloud shadow mask; performing partial convolution operation on the cloud shadow mask and the original multispectral remote sensing image to obtain the time sequence characteristics of optical data; after preprocessing the original multi-modal remote sensing image, obtaining the improved dual-polarization radar vegetation index, then combining the two kinds of data to obtain comprehensive multi-modal time sequence data; performing trend decomposition on the comprehensive multi-modal time sequence data to extract trend characteristics; taking the cloud shadow mask, the original multispectral remote sensing image, the time sequence characteristics of optical data and the trend characteristics of the comprehensive multi-modal data as the input of the reconstruction network to obtain the multispectral reflectivity time sequence reconstruction image with multispectral correlation and time sequence dependence.

The invention discloses a mineral resource intelligent prediction method based on a liquid neural network, and relates to the field of artificial intelligence and geological exploration cross technology, and the method comprises the steps: carrying out the time series reconstruction of a space-time coordinate observation data set, generating a multi-dimensional continuous space-time field function, and calculating a space gradient field and a time derivative field of the multi-dimensional continuous space-time field function, generating a continuous spatio-temporal data stream; and fusing the geological background static feature vector and the metallogenic dynamic feature vector, inputting the fused feature vector into a full connection layer where a liquid neural network branch and a convolutional neural network branch are jointly connected, mapping to generate a mineral resource predicted value, and meanwhile, generating a mineral resource quantity distribution diagram through spatial interpolation rendering. According to the method, the continuous spatio-temporal data flow is input into the liquid neural network branch for internal state evolution, the mineralization dynamic feature vector is generated, high-resolution capture of the dynamic time sequence mode in the mineralization process is achieved, and the accuracy and time sequence resolution of prediction of the mineralization dynamic process are improved.

The present invention discloses a method and system for industrial time series semantic parsing and operation and maintenance decision-making based on a large model. The method includes: collecting time series data of the normal operating status of industrial objects to construct an abnormal time series reconstruction model, reconstructing the time series data of the faulty part and screening the target variable based on the reconstruction error and the abnormal indication function; designing a time series parsing semantic library for industrial objects, and further constructing a time series semantic parsing middle platform for semantic parsing and fault knowledge recall; constructing time series semantic parsing prompt words using the time series parsing semantic library, target variables, etc. to obtain the time series feature semantic text of the target variable; recalling the corresponding fault knowledge text in the fault knowledge library through the time series semantic parsing middle platform, constructing a time series fault diagnosis prompt word input large model, obtaining the diagnosis result through multiple voting and review, and automatically generating an operation and maintenance diagnosis form. The present invention can reduce labor costs and output data-supported and interpretable diagnostic results for industrial operation and maintenance tasks.

The invention relates to the technical field of road condition analysis, in particular to a road section closing detection method and device based on double-model cooperative discrimination, and the method comprises the steps: projecting floating car track data to a road network, and obtaining the flow time sequence data of a road section through the combination of the flow information of at least one road section; inputting the traffic time sequence data into a pre-constructed time sequence reconstruction model to identify at least one deviated abnormal point, and inputting the traffic time sequence data into a pre-constructed abnormal attention model to locate at least one abnormal traffic point; and generating an abnormal evaluation score and / or an evaluation result according to the at least one abnormal point and the at least one abnormal flow point so as to obtain a cooperative discrimination result of the road section closing abnormity. Therefore, the problems that the detection precision and robustness are poor when flow data with large fluctuation and large noise are faced due to the fact that related technologies depend on heuristic rules or single-point anomaly judgment and the overall evolution trend of the closed event in a long time sequence is neglected are solved.

The invention discloses a packet capturing method for industrial equipment during wireless transmission of a control command, and particularly relates to the technical field of industrial network security. In order to solve the problems of control command repetition and artifact interference caused by network jitter, equipment failure or malicious replay attack in the prior art, an artifact interference evaluation model is constructed by starting with communication protocol identification, wireless packet capture, control command analysis and time sequence reconstruction and combining a response corresponding relation and source address consistency of a control command, so that the control command repetition and artifact interference evaluation method is provided. The interference degree is divided into a high level, a middle level and a low level, the strategies of isolation, prompt auditing and dynamic combination are adopted respectively, accurate repair of the control flow is achieved, and the technology remarkably improves the authenticity of control behavior analysis and the intelligent processing capacity of industrial network safety.

The application discloses a kind of hybrid energy storage auxiliary frequency modulation instruction prediction method and system of thermal power generating unit, belong to energy storage frequency modulation instruction prediction technical field, method: obtain original frequency modulation instruction sequence, carry out dynamic phase compensation;Phase reference value is calculated, and nonlinear correction is carried out;First correction signal is obtained;Information entropy gradient enhancement is carried out to sampling point;First correction signal is enhanced, and second correction signal is obtained;Chaotic time series reconstruction is carried out, and third correction signal is obtained;Chaotic time series reconstruction includes: constructing time delay embedding space to map to high-dimensional phase space, calculating local Lyapunov exponent, combining the time derivative of signal and chaotic adjustment factor to carry out dynamic reconstruction;Third correction signal is input to GRU neural network, and the output frequency modulation instruction prediction result is obtained.The application can predict the size of frequency modulation instruction in advance, let super-capacitor or battery act in advance, improve the yield value, and the degree of convenience and accuracy are high, with engineering applicability.

The application discloses a kind of based on multi-parameter dual-channel time series reconstruction energy storage power station operating state early warning method and system.The method of the application includes: obtaining the multi-parameter operating time series data of energy storage power station, extracting key early warning index as input feature from it, constructing original sequence channel and first-order difference sequence channel in sliding time window, constructing dual-channel time series reconstruction early warning model, and training model and calculating reconstruction anomaly score by the weighted sum of original sequence reconstruction error, difference sequence reconstruction error and time series recursive prediction error;Again, the final early warning main detection score is obtained by smoothing the reconstruction anomaly score, the early warning threshold is determined, and normal, first-level early warning, second-level early warning, third-level early warning and fourth-level early warning are output.The application realizes the stability of effective identification and hierarchical early warning of operating state anomaly by constructing original sequence channel and first-order difference sequence channel, and jointly modeling the multi-parameter operating state of energy storage power station.

The present invention relates to the technical field of safe and stable control of distribution networks, and in particular to a method for evaluating voltage-reactive coupling characteristics at a regional distribution network collection point. The method comprises the following steps: deploying synchronous phasor measurement units at multiple collection points in different areas, collecting electrical parameters and generating original state data of the collection point; performing quality verification and cleaning on the original state data to generate pre-processed state data of the collection point; performing feature extraction and time series reconstruction on the pre-processed state data of the collection point to obtain a set of key characteristic parameters of the collection point; based on the set of key characteristic parameters, constructing an initial model of the inverter group through clustering of inverter operation characteristics and analysis of control link characteristics; performing electromagnetic transient-steady-state hybrid simulation on the inverter group based on the model to generate a transient-steady-state hybrid simulation scheme. The present invention can accurately capture the dynamic interaction between the inverter group and the power grid, and effectively identify the resonant path and the reactive coupling path.

For an actual scene with limited sample quantity and serious data gap, stable and reliable space-time reconstruction is difficult to realize only depending on medium-resolution data. The invention provides a frequency domain multi-scale fusion method and device for remote sensing time sequence reconstruction, and the method is characterized in that a harmonic decomposition method comprises the steps of reference sequence construction, harmonic decomposition, high-frequency component fusion and time sequence reconstruction. High-frequency fluctuation components (amplitude and phase) are extracted from low-spatial-resolution vegetation parameter data, and then high-frequency dynamic characteristics are effectively embedded into a medium-resolution vegetation parameter time sequence by combining limited medium-resolution vegetation parameter data. By means of the method, even under the condition that medium-resolution data sampling is sparse or observation missing is serious, continuous and stable reconstruction of the vegetation space-time dynamic state can be achieved.

The invention provides a Czochralski silicon single crystal furnace lifting lead screw fault diagnosis method based on multi-source signals, and belongs to the technical field of Czochralski silicon single crystal furnace fault diagnosis. Comprising the following steps: collecting vibration information and displacement information of a lifting lead screw of the czochralski silicon single crystal furnace; performing information division and normalization processing on all the collected information in sequence to obtain a plurality of vibration normalization data sets and a plurality of displacement normalization data sets; constructing fused data information by using all the normalized data sets; performing coarse graining reconstruction, time sequence reconstruction, permutation entropy calculation and normalization operation on the fusion data information in sequence to obtain a fusion entropy corresponding to each scale factor; extracting a fusion entropy feature set from all fusion entropies, and training and verifying the support vector machine model to obtain a fault diagnosis model; and performing fault diagnosis on the lifting lead screw of the czochralski silicon single crystal furnace by using the fault diagnosis model. According to the invention, the fault diagnosis precision and efficiency can be improved.

The application discloses a large language model liver cancer prediction method and system based on an improved sampling strategy, relates to the technical field of artificial intelligence medical diagnosis, and comprises the following steps: acquiring patient electronic medical records and performing time series reconstruction to construct structured prompt words; inputting the structured prompt words into a general large language model with parameter freezing, using a random decoding strategy to generate multiple reasoning paths in parallel; constructing a target distribution model based on a gamma distribution, and using an importance sampling algorithm to calculate the normalized weight of each path to suppress low-quality paths and amplify the weight of paths conforming to medical logic; finally, the diagnostic conclusion is weighted and aggregated based on the weight to obtain a prediction result, and the reasoning process with the highest weight is output as an explainability report. The application can significantly improve the accuracy of liver cancer prediction and the transparency of clinical decision-making without model fine-tuning.

An edge flow-based hoisting equipment multi-source heterogeneous data acquisition and processing method. In view of the problems of low clock synchronization accuracy of hoisting equipment multi-source heterogeneous data, limited computing power of edge and unbalanced resource scheduling, through multi-protocol interface to access stress, vibration and operating parameters, after filtering and pretreatment, a high-precision hardware clock is set as a reference frame and a global timestamp is marked; for data with different sampling frequencies, non-uniform interpolation time series reconstruction is carried out by using cubic spline interpolation, linear interpolation and energy normalization extraction respectively, realizing strict alignment of multi-source data; the sliding window is used to dynamically intercept data stream, and the incremental principal component analysis and sequential SKL method are combined to extract low-dimensional principal component and reconstruction residual online, realizing lightweight feature extraction at the edge; a task urgency score model is constructed based on reconstruction residual, energy volatility rate and data waiting time, and when the score exceeds the threshold, the kernel priority is preempted, realizing priority transmission of key abnormal features.

The present application belongs to the technical field of meteorological data processing and remote sensing mapping, and particularly relates to a rice maturity suitability zoning method based on time series meteorological data analysis. The present application obtains an annual representative meteorological mode through annual inter-period normalization and time series reconstruction, obtains annual precipitation accumulation and temperature supply through precipitation and temperature index calculation, and combines rice growth mechanism and rice maturity law to perform threshold segmentation on large-area precipitation and temperature indexes, so as to realize rice maturity suitability zoning. The method provided by the present application effectively reconstructs the annual meteorological mode under the background of global climate change and regional climate anomaly, analyzes and extracts meteorological indexes most significantly affecting rice maturity suitability, performs threshold segmentation on key meteorological indexes in combination with rice growth mechanism and rice maturity law, can provide clear and accurate guidance and suggestions for rice maturity planning and agricultural policy making, and solves the problem of blindness in current rice maturity planning.

The present application proposes a method for diagnosing faults in the lifting lead screw of a Czochralski silicon single crystal furnace based on multi-source signals, belonging to the technical field of fault diagnosis of Czochralski silicon single crystal furnaces. It includes: collecting vibration information and displacement information of the lifting lead screw of the Czochralski silicon single crystal furnace; respectively performing information division and normalization processing on all the collected information to obtain multiple vibration normalization data sets and multiple displacement normalization data sets; constructing fused data information using all the normalization data sets; successively performing coarse-grained reconstruction, time series reconstruction, permutation entropy value calculation, and normalization operations on the fused data information to obtain the fused entropy value corresponding to each scale factor; extracting a fused entropy value feature set from all the fused entropy values, training and validating a support vector machine model to obtain a fault diagnosis model; and using the fault diagnosis model to diagnose faults in the lifting lead screw of the Czochralski silicon single crystal furnace. The present application can improve the accuracy and efficiency of fault diagnosis.

The application discloses a kind of sea surface flow field inversion and reconstruction method based on stationary satellite high frequency image, belong to marine remote sensing technical field.It includes the following steps: satellite image is preprocessed and locally enhanced;Initial inversion is carried out using dense optical flow algorithm for single time phase flow field;The flow field data cube of space-time grid uniformization is constructed;Time series reconstruction is carried out using two-way Kalman filtering, through forward recursion, backward recursion and two-way fusion, using time series information to correct abnormal vector and fill in missing data;Finally, based on the physical consistency principle of seawater approximate incompressibility, the reconstructed flow field is subjected to iterative divergence constraint, and the non-physical disturbance is weakened.The application effectively solves the problem that the traditional optical remote sensing flow measurement method is easy to produce abnormal vector and poor spatiotemporal consistency under complex sea conditions such as cloud cover and weak texture by introducing time series correlation and physical constraint, significantly improves the accuracy, stability and physical reasonableness of sea surface flow field inversion.

The application discloses a simple and robust long-time sequence reconstruction method for land satellite vegetation index, belonging to the technical field of land satellite, and comprises the following steps: acquiring a data source of a land satellite, pre-processing the data source, wherein the data source comprises Landsat NDVI data and MODIS NDVI data; performing data optimization on the Landsat NDVI data based on a 3sigma criterion and a bilinear interpolation method to obtain first optimized data; constructing a Prophet model based on the first optimized data to obtain annual, seasonal and residual components; and performing NDVI reconstruction on pixels marked as null based on the Prophet model to generate spatiotemporally continuous clear-sky Landsat NDVI data. The method has good fusion effect on the case where surface coverage (such as vegetation) changes suddenly, is simple in accurate registration, is not prone to systematic errors, is easy to accurately compare NDVI data, and is more accurate in analysis results. Meanwhile, by adjusting a prior sparsity, overfitting and underfitting phenomena are avoided.

The invention relates to the related technical field of precipitation analysis, in particular to a tropical cyclone precipitation analysis method and device based on solar terminator dynamic mapping. According to the method, a sun tracking time frame is constructed, UTC moments of any longitude and latitude points in the world are converted into STT time coordinates, and then time sequence reconstruction and daily variation statistics are carried out on tropical cyclone path points and precipitation fields around the tropical cyclone path points. The method comprises the following specific steps: acquiring a high-resolution tropical cyclone path and rainfall data; converting the world time information into sun tracking time information; determining a mask area with the tropical cyclone track point as the center; extracting an effective precipitation grid in the mask area; and analyzing the data by taking time information during sun tracking as a label to obtain tropical cyclone precipitation daily change information. According to the method, the deviation caused by different longitudes and latitudes of a traditional time frame is overcome, and the daily variation rule of the tropical cyclone precipitation system can be disclosed under a unified sun tracking reference frame.

The invention relates to the technical field of time sequence anomaly detection, and provides a dynamic confrontation cross-domain time sequence anomaly detection method and system. The method comprises the following steps: processing a time sequence from a plurality of source domains; extracting local continuous features by adopting a sequence division and vectorization method; performing segmentation and semantic coding on the domain knowledge text to obtain a semantic vector; splicing the domain knowledge vector and the time sequence feature to obtain a time sequence feature set containing semantic priori; constructing a time sequence reconstruction module based on a mixed attention mechanism, and obtaining reconstruction features maintaining causal consistency and global dependence modeling capability; standardizing a reconstruction error and adaptively distributing a sample weight; a double-path adversarial learning mechanism is introduced, a generator and a discriminator are constructed, cross-domain potential feature alignment is realized through alternate optimization, and the abnormal mode discrimination capability is enhanced; and finally, calculating an anomaly score for a newly input sequence and outputting potential anomaly.

This invention provides a method and system for predicting emergency demand based on temporal reconstruction of multi-source spatial features, comprising: acquiring multi-source data of the study area; performing spatial scale unification and fusion processing on the multi-source data to obtain processed multi-source data; using static feature temporal serialization reconstruction based on unsupervised spatial clustering to process the processed multi-source data to obtain a dynamic proxy time series vector; extracting the time dependency features of the processed multi-source data using multi-dimensional time decay and aggregation techniques to construct multi-scale sliding time window features; performing high-dimensional multimodal feature set fusion and noise reduction selection based on the processed multi-source data, dynamic proxy time series vector, and multi-scale sliding time window features to obtain a simplified target prediction feature set; using the simplified target prediction feature set to train and infer the target demand prediction model online to obtain the target demand prediction model, and obtaining the predicted value of emergency target demand from the real-time online feature tensor output.

The invention discloses a day-by-day NDVI reconstruction method combining self-supervised learning and variation, and the method comprises the steps: firstly constructing a global sample library based on a self-supervised learning strategy, effectively extracting spatial-temporal features through a global-local double-flow parallel time sequence reconstruction network, and producing seamless multi-day synthetic data, and the local branch depicts a local space-time law by adopting a convolution long-short time memory module. Therefore, a day-by-day NDVI pre-filling method with intra-year time adjacent domain priori and inter-year period similar priori as guidance is developed. For gross errors possibly introduced by factors such as surface coverage change and image quality marking in the pre-filling process, an adjacent window screening method is designed to remove noisy points. And finally, obtaining a high-quality seamless day-by-day NDVI product by adopting one-dimensional variation filtering. According to the invention, the precision and efficiency advantages of the deep learning method and the robustness advantage of the variational method are integrated, and the bottleneck problem that the conventional method is difficult to handle day-by-day NDVI large-range continuous missing is broken through.

This invention discloses an adaptive method for detecting system anomalies by fusing multidimensional sequence data. The method first collects time-related raw monitoring data during the system execution phase, digitizes the non-numerical format raw monitoring data, and then normalizes it to obtain a multidimensional time series. The multidimensional time series is then reconstructed using a time series reconstruction model, and the difference between the data before and after reconstruction is calculated to obtain a reconstructed difference sequence X. A feature space is constructed based on the data distribution of the reconstructed difference sequence X, and this process is repeated t times. Based on the feature space constructed in t experiments, the feature vector of the reconstructed difference sequence X is calculated. A fixed-length sliding window is used to divide the feature vector of the reconstructed difference sequence X into multiple subsequences. Finally, kernel mean embedding is used to calculate the similarity between different subsequences, and the time corresponding to the subsequence with a similarity less than a threshold is determined as the time of abnormal system operation. This method effectively utilizes multidimensional time series data.

The present invention relates to the technical field of data prediction, and specifically to a method for predicting the compressive strength of anode carbon blocks, including: S1. Collecting the original core dynamic characteristic data of anode carbon blocks at each time stage to construct a multi-dimensional feature data set; S2. Unifying and outputting the first data set through time series reconstruction; dividing the first data set according to the roasting and cooling process stages; S3. Performing non-uniform sampling transformation processing on the multi-dimensional data of each process stage, resampling each processed stage, and splicing them in time series to obtain a second data set; the second data set is divided into a prediction set and a training set; S4. Inputting the training set into the anode carbon block compressive strength prediction model to train the model and fit the data; S5. Inputting the prediction set into the trained anode carbon block compressive strength prediction model to predict the physical correlation and result of the anode carbon block compressive strength, and outputting the anode carbon block compressive strength value to realize the quality control of anode carbon blocks in industry.

A kind of industrial equipment running state evaluation system for multi-parameter coupling, comprising: preprocessing module, time series reconstruction module, anomaly detection and fault diagnosis module, fault unit importance mining module, fault unit division module and evaluation report generation module, the present application is established according to equipment manual, historical maintenance record and historical operation data by causal mining fault causal diagram mining coupling between parameters and evaluates the importance of each fault unit.Online state evaluation, the real-time data of equipment detection are divided and based on time series reconstruction model adaptive anomaly detection is carried out, finally, according to fault causal diagram and fault unit importance weight, equipment is scored and fault analysis is carried out, and equipment running state comprehensive diagnostic report is generated.

The application discloses a kind of structure dynamic displacement full-field reconstruction method, system and medium, it is related to structural health monitoring technical field, including steps: obtaining low-frequency displacement registration time series data and high-frequency acceleration registration time series data of target monitoring structure;Initial physical information neural network model is constructed;Composite loss function is constructed, composite loss function includes displacement data loss term, physical residual loss term and structure dynamics regularization term;Initial physical information neural network model is trained until convergence by back propagation, and network model parameter is updated;Each space-time coordinate point of the structure to be reconstructed is input into the trained physical information neural network model, and the estimated displacement value corresponding to each space-time coordinate point is output.The method of the application effectively fuses the absolute benchmark advantage of low-frequency visual displacement and the dynamic detail advantage of high-frequency acceleration, realizes structure dynamic displacement high-frequency interpolation and full-field high-frequency displacement time series reconstruction.