76 results about "Soft sensor" patented technology

Disclosed in the present invention is a lightweight active online modeling method fusing mechanism and dynamic memory, which comprises the following steps: S1, data preprocessing and model initialization; S2, online data representativeness evaluation based on dynamic memory of process data; S3, online data informativeness evaluation based on dynamic memory of label data; S4, online monitoring of a mechanism state; S5, parallel hybrid active sampling policy; and S6, online modeling of a soft sensor model. The present invention separately designs memory policies for process data and label data and only reserves high-value data, thereby satisfying a lightweight hardware requirement; in addition, the present invention designs a time delay-based mechanism state online monitoring method to detect a real concept drift of the process, and further provides the parallel hybrid active sampling policy for detecting both a real concept drift and a false concept drift; in combination with dynamic memory and a kernel incremental regression method, the computation amount for model updating and the hardware overhead for edge deployment are reduced.

A method to perform oil in produced water analysis allows measuring the large volume of oil in produced water reliably. In the method, a time-series and physics based machine learning model of a gas oil separation plant is generated, advisory actionable items for maintaining a crude oil quality within a pre-determined threshold are generated based on machine learning model coefficients and outputs of soft sensors, and then the advisory actionable items are presented to a user.

The invention relates to the technical field of production line control, and discloses a drinking water production line water quality real-time monitoring and intelligent adjusting system based on the Internet of Things. Registration equipment is unified with units, and NTP time synchronization is performed to eliminate deviation; the shift / CIP calibrates the zero range, and the EWMA monitors the drift weight drop; constructing state vectors containing IDs / batches, and performing auditing and warehousing; median / bilateral filtering denoising is carried out, and grid interpolation is unified; sliding window multi-scale aggregation is carried out, and self-adaptive sampling is carried out according to curvature and events; normalizing and interpolating a near-time training window; estimating indexes difficult to measure by a soft sensor, and depicting residual chlorine by a mechanism model; and carrying out uncertainty weighted fusion, and outputting estimation and intervals and auditing. According to the method, the whole process of real-time monitoring, quick response and traceable compliance is realized, the risk caused by chemical shortage or ratio error and environmental protection constraint is remarkably reduced, the capacity loss caused by cleaning and plan conflict is reduced, and the factory water quality stability and the treatment transparency are improved.

A digital process twin system and method for operating a process plant with at least one automation component to control an industrial process within the process plant with at least one input ingredient and at least one output product, wherein a non-real-time simulation model of the industrial process is used to generated quality attributes as a function of process variables and process parameters, the generated quality attributes and related process variables are used as an input for a machine learning model serving as a soft sensor to estimate quality attributes of the output product as a function of measured or simulated process variables of the industrial process, and the performance of the process plant is optimized based on the estimated quality attributes of the output product, whereby the method and system allow process operations and control that are faster, more efficient, and more reliable.

The invention provides an intelligent control method for natural gas continuous heat treatment equipment based on digital twinning, and relates to the technical field of heat treatment equipment control. The method comprises the steps that a digital twinning model of the natural gas continuous heat treatment equipment is established, and production monitoring data collected on line are utilized; the method comprises the following steps: predicting combustion process characteristics and unmeasurable states through a digital twin model, performing multi-layer strategy search by adopting a hierarchical control strategy according to a state prediction result, performing safety constraint simulation verification on the multi-layer strategy in the digital twin model, and predicting unmeasurable state parameters through a soft sensor. And introducing the quality index as a soft constraint condition into optimization control, and generating a control execution parameter meeting the constraint condition. The technical problems that in the prior art, natural gas continuous thermal treatment equipment is poor in control effect and unstable in workpiece quality are solved. The technical effects that the natural gas continuous heat treatment equipment is intelligently, accurately and reliably controlled, and the product quality is improved are achieved.

The invention discloses a semiconductor packaging process dynamic optimization system and method, and relates to the technical field of semiconductor packaging process optimization. A dynamic optimization system for a semiconductor packaging process comprises a data acquisition module, a fingerprint generation module, an index evaluation module, an optimization model construction module, a parameter solving module, a parameter bias module, a feedback detection module and a test analysis module. According to the method, a soft sensor model taking batch fingerprints as a core is introduced into a semiconductor packaging process, and dynamic process data and equipment state information are combined, so that high-precision prediction of a key process quality index and an energy consumption index is realized; through fusion of an extreme value theoretical model and a tail risk quantile evaluation method, deviation correction is carried out on a preliminary prediction result of the soft sensor, and robustness and safety of process quality prediction are improved.

This invention provides a method and system for small-sample industrial time-series data augmentation and quality prediction. The method includes the following steps: acquiring a historical dataset as a training dataset; inputting the historical dataset into a variational autoencoder (VAE) model incorporated into the time-series decoding process to generate virtual samples; processing the virtual samples to obtain supplementary process samples; inputting the supplementary process samples into a pre-trained soft sensor model to obtain virtual quality samples; using an extended training dataset composed of the above data to update and train the soft sensor model; and inputting online input variables into the updated soft sensor model to predict online quality variables. The system includes a sample generation module, a hybrid training module, and a prediction module electrically connected in sequence. Compared to traditional VAE-based methods, the addition of a time-series decoder ensures that the generated virtual samples more closely approximate the distribution of real samples, exhibiting better diversity.

The invention relates to the technical field of fluid machinery intelligent control, and discloses a centrifugal circulating pump non-differential pressure sensor variable pressure control method based on edge intelligence. Water pump rotating speed and power signals are collected, subjected to normalization processing and then input to a lightweight recurrent neural network (RNN) model deployed at an edge device end; the RNN model serves as a soft sensor, and the flow and the lift value of the most unfavorable point of the pipe network are deduced in real time; and the system dynamically calculates the target lift based on a pipe network characteristic curve according to the reasoned flow, the target lift serves as a set value and forms a closed-loop control loop with the actual lift fed back by the RNN, the rotating speed adjustment amount is output to the frequency converter through a PID algorithm, and finally accurate variable pressure control without the physical differential pressure sensor is achieved. According to the method, the dependence on a tail end physical sensor is eliminated, the cost and the complexity are reduced, meanwhile, the real-time control and the data security of ultra-low delay are realized by utilizing edge calculation, and the response speed, the control precision and the energy-saving effect of the system are effectively improved.

A soft sensor for determining a frictional power of a mechanical seal by estimation has a data interface designed to receive data signals regarding a ring temperature of a sliding or mating ring of the mechanical seal and a medium temperature of the flow medium on the side of the mechanical seal facing the sliding ring and outside of a sealing gap between the sliding ring and the mating ring as an input variable of the soft sensor. The soft sensor is designed to estimate frictional power by evaluating the data signals using a heat flux model, the model configured to describe heat fluxes in the mechanical seal on the basis of a temperature difference between the ring temperature and the medium temperature, the model configured to deduce frictional power by way of the heat fluxes. The soft sensor is used with a pump and method for determining frictional power.

The invention discloses an aircraft flight process soft measurement algorithm based on an improved flow model. According to the method, a conditional cycle real value non-volume preserving model (CP-RNVP) and a conditional streaming soft sensor (CFSS) framework are provided for solving the problems that the data distribution difference is obvious and samples are unbalanced under the multi-working-condition operation condition of an aero-engine. The CP-RNVP model comprises three key components: (1) an adaptive mask mechanism based on feature correlation, wherein the range and granularity of feature selection are dynamically adjusted according to feature importance and correlation; (2) the multi-period affine coupling layer is used for modeling a complex period mode in data by introducing learnable multi-frequency parameters; and (3) a non-explicit condition injection method is adopted, condition information features are extracted through the sub-network and are adopted as modulation items in affine coupling transformation, and joint learning of multi-working-condition distribution is realized. Based on a CFSS soft measurement framework developed by a CP-RNVP model, multi-level training verification of soft measurement is realized by utilizing reversibility of a flow model, a soft measurement head is designed in the process of data mapping and inverse transformation, and multi-level soft measurement prediction and verification are carried out. Experiments show that the method has a remarkable effect on solving the multi-working-condition imbalance problem, compared with an existing method, the soft measurement precision is remarkably improved, and meanwhile the data generation quality and the distribution fitting capacity are better. The method is suitable for the fields of aero-engine state monitoring, health management and the like.

The invention discloses an automatic control method and system for fumaric acid production. The method comprises the following steps: collecting process variables in a fermentation process; constructing a soft measurement model fusing physical modeling and a residual neural network to obtain an acid production rate prediction value and prediction confidence; calculating a feed-forward control quantity, and adjusting the weight according to the prediction confidence to obtain a feed-forward control output; constructing a state vector, and establishing a Markov decision process; training a neural network control strategy by adopting a near-end strategy optimization algorithm, and outputting a proportionality coefficient, an integral coefficient and a differential coefficient; calculating feedback control quantity, fusing the feedback control quantity with feed-forward control output, and applying change rate and amplitude limitation to obtain final control output; and switching to a feedback priority control mode under a specific condition, and regularly updating a soft measurement model and a control strategy parameter. According to the invention, high-precision stable control of pH in the fumaric acid fermentation process is realized, the production efficiency and the process robustness are improved, and the method is suitable for automatic control of the large-scale fermentation process.

The invention discloses a water supply network abnormal data diagnosis method based on duration and support probability, and the method comprises the steps: calculating a local abnormal factor, quantitatively measuring the abnormality at each time point, and enabling the process of abnormality from start to end to be measured through an accumulation sum algorithm. By constructing a soft sensor, a sufficient reference sample is provided for a time sequence recorded by the sensor in real time from two aspects of time and space. By using the surplus samples, the calculation of the abnormal support rate is realized. And finally, distinguishing the fault of the sensor from the abnormal event of the water supply system by using the two indexes of the calculated abnormal duration and support rate.

The invention relates to a method for training a soft sensor for determining a target measured value of a fuel cell system (100), the fuel cell system (100) comprising a first sensor (10) for detecting a first measured value, a second sensor (20) for detecting a second measured value, and a control device (50), the control device (50) is designed to determine a target measured value from at least the first measured value and the second measured value on the basis of a model stored on the control device (50). Furthermore, the fuel cell system (100) comprises a target value sensor (40) for determining a target measurement value. The method comprises, as a step, operating (320) at least one system unit (110) of the fuel cell system (100) that affects a target value in at least one first operating point. Furthermore, the method comprises, as a step, adjusting (340) at least one model parameter of the model stored on the control device (50) in such a way that a deviation between the target value determined by the soft sensor and the target value determined by the target value sensor (40) at the first operating point of the system unit (110) is at least reduced.

A system and method for estimating a parameter for a reactive absorbance unit are provided. An exemplary method includes creating a kinetic model of an absorbance process, setting a range for each of a plurality of input parameters, based, at least in part, on operational data measured from the reactive absorbance unit. A sampling technique is used to generate a plurality of input vectors in the range of each of the plurality of input parameters. A plurality of output vectors is generated from the plurality of input vectors. A predictive model is trained with the plurality of output vectors and the plurality of input vectors. The parameter is estimated from the predictive model. The parameter is used in a control model for the reactive absorbance unit.

The present invention discloses a new soft measurement method based on one-dimensional convolution-stacked autoencoder, which aims to study and construct a soft measurement model for high-dimensional noisy data. The main core of the method of the present invention is to construct an autoencoder (AE) module that introduces one-dimensional convolution, remove the pooling layer and use a double-step convolution layer. Multiple modules are stacked to form a stacked autoencoder pre-training, minimize the reconstruction error of the module input and output and extract and save the parameters. The network is connected to the fully connected layer, and the model is initialized with the saved parameters and then fine-tuned to minimize the output and actual quality data errors, thereby completing the construction of the soft measurement model. In addition, the method of the present invention takes into account the local characteristics of the data while taking into account the data correlation. Compared with other traditional methods, this method can better process high-dimensional noisy data and obtain better results. It can be said that the method of the present invention is a more excellent method for building a soft measurement model.

Provided is an industrial process soft sensor method based on a federated stochastic configuration network, including: acquiring historical industrial process auxiliary data and corresponding product quality data; finding out optimal hidden layer parameters; processing, by a central server, to obtain global parameters, and downloading same to each factory as hidden layer parameters for a local model; obtaining output weights of a current network through an optimization algorithm, and uploading same to the server for weighted aggregation; and when the number of hidden layer nodes in a current network exceeds a maximum given value or a residual in current iteration meets an expected tolerance, completing modeling to obtain the global federated stochastic configuration network without adding new nodes. The present disclosure effectively improves prediction performance of models and protects data privacy, thus meeting industrial process soft sensor requirements.

A method of generating a soft sensor for a mechanical system includes defining a training dataset of sensor values received from at least one first sensor corresponding a first physical quantity. The method further includes training a Sparse Identification of Nonlinear Dynamics (SINDy) model using the training dataset to generate a mathematical function defining a relationship between the sensor values and soft sensor values corresponding to a second physical quantity. The method further includes embedding machine readable instructions based on the mathematical function into an embedded system of the mechanical system including at least one first sensor. The machine readable instructions, when executed by the embedded system, cause the embedded system to receive sensor values corresponding to the first physical quantity and generate soft sensor values based on the sensor values and the mathematical function.

Proposed is a multi-channel soft-sensor-based interface device and an operating method therefor. The glove-type multi-channel soft-sensor-based interface device can recognize the degree of flexibility of finger joints. The device may include a sensing module which includes a plurality of strain sensors and which collects strain sensing signals of the plurality of strain sensors. The device may also include a switching unit configured to select any one of the strain sensing signals of the plurality of strain sensors. The device may further include an amplification module configured to amplify the strain sensing signal selected by the switching unit. The device may further include an analog-to-digital (A / D) converter configured to convert the amplified sensing signal into a digital signal.

The invention designs a complex industrial process soft measurement modeling method based on double-layer dynamic random configuration width learning, and aims to better extract dynamic characteristics from industrial data and improve the performance of a soft sensor. According to the method, a double-layer dynamic feature analysis model is established in a mapping feature layer, a slow feature analysis method is used in the first layer to extract intrinsic essential features in industrial data, and the influence of noise in the data on the model is reduced; in the second layer, dynamic features in data are further extracted from features extracted by a slow feature analysis method by adopting a deterministic jump cyclic reserve pool, and meanwhile, nodes of an enhanced feature layer are obtained by adopting a random configuration method in the enhanced feature layer, so that the quality of the enhanced nodes is improved. The method designed by the invention not only has good dynamic feature extraction capability, but also improves the prediction performance of the soft sensor.

The invention belongs to the crossing field of biomedical engineering, robots and acoustic metamaterials, and discloses a miniature magnetic control robot with ultrasonic sensing feedback, a magnetic soft clamp robot comprises a clamp main body and two embedded ultrasonic soft sensors, and the clamping or loosening state of a U-shaped opening of the clamp main body can be adjusted under the action of an external magnetic field; and the clamping force of the clamp main body can be wirelessly monitored in real time by ultrasonically detecting the deviation of the natural resonant frequency of the embedded ultrasonic soft sensor. According to the invention, the embedded ultrasonic soft sensor EUSS and the magnetic driving module are coupled and integrated, so that the flexible control of the magnetic driving module and the real-time feedback function of the photonic crystal sensor are integrated; the problems that a traditional magnetic control robot lacks wireless and real-time feedback capacity, and an existing capsule robot lacks a monitoring feedback function in the medicine release process can be effectively solved.

The invention belongs to the technical field of biological process monitoring and optimization, and discloses a fermentation-hydrolysis process parameter automatic control system and method, and the method comprises the steps: collecting online process parameter time sequence data for pre-selected key process parameters; taking the online process parameter time sequence data as real-time input, and utilizing a pre-trained soft measurement model to obtain a key metabolite estimation concentration value sequence; calculating to obtain a key metabolite estimation generation rate value sequence based on the key metabolite estimation concentration value sequence; forming a real-time two-dimensional state vector based on the current estimated concentration value of the key metabolite and the current estimated generation rate value of the key metabolite; comparing and evaluating the real-time two-dimensional state vector with a predefined decision logic; when the decision logic judges that a preset triggering condition is met, a decision signal for starting hydrolysis is obtained; and the optimal opportunity for starting hydrolysis is accurately captured.

The invention discloses a water supply pipe network sensor data recovery method based on a soft sensor, and the method comprises the steps: constructing the soft sensor, creating analysis redundancy through the time series of other sensors in a pipe network and the historical series obtained by the sensors, and providing rich reference samples for data filling; and the accuracy of data filling is improved. And meanwhile, the data in the soft sensor is screened and strengthened again by adopting the maximum information coefficient and principal component analysis, so that the linear model can efficiently fill the data, and sensor data filling in the water supply network becomes an accurate and rapid process.

A sensor array for measuring a force is provided. The sensor array includes a plurality of soft sensor elements, each comprising: a sensor; a deformable element, the deformable element being deformable due to a force; and at least one element for interacting with the sensor for measuring a force by deformation of the deformable element, the deformable element extending at least partially between the sensor and the at least one element. Further, a method for measuring a force using the sensor array is provided.

Approaches for conditioning one or more process variables for configuring soft sensors, are described. According to one example, a processor may receive unconditioned data comprising data points or process variables indicating one or more characteristics associated with a process. The unconditioned data may be supplemented with an auxiliary set of process variables on detecting one or more missing process variables within the unconditioned data. A modified unconditioned data may thus be obtained. Further, a conditioned set of process variables, empirically representing the one or more characteristics associated with the process, may be identified from within the modified unconditioned data. The conditioned set of process variables may be provided to a plurality of inferential modellers to configure one or more soft sensors. A soft sensor, from among the one or more soft sensors, may then be selected for predicting runtime conformance metric associated with an outcome of the process.

The invention discloses a method, device and medium for soft measurement and multi-objective optimization control of sewage treatment process, collects dissolved oxygen, nitrate nitrogen, effluent water quality and energy consumption data of sewage treatment process; establishes a soft measurement model based on regularized online sequential extreme learning machine to predict effluent water quality and energy consumption of sewage treatment process, adopts piecewise linear chaotic mapping initialization and multivariate learning strategy to improve adaptive guided differential evolution algorithm, and obtains IMOAGDE algorithm; takes water quality and energy consumption as optimization targets, and obtains optimized set values ​​of dissolved oxygen and nitrate nitrogen through IMOAGDE algorithm; takes integral time absolute error, adjustment time and overshoot as optimization targets, uses IMOAGDE algorithm to perform parameter tuning on FOPID, finds the optimal parameter combination, and tracks and controls dissolved oxygen and nitrate nitrogen. The invention realizes online, high-precision soft measurement and multi-objective optimization control of key parameters of sewage treatment process, which is of great significance for improving sewage treatment process and energy saving and emission reduction.

A material and energy requirements analysing arrangement for influencing a mining process performed at a production site by a mining system using energy generated by an energy supply system obtains (S100) measurements of the mining process, applies (S110), in a soft sensor, available data of at least a part of the mining process in at least one dynamic mining model, the available data comprising the measurements, determines (S120), in the soft sensor, at least one key performance indicator based on the application of the available data in the at least one mining model together with dynamic properties of the mining process, the at least one key performance indicator comprising a sustainability key performance indicator, applies (S130) the at least one key performance indicator in a control problem linked to the production of the mining process and energy available from the energy supply system, optimizes (S140) the control problem using production goals and energy supply limitations as constraints and influences (S150) the mining process based on the optimized control problem.

Proposed is a multi-channel soft-sensor-based interface device and an operating method therefor. The glove-type multi-channel soft-sensor-based interface device can recognize the degree of flexibility of finger joints. The device may include a sensing module which includes a plurality of strain sensors and which collects strain sensing signals of the plurality of strain sensors. The device may also include a switching unit configured to select any one of the strain sensing signals of the plurality of strain sensors. The device may further include an amplification module configured to amplify the strain sensing signal selected by the switching unit. The device may further include an analog-to-digital (A / D) converter configured to convert the amplified sensing signal into a digital signal.

The application discloses a kind of microbial manufacturing process bacteria concentration perception method and system based on experience transmission, method includes: according to the historical data of microbial manufacturing process, construct bacteria concentration prediction general model, obtains its model parameters based on source system historical data training and establishes target system model structure;By introducing latent variable quantification model difference between source system and target system, a unified experience transmission probability framework is constructed;Based on the framework, the iterative algorithm combining expectation and gradient descent is used to estimate the target system parameters until convergence;The converged parameters are substituted into the general model to construct an online soft sensor, and process auxiliary variable data is collected in real time and the bacteria concentration perception value is output.The application effectively suppresses negative transfer by explicitly modeling the differences between systems, and in the industrial scenario where data is scarce and distribution is inconsistent, it realizes robust, accurate and real-time perception of bacteria concentration.

A plant process interface method according to one embodiment of the present invention comprises the steps of: loading a process design drawing; assigning attribute data to an entity shown in the process design drawing; loading a widget for displaying digital twin data via the attribute data; and displaying the process design drawing, the entity, and the widget on one canvas. The digital twin data may include sensor data, soft sensor data calculated from the sensors, or external environment data, and the widget may be a modal window expressing the digital twin data as text, an image, a table, a graph, or a 3D model.

This invention discloses a soil volumetric water content (VWC) measurement system based on LoRa-RSSI and a drone. The system comprises several underground LoRa nodes for collecting environmental monitoring data; a host computer embedded in the soil; and an aerial node carried by a small drone. Combining IoUT technology, this invention designs an innovative system for soil VWC measurement based on LoRa received signal strength and a drone. The system utilizes the changes in LoRa-RSSI between the soil's internal transmitter and the drone's aerial receiver during the drone's angular rotation. Combined with a Long Short-Term Memory (LSTM) network, it collects differential LoRa-RSSI values ​​and uses a deep learning (DL) algorithm to calculate soil VWC, achieving relatively accurate soil VWC data. This invention eliminates the need for depth measurement of VWC data, utilizes soft sensors to measure soil VWC, and offers low cost, high efficiency, and small measurement error, providing a novel approach for soil VWC measurement design.