141results about How to "Improve fitting accuracy" patented technology

The invention discloses a high-speed rail system sliding layer peak displacement envelope curve and rail direction irregularity prediction method considering seismic oscillation and pier height uncertainty and application. The method comprises the following steps: constructing a multi-mode hybrid neural network model TransTCN embedded with field priori knowledge; feature extraction is carried out on multi-modal input features of a time sequence based on seismic oscillation fusion structure features and a space sequence for outputting peak displacement based on an elastic response spectrum method; the pier height features and the seismic oscillation features are subjected to dimension fusion to serve as time sequence features, and a time dependency relationship is processed; processing a spatial dependency relationship by taking an elastic response spectrum analysis result as a spatial feature; and jointly inputting the extracted features into a constructed multi-mode hybrid neural network model TransTCN, and realizing prediction of the peak displacement envelope curve and rail direction irregularity of the sliding layer of the high-speed rail system under the consideration of seismic oscillation and pier height uncertainty. The prediction precision is improved, the training sample demand is reduced, the regional high-speed rail system damage is rapidly evaluated, and the practical value is good.

The invention discloses a lithium battery state estimation method based on relaxation time scale modeling, and the method comprises the steps: 1, carrying out a sudden pulse charging test on a lithium battery, collecting voltage and current data, and calculating the open-circuit voltage, direct-current internal resistance and a time-domain impedance sequence of the lithium battery; and 2, converting the time domain impedance sequence of the lithium battery into a frequency domain impedance sequence after discrete Fourier transform, establishing a fitting equation set of a distribution function of the relaxation time scale DRT of the lithium battery and the frequency domain impedance sequence, and solving by adopting a regularized least square method. And 3, taking the solved DRT model distribution function as the input of a neural network, taking the state parameters of the lithium battery as the output of the neural network, and establishing a prediction model of the state parameters of the lithium battery. According to the method, EIS test required by traditional relaxation time scale modeling is avoided, the control complexity and design cost of test acquisition equipment are effectively reduced, and the method can be effectively applied to occasions such as battery life prediction, battery classification and screening, battery echelon utilization and the like.

The invention discloses a mobile terminal multi-size webpage adaptive display method based on a browser viewport, which belongs to the field of browsers and comprises the following steps of: firstly, selecting iPhoneX (375px width) as a reference viewport size; and then, calculating a zoom factor according to the proportion of the viewport width of the current equipment to the reference viewport width, and dynamically adjusting the webpage content. And then, zooming the page root element by using the zooming factor. In order to adapt to screen widths and resolutions of different devices, breakpoints are set, a compact layout is applied to a narrower device, and more blanks and spaces are reserved on a wider device. And finally, recalculating the scaling in real time and updating the page. The method has the beneficial effects that a finer self-adaptive design is realized by dynamically calculating the zoom factor, adjusting the page layout in real time and optimizing display according to the breakpoint of the equipment.

The invention discloses a Raman spectrum fluorescence baseline correction method and device, and the method comprises the steps: collecting a time domain signal of a to-be-detected sample, and obtaining an IRF function of a to-be-detected sample collection system; pre-processing the time domain signal, and carrying out deconvolution on the pre-processed time domain signal and an IRF function to obtain a deconvolution result; fitting the second half data of the deconvolution result by using a support vector regression model to obtain a fitted curve; wherein a radial basis function is adopted as a kernel function, and hyper-parameters of the support vector regression model are optimized through a grid search and cross validation method; performing difference operation on the deconvolution result and the fitting curve to obtain a Raman attenuation curve; and windowing the Raman attenuation curve and averaging to obtain a Raman spectrum. According to the method, a fluorescence attenuation model and parameters thereof do not need to be preset, the IRF deconvolution result is processed and the time domain baseline is predicted by utilizing the characteristics of the time domain signal data and the support vector regression model, the signal-to-noise ratio of the extracted Raman signal can be greatly improved, a good effect is achieved, and the method is suitable for practical application.

The invention discloses a chamfer defect detection method, device and equipment and a storage medium, and relates to the technical field of defect detection.The method comprises the steps that an obtained to-be-detected device image is preprocessed, and chamfer contour points and chamfer contour point pixel coordinates are determined in the preprocessed image; correcting the pixel coordinates of the chamfer contour points, and fitting a chamfer edge straight line by using the corrected pixel coordinates of the chamfer contour points; according to the chamfer edge straight line, screening the corrected chamfer contour point pixel coordinates, and determining a target chamfer end point; and calculating the actual chamfer length based on the target chamfer endpoint, and comparing the actual chamfer length with the preset chamfer length to judge whether the chamfer has defects or not. According to the method, the detection requirements of chamfers of different shapes and sizes can be met by preprocessing and correcting the images, fatigue and subjective judgment errors possibly existing in traditional manual detection are avoided through automatic image processing and data analysis processes, and the detection efficiency is improved.

The invention provides a heat source tower heat pump applicability evaluation model construction and applicability evaluation method and system. The heat source tower heat pump applicability evaluation model construction method comprises the steps that historical state data of a heat source tower heat pump, all index values of the historical state data under a preset applicability evaluation index system and real applicability evaluation results corresponding to all the index values are obtained; taking each index value of the historical state data under the applicability evaluation index system as input data of training data; taking a real applicability evaluation result corresponding to each index value as output data of the training data; based on the input data and the output data, training a neural network model by using a particle swarm algorithm to obtain an applicability evaluation model; the constructed applicability evaluation model is based on the applicability evaluation index system, the subjectivity of manual judgment can be reduced, the neural network model is trained by adopting the particle swarm optimization, the fitting precision of the applicability evaluation model can be improved, and the accuracy of model prediction can be improved.

The invention relates to the technical field of numerical control system information processing, in particular to a curved surface numerical control machining information extraction method and device based on B-spline curved surface refactor.The method comprises the steps that after three-dimensional space position information of a curved surface machining cutter location point is obtained, cutter location point parameterization is carried out to determine B-spline curved surface parameters of a cutter positioning curved surface; determining a precision index according to the sum of squares of deviations of all cutter location points and the cutter positioning curved surface to calculate a regression matrix, determining a fairing energy index to calculate a fairing energy matrix, combining the two indexes to determine a comprehensive index, generating a curved surface reconstruction problem, solving the problem to obtain a reconstruction structure, and extracting curved surface numerical control machining information. Therefore, the problem that in the related technology, when a curved surface reconstruction algorithm processes disordered spatial discrete cutter location points, the requirements for rapidity, fairness and precision are generally difficult to consider is solved.

The invention discloses a novel power system output regulation and control strategy generation method and device and a control system, and belongs to the technical field of power system regulation and control. The output regulation and control strategy generation method comprises the following steps: representing net load historical data corresponding to net load climbing demand uncertainty characteristics by adopting a Gaussian mixture model; the uncertainty of the net load climbing requirement is reduced; inputting the net load day-ahead prediction data corresponding to the net load historical data and the net load uncertainty scene set into a two-stage risk avoidance model based on the lowest operation cost and the worst condition risk probability assessment theory; the two-stage risk avoidance model divides the whole strategy generation process into day-ahead optimization scheduling plan generation and intra-day real-time scheduling plan generation; the day-ahead optimal scheduling plan is optimized by using the obtained intra-day real-time scheduling plan to obtain a safe and economic output scheduling strategy; and furthermore, the novel power system is controlled in different scenes to realize efficient, stable and economical operation.

The invention relates to a deformation measuring device and method capable of deforming along with the rear edge of a variable-camber wing, and belongs to the field of wing deformation measurement. The key component of the deformation measuring device is a deformation beam provided with a fiber grating sensor, the deformation beam is connected with the wing through a sliding support, a positioning pin and other structures, according to the deformation characteristics of the rear edge of the variable camber wing, a sliding connection mode is adopted in a large deformation area, and a fixed connection mode is adopted in the rest area. The deformation beam can deform along with a wing after being connected with the wing structure, and the length change of a sliding connection area can be measured through a stay wire displacement sensor. The problems that in an existing wing deformation measurement technology, large deformation generated by a flexible structure is difficult to directly measure, large deformation measurement and small deformation measurement are difficult to consider, errors exist in the sensor manufacturing and installing process and the like can be solved, and efficient and accurate wing trailing edge profile measurement is achieved.

The invention provides a multi-feature fusion-based black box attack detection method (MFBA), device and system, different features are extracted by using a multi-model fusion strategy, the effectiveness and quality of opposite samples are improved, and the MFBA comprises the following three key steps: firstly, obtaining an initial data set from a target model; a decision boundary is approached by using a multi-feature fusion model; secondly, the stability of the model is enhanced through multiple rounds of training application data expansion and incremental learning technologies; and finally, generating opposite examples by using the optimized fusion model. By constructing the comprehensive attack framework, how to combine various model features to improve the performance against the resistance attack is demonstrated; compared with a single substitution model, the antagonism examples generated by the MFBA have higher mobility under different disturbance levels and have better attack effects; according to the method, an effective solution is provided for black box antagonism attacks, and the model in the generation of the advantage opposite paradigm of multi-feature fusion is verified.

The utility model provides a single-station air bag chip mounter. The single-station air bag chip mounter comprises a machine body, an operation panel, a heating carrier seat, a lifting driving device, a mounting seat, a rubber mat, a soft air pipe and an air pump, the operation panel is installed on the machine body and connected with the heating carrier, the lifting driving device and the air pump through electric signals. The heating carrier is fixed at the lower part of the machine body; the lifting driving device is mounted at the upper part of the machine body and a power output end is connected with the mounting seat; the mounting seat is positioned above the heating carrier seat; a downward opening is formed in the mounting seat; the outer periphery of the rubber mat is fixed on the mounting seat; an air chamber is formed between the rubber mat and the mounting seat; an air pipe connector communicated with the air chamber is arranged above the mounting seat; and the air pipe interface is connected with an air pump through a flexible air pipe. According to the design, the wafer can be tightly attached to the ceramic plate.

The invention relates to a battery spare part prediction method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring various data of a plurality of in-use battery packs to obtain total data; wherein the plurality of in-use battery packs are arranged in different electric equipment; determining a failure change relationship according to the total data; the failure change relationship comprises a relationship that the failure probability of the battery pack changes along with the first survival duration; according to the failure change relationship, predicting the failure number of the plurality of in-use battery packs in a preset future time period; and determining the number of battery pack spare parts required in the preset future time period according to the failure number. According to the invention, the accuracy of battery spare part prediction can be improved, so that spare parts can be better supplied.

The application provides a heterogeneous computing power automatic adaptation method and system for an intelligent chip, relates to the technical field of heterogeneous computing adaptation of the intelligent chip, and acquires computing graph structure and operator feature data and hardware state data of a to-be-deployed model; an execution knowledge base is constructed, a scheme list is generated in combination with real-time data and the execution knowledge base; execution code is distributed according to the decision weight of the scheme list; the current execution state is evaluated at a predefined execution decision point to determine whether to switch the adaptive path; the actual execution effect and environment data generated after each decision are summarized, execution track records are formed, and the execution knowledge base is updated to periodically optimize the basis for scheme generation. The application can realize automatic perception, decision, optimization and online learning of a computing task in a heterogeneous computing power environment for an intelligent chip, significantly reduces the adaptation cost and improves the resource utilization efficiency.

A broadband voltage transformer error compensation method is proposed. First, a broadband test platform is constructed using a signal generator, power amplifier, standard voltage transformer, and high-bandwidth data acquisition system. The ratio and phase errors of the transformer at multiple frequency points are acquired through primary-side excitation and secondary-side synchronous acquisition. Based on the slope, first derivative, and second derivative characteristics of the error curve, the entire frequency band is adaptively divided into segments. Within each segment, multiple candidate models are fitted using a polynomial or exponential-linear combination model. The optimal model is selected based on the coefficient of determination, mean absolute error, and root mean square error. Ratio and phase error compensators are constructed based on the fitting functions of each segment, and error compensation is achieved using a feedforward approach. This method can significantly reduce amplitude and phase errors over a wide frequency range, improving the dynamic measurement accuracy of the voltage transformer.

This invention relates to the field of CNC machine tool technology, and provides a nonlinear error control method and control system for CNC machine tools. The method includes: constructing a nonlinear error dynamic calculation model integrating axis speed and acceleration constraints; traversing the time nodes of the interpolation cycle between two adjacent tool positions; obtaining the nonlinear error value at each time node based on the nonlinear error dynamic calculation model; obtaining the maximum nonlinear error and its corresponding actual position; using the actual positions of two adjacent tool positions and the maximum nonlinear error as fitting feature points, constructing an equivalent curve equation for the nonlinear error using a piecewise quadratic function fitting method; and performing error control based on the equivalent curve equation and the allowable error value. If the nonlinear error of the interpolation segment meets the accuracy requirements, no optimization is performed; otherwise, interpolation point optimization is performed to obtain the theoretical trajectory. This invention can improve the accuracy of error control.

This invention relates to an intelligent method and system for collecting reconnaissance data for power distribution network projects. The method includes: collecting real-time audio data, historical audio data, and geographic coordinate information from the reconnaissance site; constructing a dedicated corpus for power distribution network reconnaissance based on the historical audio data; fine-tuning a general reconnaissance speech recognition model using the dedicated corpus; inputting real-time audio data into the fine-tuned general reconnaissance speech recognition model to generate original reconnaissance text data; constructing a knowledge base for the power distribution network reconnaissance domain, performing domain matching on the original reconnaissance text data within the knowledge base to generate structured reconnaissance data that conforms to power distribution network projects; and fusing geographic coordinate information and the structured reconnaissance data to obtain a reconnaissance dataset with geographic coordinate location information.

This application relates to the field of gas sensor packaging technology, specifically to an automated packaging method, system, and packaging device for gas sensors. The method includes: pre-establishing a packaging process parameter-gas-sensitive channel adaptation model; calculating the actual coaxiality deviation and correcting the sealing pressure reference value based on the gas-sensitive channel aperture; starting the integrated packaging unit, applying pre-pressure according to the corrected sealing pressure, and simultaneously collecting the temperature of the sealing area through an infrared temperature measurement module to determine whether dynamic adjustment of the welding temperature is needed; after welding, maintaining the corrected sealing pressure, and dynamically adjusting the pressure holding time based on the gas-sensitive channel length and the creep coefficient of the sealing material to complete the sealing process. This application achieves a precise balance between sealing performance and conduction efficiency by establishing a dynamic adaptation model of process parameters and gas-sensitive channel structure, overcoming the coupling imbalance bottleneck of existing technologies.

The present invention provides a method, device, equipment and storage medium for predicting the core loss of a magnetic element, including obtaining the actual core loss value corresponding to the magnetic element when it works within the working temperature range; partitioning the working temperature range of the magnetic core of the magnetic element to construct a piecewise function that characterizes the core loss in each temperature range; constructing a core loss model based on the piecewise function; estimating the parameters of the core loss model with the goal of minimizing the sum of squares of the error between the predicted core loss value and the actual core loss value to obtain the core loss prediction model; based on the core loss prediction model, the prediction of the core loss of the magnetic element is achieved. By using the piecewise function, the temperature is converted into a qualitative variable, and its influence on the core loss is divided into different intervals. The core loss model is constructed based on the piecewise function, and the model can better capture the local changes of experimental data in different temperature ranges, thereby improving the fitting accuracy.

The present application relates to the technical field of magnetic element modeling and control, and particularly relates to a multi-factor magnetic core loss prediction method based on machine learning. The method first acquires the magnetic flux density data of the magnetic element, extracts its distribution characteristics and shape characteristics, and realizes excitation waveform recognition by constructing a classification model; secondly, a temperature index correction term is introduced based on the Steinmetz equation to establish a temperature-sensitive loss model; further, the entropy weight method and statistical analysis are used to evaluate the importance of multiple factors; an integrated learning model of random forest and LightGBM is used to construct a magnetic core loss prediction model; and a single-objective optimization model is established by combining differential evolution and simulated annealing algorithm with the objectives of minimum loss and maximum magnetic energy transmission. The method improves the generalization ability and prediction accuracy of the model, can adapt to the modeling needs of magnetic core loss under multiple material, multiple waveform and multiple working condition conditions, and is suitable for application scenarios such as high-frequency magnetic element design and power electronic system optimization.

The invention relates to a one-dimensional function fitting method and system based on input dimension expansion, and the method comprises the steps: firstly obtaining a target function and a training sample thereof, carrying out the dimension expansion of each input scalar according to an odd-dimension central symmetry rule: placing the scalar at the center of an expansion vector, and filling the two sides with the same constant related to a domain of definition; and constructing a neural network taking the extended vector as input, and adjusting the width of the hidden layer to enable the deviation between the total parameter number and the reference model not to exceed a set value, thereby realizing parameter redistribution. Then, a second-order optimization algorithm is used to train the network. During prediction, new input is input into the trained model after dimension expansion according to the same rule, and a predicted value is obtained. According to the method, parameter arrangement symmetry is actively broken, and the loss function landscape is remarkably improved.

The invention relates to the technical field of high-strength steel production, in particular to an ultrahigh-strength steel cold-rolled material constitutive hardening model construction method. Simulating rolling; measuring mechanical properties; data processing and drawing; and performing linear fitting. The method has the advantages that based on the idea of combining physical experiments and mathematical modeling, a constitutive model is established through the systematized process of'simulated rolling + mechanical property measurement + double logarithmic linear fitting ', data processing software is adopted for data processing and linear regression, a power function hardening model is obtained, and the method has the advantages of being simple in structure, convenient to operate and high in practicability. The method shows excellent fitting precision in 980MPa and 1180MPa ultrahigh strength steel, the model has clear statistical support, and the prediction accuracy and reliability of the material hardening behavior are remarkably improved; the whole modeling process is clear in step and conventional in equipment, only a conventional two-roller cold rolling mill, a universal tensile testing machine, a wire cutting machine and general data processing software are needed to complete the modeling process, and the whole process from sample preparation to model output is high in standardization degree.

The utility model relates to the technical field of wafer production equipment, in particular to a double-station film sticking machine which comprises a working platform, a film sticking table installed on the end face of the working platform in a sliding mode, a film sticking mechanism arranged on the film sticking table and a control table arranged on one side of the working platform and electrically connected with the working platform. A rotating platform is arranged between the film pasting mechanism and the working opening, a film pasting station and a preparation station are arranged on the end face of the rotating platform, the rotating platform can rotationally drive the film pasting station and the preparation station to be alternately changed, and the film pasting table can slidably drive the film pasting mechanism to the top of the film pasting station. A feeding reel used for feeding a film to the film pasting station and a waste film collecting reel used for collecting backing paper are further arranged on the side, back on to the rotating platform, of the film pasting table, and a film separating assembly used for separating the film from the backing paper is further arranged between the feeding reel and the film pasting mechanism. The utility model aims to solve the problem of manual film pasting errors and improve the film pasting efficiency.

The invention relates to the technical field of land pollution treatment, solves the problem that in the prior art, when a large-area medium polluted by hexavalent chromium is treated, the addition amount of ferrous sulfide is prone to being inaccurate, and consequently the remediation effect is not ideal, and provides ferrous sulfide particles with shells and a preparation method and application of the ferrous sulfide particles. The prepared ferrous sulfide particles with the shells can keep efficient removal of chromium within the pH value range of 5-9, the problem that the adsorption behavior of a traditional model in a low-chromium-concentration area is difficult to predict is effectively solved by means of a correction adsorption model with higher fitting precision, and therefore the actual demand quantity of target repairing liquid can be accurately predicted, and the application prospect is wide. The target repairing liquid can be injected more accurately, and excessive or insufficient injection is avoided. After hexavalent chromium is reduced into trivalent chromium by ferrous sulfide, the trivalent chromium and carboxymethyl cellulose can form a stable chelate, and zero-valent sulfur can also form a compact passivation layer on the surface of a polluted medium, so that the trivalent chromium is prevented from being reoxidized, and long-term stable repair of hexavalent chromium is realized.

The application provides an SDA spectrum design automation method, comprising: setting a database, the database storing curve data and radiation flux parameters of standard lamps, and importing corresponding target curves; generating a plurality of formula groups according to the target curves, each formula group containing at least one setting parameter of a standard lamp, the setting parameter comprising a lamp name, a current value and an occupancy ratio; performing curve fitting calculation on each formula group to generate a fitting curve, respectively acquiring an absolute amplitude curve of each group of standard lamps, and performing area integral superposition on all absolute amplitude curves participating in fitting to obtain the fitting curve; verifying the fitting curve through a multi-dimensional verification strategy to obtain a score result, selecting a formula group with the highest score as an optimal formula group, and judging whether the optimal formula group meets an expectation; if the optimal formula group meets the expectation, outputting the optimal formula group; if the optimal formula group does not meet the expectation, initializing a formula and regenerating a new formula group.

The present application relates to the technical field of strain sensor detection calibration, and particularly relates to a sub-micron FBG strain sensor calibration device and a calibration method. In terms of the calibration device, the first guiding clamping mechanism and the second guiding clamping mechanism realize synchronous displacement movement by relying on the guiding action of the first force bearing plate and the second force bearing plate respectively, the micro-displacement required for calibration operation is generated by changing the clamping interval of the first guiding clamping mechanism and the second guiding clamping mechanism, and combining the geometric constraint relationship of the first force bearing plate and the second force bearing plate fixed at a preset angle. During the calibration operation, the displacement detection mechanism collects real clamping interval change data of the first guiding clamping mechanism and the second guiding clamping mechanism in real time, the control mechanism dynamically corrects the output parameters of the micro-displacement driving part according to the real-time displacement feedback signal, realizes the sub-micron high-precision closed-loop control of the calibration displacement, and is beneficial to greatly improving the fitting accuracy of the wavelength parameters and the strain response parameters of the FBG strain sensor.

The invention discloses a preparation method of a halftone photomask, which comprises the following steps: S1, a binary mask substrate is subjected to laser photoetching, and a pattern opening on the substrate comprises a binary mask area and a halftone mask area; s2, developing and etching the substrate, removing the photoresist, and cleaning the substrate; s3, coating an ultraviolet light curing adhesive; s4, carrying out ultraviolet laser photoetching on a binary mask region; s5, developing and removing the uncured protection area; and S6, depositing a halftone layer. According to the preparation method disclosed by the invention, the damage to the binary mask chromium layer in the halftone film layer deposition process is reduced, the process yield is improved, the halftone mask region is photoetched, and the binary mask region and the mask region are photoetched at the same time, so that the process flow is simplified, and the process does not need to be returned to a factory for secondary photolithography after being sent out for coating, thereby shortening the process and improving the production efficiency. The nesting precision and the line precision of the halftone mask plate product are improved, the product quality is improved, and the production efficiency is improved.

The invention is suitable for the technical field of video processing, and particularly relates to a video compression method, a video decompression method, an intelligent terminal and a storage medium, and the method comprises the steps: obtaining a to-be-compressed video, selecting two groups of independent picture frames for cutting, and obtaining two groups of cut image blocks; constructing two groups of pattern recombination functions, and generating recombination cardinal numbers according to the two groups of pattern recombination functions; based on the first recombination cardinal number and the second recombination cardinal number, carrying out pixel exchange on each group of independent picture frames, generating a pixel exchange image, calculating the pixel order degree of the pixel exchange image, selecting a group of exchange sorting images, carrying out pixel partitioning on the exchange sorting images, and carrying out multi-function fitting on the partitioned pixels; and an image fitting function is obtained and stored. According to the method, the pixels are partitioned and fitted to obtain the image fitting function, and through multi-function fitting, the fitting precision of the pixel coordinates is greatly improved, lossless compression is realized, and the data compression efficiency is improved.

The invention relates to the technical field of battery sealing nail welding, and provides a battery sealing nail welding defect repairing method. The repairing method comprises the following steps: acquiring an aluminum wafer of which the diameter is greater than that of a liquid injection hole of the battery top cover; removing the sealing nail at the liquid injection hole of the top cover of the battery; cleaning a liquid injection hole area; carrying out penetration type stitch welding by taking the wafer as an upper layer and the battery top cover as a lower layer to form a welding seam for melting the wafer and the battery top cover into a whole; and treating the welding seam and the wafer, enabling the outer surface of the wafer to be flush with the outer surface of the battery top cover, and enabling the liquid injection hole area not to be visually recognized. According to the repairing method, the aluminum wafer and the battery top cover are subjected to penetration stitch welding, and the welding seam and the surface of the wafer are treated, so that the welding defect of the sealing nail is repaired, the situation that the battery is degraded and scrapped due to poor welding or appearance defect of the sealing nail is avoided, and the quality requirement of the battery is met.

The utility model discloses a material stripping module. The material stripping module comprises two supporting seats, and a first compression roller assembly, a second compression roller assembly and a stripping knife assembly which are connected between the two supporting seats, the second compression roller assembly is located on one side of the first compression roller assembly. The stripping knife assembly is located on one side of the second compression roller assembly. A sliding driving device capable of driving the stripping knife assembly is further arranged on the supporting seats, and the stripping knife assembly is arranged on the two supporting seats; a material belt sequentially passes through the first pressing roller assembly, the stripping knife assembly and the second pressing roller assembly, and a product on the material belt is stripped from the material belt at the stripping knife assembly. The stripping module has the beneficial effects that in the advancing process of the material belt, the first pressing roller assembly and the second pressing roller assembly are matched to fix the material belt, the stability of the stripping module in the stripping process is improved, and semi-finished glue can be effectively prevented from accidentally falling off in the stripping process.

The invention discloses a method for measuring a magnetic field of an inertial measurement system, which realizes accurate reconstruction of a global magnetic field vector field by planning a discrete point test layout, synchronously acquiring multi-dimensional data and constructing a magnetic field vector fitting model. The method comprises the steps of test area division and discrete point layout, multi-source data synchronous acquisition, data preprocessing and abnormal value elimination, global magnetic field vector field fitting modeling, and fitting result verification and optimization. According to the method, a discrete point layering layout strategy is adopted, a multivariable fitting algorithm is combined, magnetic field space change characteristics can be effectively captured, global magnetic field vector field fitting errors are controlled, and the method is suitable for inertial measurement system magnetic field performance detection and optimization design and has the advantages of being high in measurement speed, high in representation precision and high in practicability.