702 results about "Benchmark data" patented technology

The invention provides an open vocabulary industrial defect detection method based on multi-modal prior prompt, which comprises the following steps: S1, collecting and sorting defect images of industrial products to be detected, and constructing a large industrial reference data set; s2, building a visual priori prompt pool, and injecting refined and fine-tuned priori into multi-scale industrial features; s3, designing a significance Gaussian distribution modeling mechanism, and capturing a complex spatial mode of position prior, so that fine-grained prior has better expression ability, and the generalization of prior is improved; s4, establishing a decoupling LoRA text encoder, and extracting an industrial semantic basis in the industrial text template through a hierarchical prompt template and a hierarchical decoupling LoRA mechanism; and S5, constructing a visual text fusion unit, and performing multi-modal fusion on text prompt embedding and visual features. According to the method, high-precision and strong-generalization defect detection is realized by fusing visual and text prior prompts, and the robustness and adaptability in a complex open environment are remarkably improved.

The invention discloses a PCB laser drilling deviation detection method and system, and the method comprises the following steps: obtaining original image data corresponding to a PCB laser drilling region, standard hole site coordinate data in a drilling design file, and real-time operation parameter data corresponding to laser drilling equipment; performing preprocessing to generate preprocessed image data, a standard coordinate data set and an equipment operation reference data set; extracting edge contour feature data of a drilling area based on the preprocessed image data, and generating an initial deviation vector of an actual drilling contour and a design contour in combination with the standard coordinate data set; constructing a drilling deviation influence factor matrix based on the equipment operation reference data set in combination with the initial deviation vector, and generating a weight coefficient of each influence factor for drilling deviation; and correcting the initial deviation vector based on the weight coefficient, judging whether deviation exists in the drill hole or not, and generating a drill hole deviation detection result. According to the invention, the drilling deviation detection precision of the PCB can be improved.

The invention discloses a machine vision-based numerical control machine tool cutter wear identification method and system, and relates to the technical field of machine vision. A numerical control machine tool cutter two-dimensional image sequence is collected, high-precision point clouds are obtained through three-dimensional reconstruction and calibration, and high-precision cutter reference three-dimensional point clouds are obtained; constructing a functional region division rule, segmenting the high-precision tool reference three-dimensional point cloud to obtain a tool functional region, and generating tool region structured reference data through geometric shape quantitative description algorithm analysis; collecting regional structured data of the cutter in different machining periods, and obtaining a shape evolution track of the cutter through space-time registration; constructing a tool wear recognition model based on the time sequence convolutional network, inputting a tool morphology evolution trajectory, and outputting a tool wear type; and collecting tool historical case data to construct a wear mechanism knowledge base, performing matching mapping on a tool wear type and the knowledge base, and generating a tool wear identification result. And based on the wear identification result, realizing the identification of the wear of the numerical control machine tool cutter.

The invention discloses a knowledge graph completion method based on combined prediction of large and small models, which comprises the following steps: 1, constructing and preprocessing a knowledge graph completion reference data set, and training by adopting a RotatE model to obtain candidate entities generated by the model and confidence scores; 2, constructing a related triad, an adjacent triad and entity long text description based on the query to form a context prompt, inputting the context prompt and the query into a large language model, and performing semantic reordering and scoring by the large language model; and 3, constructing a fine tuning data set, and performing fine tuning on the large language model to obtain the KGC task optimization-oriented large language model. And based on the KGC score, the LLM score and the dynamic weight, outputting a complementation result through joint prediction of a fusion result. According to the method, the structured reasoning ability of the small model and the deep semantic understanding of the large language are effectively combined, the prediction accuracy, robustness and specialty are remarkably improved, and the defects that a single model is weak in generalization ability and insufficient in semantic utilization are overcome.

The invention relates to the technical field of large model algorithms, in particular to a large model dynamic compression optimization method and system based on sparse pruning, and the method comprises the steps: capturing original weight fluctuation data generated by resource fluctuation in reasoning, and obtaining sparse weight reference data through sparse processing; analyzing calculation complexity through model reasoning delay data, and separating reasoning delay amount caused by a model scale; dynamically controlling the model compression ratio within a preset performance range based on the delay amount and the sparse reference data, and collecting reasoning precision distribution data under different compression parameters; evaluating a model performance state under each parameter by means of a neural network simulation method, and generating a performance state simulation result; determining a model quality optimization compensation parameter based on a simulation result by combining resource fluctuation data acquired in real time in a compression process; and finally, the compression strategy is adaptively regulated and controlled through the compensation parameters, and collaborative optimization of model calculation complexity, reasoning precision and delay during dynamic change of hardware resources is realized.

A method, a system, and a computer program product for generating a benchmarking dataset. One or more queries for generation of one or more summaries of one or more electronic documents are received. The queries are modified using one or more parameters associated with the electronic documents to generate modified queries. The electronic documents are sent to a generative artificial intelligence (AI) model. The generative AI model generates summaries of the electronic documents based on at least one of: the initial queries and the modified queries. One or more labels for the electronic documents are generated using the summaries.

The invention discloses an aerospace equipment data-real fusion performance test method and device based on reinforcement learning, and the device comprises a digital twin model building module which is used for building an aerospace equipment high-fidelity mechanism model fusing multi-physical field features, a lightweight digital twinborn body adaptive to reinforcement learning training is generated through a model order reduction technology; the digital twinborn model correction module is used for carrying out consistency calibration of multi-dimensional features and key physical parameter correction on the reduced-order model based on object reference data; the performance test module based on reinforcement learning actively explores a coupling scene of an extreme working condition and multiple faults in an interaction environment by constructing an attacker agent, and autonomously explores a performance boundary and a limit envelope of equipment; and the number-real fusion verification module is used for carrying out a hardware-in-the-loop point compensation test by taking the performance limit obtained by the digital test as guidance so as to assist in decision making and optimization of a test scheme. According to the method, the coverage, the safety and the test efficiency of the limit performance test of the aerospace equipment are improved.

The invention relates to the technical field of power system relay protection, and discloses a relay protection device mutual inductor adjustment system and method, and the system comprises a system connection initialization module, a data collection reference data generation module, a power frequency background signal offset module, a pilot frequency current injection error calibration module, and a calibration result verification output module. The method comprises the following steps: initializing sampling frequency of a to-be-calibrated mutual inductor and calibration source output precision of a calibration system, collecting amplitude and phase data of real-time power frequency current of the mutual inductor, generating offset current with equal amplitude and opposite directions through power electronic compensation based on reference data, optimizing offset current parameters through a closed-loop adjustment algorithm, and adjusting the offset current parameters through a closed-loop adjustment algorithm. Power frequency background signal offset is completed, pilot frequency calibration parameters are set, a difference value between a mutual inductor pilot frequency output signal and a calibration source standard signal is extracted by adopting a zero pointing technology, an optimal calibration coefficient is automatically calculated, verification of a calibration result is completed, and accuracy and reliability of the calibration result are ensured.

The invention provides a printing color consistency automatic calibration and management system and method, and the method comprises the steps: collecting a current environment temperature and humidity ink matching parameter density curve data dot increase compensation parameter in real time through a state detection device, and forming a current state data set; performing similarity analysis on the reference data set and the current state data set by adopting a matching degree evaluation algorithm, and judging a relationship between a matching degree numerical value and a preset threshold value; adjusting and correcting the calibration reference value and the color control parameter in the reference data set according to the offset compensation set to obtain a compensated and adjusted parameter set; on the basis of the compensated and adjusted parameter set, generating update code data by combining a current time identifier and an operator identifier, and forming new two-dimensional code information; and outputting the new two-dimensional code information to a subsequent work order carrier by adopting a code generation module, and establishing a parameter transmission tracing chain.

The invention provides an industrial air conditioner energy-saving optimization control method based on multi-source data fusion and related equipment, and is applied to the technical field of data processing. Industrial air conditioner energy-saving optimization is taken as a core, and multi-source data such as equipment parameters and environment variables and system-level reference data such as a fusion layer processing standard and an AI decision operation reference are acquired. And combining the multi-source data fusion parameters and the control strategy parameters of each level, matching data processing and instruction execution logic, and establishing a collaboration rule and generating an energy-saving constraint condition by means of hierarchical processing and multi-modal integration. And based on the data and constraint acquisition energy consumption, operation, fault early warning and other data, energy consumption prediction, supply and demand adjustment and other functional parameters are combined to generate energy-saving effect and operation state associated data. And finally, through visual analysis of an energy management platform and ISO50001 compliance auditing, an energy-saving optimization and system reliability evaluation result is output.

The invention relates to the technical field of deep counterfeit image detection, and discloses a deep counterfeit detection method and system, a storage medium and computer equipment. The method comprises the following steps: firstly, constructing a reference data set containing a forged image and an original real image; secondly, through an integrated model, generating antagonistic samples for the reference data set, and integrating the successfully attacked antagonistic samples into an antagonistic sample set; and finally, merging the reference data set and the adversarial sample set, and constructing a robustness enhanced data set containing four types of samples. In the model training stage, multi-classification cross entropy loss and comparative learning loss are combined, and expression of the model in a feature space is optimized through comparative learning constraint, so that the model learns discriminative features with more compact intra-class features and more dispersed inter-class features. The model trained by the method not only can effectively defend against attack and improve robustness, but also surpasses original detection performance on clean samples, and has remarkable technical advantages and application value.

The invention discloses a distribution box room environment parameter integrated measurement method, and particularly relates to the field of multi-parameter measurement, and the method comprises the steps: firstly constructing a machine room three-dimensional model and a sensor correlation degree matrix, then building a heat balance, humidity diffusion and airflow motion model, and calibrating parameters; deploying a measurement system and calibrating through dual synchronization and cross check; establishing a sensor confidence evaluation system and an adaptive threshold based on 72-hour reference data; fusing calibration data by adopting a three-level correlation calibration mechanism; four types of measurement modes and conversion logics are designed, and measurement resources are dynamically allocated through environmental risk assessment; multi-dimensional state parameters are extracted, a comprehensive evaluation value is calculated through normalization, dynamic weighting and combinatorial algorithms, and five-level early warning response is achieved; according to the method, multiple models and an intelligent algorithm are integrated, the parameter measurement precision and the environment risk identification efficiency are improved, the response delay is reduced, the fault diagnosis accuracy is improved, and reliable technical support is provided for safe operation and maintenance of the distribution box room.

The invention relates to a multi-dimensional intelligent cooperative detection method, system and device based on a tunnel inspection robot and a medium. The method comprises the steps that initialization is carried out based on tunnel environment parameters and task requirements, and a standardized initialization reference data set is generated; based on the data set, starting an inspection robot to carry out data acquisition, preprocessing original data, and generating a synchronized multi-source data acquisition stream; then abnormal event monitoring, structure health detection and environment perception detection are executed in parallel, and a standardized detection result set is generated; and performing multi-source data fusion processing on the standardized detection result set to obtain a core detection result and a comprehensive confidence coefficient, and performing matching in combination with a preset decision rule base to generate a standardized decision instruction set. According to the method, through multi-module collaborative data acquisition and multi-source data fusion processing, the comprehensiveness, accuracy and efficiency of tunnel inspection can be improved, and then the safety and efficiency of tunnel operation and maintenance management are effectively improved.

The invention relates to the technical field of intelligent power monitoring, in particular to an electric energy meter data acquisition method and system based on edge computing, and the method comprises the following steps: obtaining a data reading signal and task time of an electric energy meter, collecting a continuous state sequence according to a reference data set, analyzing the variation amplitude, verifying that the task completion is consistent with the time range check amplitude, and judging whether the task completion is consistent with the time range check amplitude. And updating an execution parameter based on the state record adjustment reference value, filling a control configuration activation rhythm signal write-in interface, and obtaining an electric energy meter data acquisition automation control information set. According to the method, real-time verification and self-correction of the acquisition process are realized through benchmark and revision processing of the task triggering process and state fluctuation analysis, so that the execution time and the state change are coordinated and consistent, the acquisition timeliness and stability are effectively improved, and the continuous monitoring capability of the running state is enhanced; adaptive updating of acquisition and control signals is realized, continuous and accurate information acquisition is ensured, and operation efficiency and response flexibility are improved.

The invention discloses a power distribution system planning operation collaborative optimization method integrated with multi-flexibility resources. Comprising the steps of obtaining a reference data set of a power distribution system, and fitting wind power and photovoltaic edge distribution to construct a joint probability model; and fitting the relationship between the energy storage discharge depth and the capacity attenuation, establishing a full life cycle cost model, and outputting equivalent operation cost parameters changing along with the operation strategy. Based on wind-solar combined distribution parameters, operation cost parameters and load electricity price prediction, a multi-temporal-spatial-scale scene set is generated by adopting Latin hypercube sampling, and a typical scene is obtained through clustering and scene reduction. And constructing a two-stage robust capacity planning model in combination with the typical scene and the equivalent operation cost. The model is solved through a main-sub problem decomposition coordination strategy, an installation and capacity configuration scheme is obtained, and integrated collaborative optimization of long-term planning and short-term operation of the power distribution system is achieved. According to the invention, the safety, economy and flexibility of the power distribution system in a high-volatility environment can be effectively improved.

The invention discloses a three-dimensional terrain construction method based on multi-modal semantic constraints, and relates to the technical field of three-dimensional terrain modelling, and the method comprises the steps: obtaining an optical remote sensing image and a digital earth surface model, constructing unified space reference data, and carrying out the joint semantic analysis of ground features under the guidance of the multi-modal semantic constraints; generating a semantic mask set with structural consistency; further performing regional geometric reconstruction on the digital earth surface model on the basis of a mapping relation between semantic categories and earth surface physical attributes to form a three-dimensional terrain entity conforming to ground feature space expression characteristics; meanwhile, by analyzing the project cost document and extracting the project cost data with spatial directivity, under the constraint of spatial consistency and semantic consistency, the project cost information and the three-dimensional topographic entity are subjected to association mapping, and finally a three-dimensional topographic data result fusing topographic geometric information and project economic attributes is formed.

The invention discloses an electric meter historical data restoration method based on a multi-dimensional incidence relation, and relates to the technical field of electric power data management and intelligent restoration, the accuracy and traceability of data restoration are realized by establishing a multi-source exception type library and an element time-varying degradation model, and the accuracy and traceability of data restoration are improved for multi-source data conflicts. The method comprises the following steps: pre-defining a plurality of abnormal mode combinations and corresponding repair strategies, positioning a fault source through mode matching of a real-time index and an abnormal type library, selecting reference data in combination with a historical average deviation rate of a target ammeter and an associated ammeter, avoiding deviation caused by single judgment, and for nonlinear degradation, firstly determining fault elements of a shunt and an ADC chip, a resistance degradation model and a conversion error model are constructed, a theoretical correction value is calculated, a result is optimized through a machine learning model trained by historical normal data, meanwhile, original readings, model parameters and calculation bases are stored in an associated mode, a traceable log is formed, and the repair accuracy is guaranteed.

The invention is suitable for the technical field of computer vision, and provides a polyp image segmentation method based on a basic model and feature guidance. Two innovative lightweight modules are specifically introduced into an SAM image encoder to compensate for information loss: 1) a long-distance context injection module (LCI) is introduced behind a local window self-attention layer, and global context information is efficiently injected for local features by using a shared learnable token (FeatureToken); and 2) introducing a frequency-guided attention module (FGA) behind the global self-attention layer, and guiding an attention mechanism by the module by combining high-frequency component (HFC) features of the image with a low-rank (LoRA) learnable semantic token so as to supplement key edge and texture details for global features. According to the method provided by the invention, the leading segmentation performance can be obtained on a plurality of public polyp segmentation reference data sets only by finely adjusting a small number of parameters, and the method is obviously superior to the prior art.

The invention relates to the technical field of data processing, in particular to a pipeline digital twin modeling and working condition deduction method and system and a storage medium. Comprising the following steps: acquiring temperature, pressure and corrosion data through a multi-source sensor, and performing fusion processing to obtain a calibration measurement value; comparing with historical reference data to determine systematic error characteristics, and if an error exceeds a preset threshold value, utilizing Monte Carlo simulation to generate an error distribution sample and calculating an accumulated deviation sequence to obtain an error amplification coefficient; based on this, adjusting risk assessment model parameters, constructing a digital twinborn model, generating a simulation scene, and assessing a preliminary risk level; further analyzing the trend characteristics of the risk quantitative indexes, calculating the potential failure probability in combination with the prediction model, and generating a pipeline maintenance priority sequence; and optimizing a propagation path according to the priority sequence, reasonably distributing maintenance resources, and formulating an accurate maintenance scheduling scheme. The accuracy and efficiency of pipeline maintenance are effectively improved, and the failure risk is reduced.

A road disease identification method and system based on geological radar data relates to the field of electrical digital data processing, and the method comprises the steps: transmitting electromagnetic waves to the interior of a target detection road through a geological radar transmitting antenna along a preset detection route, and receiving electromagnetic wave signals reflected by different media in the target detection road through a receiving antenna, obtaining original reflected wave data; after preprocessing the original reflected wave data, aligning ground reflected waves of each signal channel of the original reflected wave data through zero point adjustment to obtain reference data; extracting signals in an effective frequency range in the reference data, and averaging by adopting a sliding time window to obtain a wave group morphological feature, an amplitude feature and a phase spectrum feature of the target area; and matching with a preset feature threshold according to a combination relationship of the wave group morphological features, the amplitude features and the phase spectrum features to obtain a disease type of each target area in the target detection road. By implementing the method, the accuracy of road disease identification can be improved.

The invention discloses an unmanned aerial vehicle multispectral system error calibration and correction method based on a Boolean model, and belongs to the technical field of photogrammetry and remote sensing. Comprising the following steps: acquiring a multispectral image, acquiring original POS data, and performing band registration preprocessing on the image to generate a first orthoimage to confirm a system error magnitude; carrying out bundle method strict adjustment by using ground control points to obtain reference POS data, and solving seven-parameter conversion parameters; correcting the original POS data by using seven parameters and generating a high-precision orthoimage; and generating a simplified multispectral camera calibration and verification formula according to the difference value of the POS data before and after correction. Complex coordinate conversion is simplified into fixed value addition and subtraction operation, one-button calibration and batch correction of system errors of the multi-spectral camera of the unmanned aerial vehicle are achieved, the positioning precision is improved to the centimeter level from the meter level, dependence on ground control points is greatly reduced, and an efficient standardized solution is provided for large-scale engineering data processing.

The invention relates to the technical field of semiconductor detection and automatic packaging, discloses a vision detection and accurate placement system and method for MINI chips, and aims to solve the problems that in the prior art, due to the lack of multi-source sensing fusion and closed-loop control, the chip positioning precision is insufficient, the grabbing and placement reliability is low, and the whole-process quality monitoring is lack. The method comprises the following steps: acquiring three-dimensional coordinates and attitude data of a chip through a distance measurement detection unit; the vacuum pressure of the suction nozzle is monitored in real time to confirm the grabbing state; the actual posture of the chip and the burning station reference data are fused, a six-dimensional space compensation vector is calculated, a manipulator is driven to conduct fine adjustment, and accurate placement is achieved; and visual inspection is carried out at the discharging braid end, and comprehensive judgment of position, appearance and character recognition is completed. According to the scheme, submicron positioning correction, closed-loop grabbing control and whole-process quality monitoring are achieved, and the chip placement precision, the production yield and the automation efficiency are remarkably improved.

The invention discloses a multi-source information fusion-based tunneling equipment vision-assisted pose detection system and method, and relates to the technical field of exploration and engineering surveying, and the system comprises a binocular vision module, a strapdown inertial navigation module, a laser orientation instrument module and a data fusion processing module. The binocular vision module solves three-dimensional coordinates through stereoscopic vision; the strapdown inertial navigation module provides high-frequency dynamic attitude data; the laser orientation instrument modules are triangularly arranged to provide a global absolute position reference; and the data fusion processing module adopts an extended Kalman filtering dynamic fusion algorithm and combines dynamic weight adjustment to fuse multi-source data. The method comprises the steps of data acquisition, preprocessing and time-space synchronization, fusion calculation and pose information output. Through the multi-source collaborative fusion and dynamic environment adaptation technology, the detection precision and stability are improved, the automatic tracking and remote monitoring requirements of the tunneling equipment are met, and intelligent upgrading of coal mine tunneling is promoted.

The invention discloses a building reconstruction and reinforcement whole process monitoring method, device and equipment and a medium, and relates to the field of building reconstruction. According to the method, a digital twinborn-driven closed-loop monitoring system is constructed to link all parties, firstly, a dynamic digital twinborn body is constructed through multi-source data fusion, and macroscopic-microscopic two-dimensional reference data is collected; secondly, a differentiated monitoring scheme is cooperatively designed, and a physical and data-driven dual early warning model is established; the construction stage depends on a perception-transmission-platform-application four-layer system, multi-parameter real-time monitoring, four-level intelligent early warning and linkage regulation and control of the construction process are achieved, the problems of traditional monitoring fragmentation, early warning lag and information isolated island are solved, data of design, construction and monitoring parties can be effectively linked and combined, and the construction efficiency is improved. All parties are pushed to make accurate dynamic decisions under structural state changes generated by construction disturbance, full-period intelligent management and control are achieved, engineering safety and quality controllability are improved, and the method is suitable for various existing building reconstruction and reinforcement projects.

The invention discloses a Non-IID federated learning backdoor attack defense method and system and a medium, and relates to the field of federated learning and network security. In order to solve the problems that an existing method depends on model parameters and is easy to avoid by a malicious client, and adaptive Non-IID scenes are poor, truncated singular value decomposition is executed through the client to extract first p left singular vectors, and the first p left singular vectors are uploaded to a server; the server constructs a matrix based on left singular vector cosine similarity and performs hierarchical clustering, and calculates a client similarity score and an aggregation weight by using a zoom dot product attention mechanism in combination with a left singular vector of the clean reference data set; the server distributes a global model, the client uploads parameters after local training, and the server weights and aggregates the model parameters in the cluster according to the weight and iteratively optimizes the model parameters. According to the method, the malicious client is identified from the data essential features, the malicious proportion does not need to be preset, the good client contribution and the data privacy are guaranteed while the backdoor attack is inhibited, the method is suitable for a Non-IID scene, and the model robustness and the main task performance are improved.

The invention relates to the technical field of sterilization equipment control, in particular to an intelligent control method and device for sterilization equipment and electronic equipment. The method comprises the following steps: deploying a sensing network in a disinfection space and executing environment calibration to generate sensing reference data; synchronously acquiring space environment data through a sensing network; constructing a target recognition model based on the sensing reference data and the space environment data so as to perform dynamic target recognition and contact judgment; determining environment state parameters according to dynamic target recognition and contact judgment results; executing risk partitioning according to the environment state parameters to determine a region type and a risk score; and establishing a killing strategy rule base and executing killing operation according to the area type and the risk score. According to the invention, a differential disinfection strategy is formulated and executed based on dynamic environment perception and partition recognition.

The invention discloses an on-line monitoring and compensating method and system for the thickness of an electroplated layer of an automobile part based on current self-adaptive adjustment, and relates to the technical field of automation and intelligent control. Correcting the original thickness data according to the deviation degree between the current temperature and the real-time PH value and the preset reference data to obtain the real coating thickness of each subarea under the reference condition, judging that the real coating thickness reaches the standard according to the preset trigger compensation condition, if the real coating thickness reaches the standard, continuously monitoring on line, and if the real coating thickness does not reach the standard, determining the current increment of each subarea. According to the method, the current increment is compensated according to the influence of the PH value deviation, the reference current is adjusted, the final adjustment current is output, and the plating thickness of each subarea is dynamically compensated, so that the multi-dimensional process variable can be sensed, and the plating thickness of each subarea can be dynamically compensated by dynamically adjusting the current and compensating the thickness deviation caused by parameter fluctuation. And finally, the whole-course accurate control of the coating thickness of the automobile part is realized.

The invention discloses an LED lamp strip dynamic image quality intelligent optimization method and system based on deep learning driving, and relates to the technical field of LED intelligent control. The LED lamp strip dynamic image quality intelligent optimization method and system based on deep learning driving comprises the steps of S1, preprocessing real-time LED working condition data, memory reference data and historical fluctuation abnormal data; s2, adaptively dividing the LED lamp strip into different independent partitions according to spatial and physical attributes through multi-dimensional feature analysis and clustering; s3, based on a time sequence derivative sensitivity theory, judging thermal drift instability and color temperature abnormity, and entering a sudden change criterion process; s4, executing a sudden change criterion process according to a trend sudden change composite criterion result, and entering an offset compensation process; and S5, based on a criterion compensation theory, executing an offset compensation process. The problems of image quality degradation and color difference accumulation after long-term operation due to the fact that an existing LED lamp strip image quality intelligent optimization system is difficult to achieve real-time self-adaptive compensation of color temperature drifting and color cast phenomena caused by chip junction temperature rise are solved.

The invention relates to the technical field of textile bio-manufacturing and industrial process automation control, in particular to a bio-based fabric production temperature and humidity intelligent regulation and control system which comprises a data acquisition unit used for acquiring temperature and humidity sensor values in a fermentation curing workshop in real time and reading current heating power of an actuator and a fan rotating speed instruction; the reference state reconstruction unit is used for receiving the input state vector and generating an interference-free reference vector; the disturbance simulation unit is used for generating a theoretical damaged state vector representing a specific fault mode; the double-track difference unit is used for generating a real difference feature vector and a theoretical difference feature vector; the coupling regulation and control unit is used for calculating a similarity value of the real difference feature vector and the theoretical difference feature vector in a number domain space; according to the method, non-ideal factor interference is effectively eliminated, and the physical interpretability and numerical traceability of the reference data are ensured.

The invention relates to the field of shield tunneling machines, and discloses a fault diagnosis method for a quasi-rectangular shield tunneling machine based on artificial intelligence, and the method comprises the steps: collecting the operation state data of the shield tunneling machine through a sensor array, so as to obtain the operation data of the shield tunneling machine; performing component operation state difference analysis on the operation data of the shield tunneling machine to obtain shield component state difference data; performing abnormal feature distribution density analysis according to the shield part state difference data to obtain abnormal feature distribution density data; according to the method, sensor data collection, state difference analysis and abnormal feature analysis are combined, so that the operation state and the abrasion condition of each part of the shield tunneling machine can be accurately monitored, and the reliability of the shield tunneling machine is improved. A fault prediction model is established through a deep reinforcement learning method, potential faults can be recognized in advance, sudden faults are reduced, and the reliability and safety of equipment are improved.