97results about How to "Remove uncertainty" patented technology

The invention relates to the field of new energy vehicles, in particular to a vehicle state prediction method and related equipment. The vehicle is provided with a plurality of sensors, and the method comprises the following steps: collecting a plurality of sensor data from the plurality of sensors respectively; carrying out data fusion on various sensor data to obtain multi-modal fusion data; and predicting the vehicle state of the vehicle based on the multi-modal fusion data and a preset vehicle state prediction model to obtain a predicted vehicle state. According to the method and the device, the vehicle sensor data of different sources and different types can be integrated, so that the vehicle state prediction model can fully utilize the multi-modal fusion data to carry out operation and analysis so as to adapt to different driving environments and working conditions to carry out vehicle state prediction, and the accuracy of vehicle state prediction is improved.

The invention discloses a vital sign monitoring method based on multilayer variational mode decomposition and reconstruction, which comprises the following steps: sending a radar microwave signal to a target human body, receiving a path of returned radar microwave reflection signal, and extracting frequency domain characteristics of the radar microwave reflection signal; judging whether the current human body is in a resting state or not according to the radar microwave reflection signals, and if so, extracting the radar microwave reflection signals in the resting state; carrying out respiratory signal and heart rate signal separation reconstruction on the radar microwave reflection signal in the resting state to obtain a reconstructed respiratory signal and a reconstructed heart rate signal; phase data of the reconstructed respiration signals and the reconstructed heart rate signals are extracted respectively, wavelet transformation is conducted on the phase data of the respiration signals and the phase data of the heart rate signals respectively, respiration phase characteristics and heart rate phase characteristics are obtained, and the respiration frequency and the heart rate are calculated according to the respiration phase characteristics and the heart rate phase characteristics.

The invention belongs to the field of big data processing, and discloses a crowd travel mode identification method and system for multi-view subspace learning based on dynamic and static feature fusion, and the method comprises the steps: carrying out the preprocessing of original signaling data, and cleaning invalid data; extracting a travel track of each user, and processing possible abnormal conditions of ping-pong data and drift data; after track data of a single user in a sample time period is obtained, segmenting the track data, and extracting a single trip chain track; extracting static characteristics such as travel behaviors and travel traffic of a travel chain in combination with multi-source data such as mobile phone signaling data, an OSM road network and public traffic data; extracting time sequence features by using an LSTM model, and modeling dynamic features of a travel track; static and dynamic features are fused, and travel modes are captured from multiple views. According to the method, the complex travel mode can be more accurately captured, data redundancy and noise are reduced, and the accuracy of travel mode identification is remarkably improved.

The invention discloses an unmanned vehicle trajectory tracking control method based on adaptive time domain model predictive control, and relates to the technical field of automatic driving, and the method comprises the steps: obtaining a system predictive equation, converting a target function into a standard quadratic programming form function, and obtaining a system predictive equation based on a constraint condition and the standard quadratic programming form function; calculating to obtain a control increment sequence in a control time domain; constructing a model prediction trajectory tracking controller; a model prediction trajectory tracking controller is used for tracking the trajectories of the tested vehicle under different working conditions, and an optimal time domain data set is determined; training the optimal time domain data set by using an adaptive neural fuzzy inference system to obtain a time domain optimization module; and based on the time domain optimization module and the model prediction trajectory tracking controller, generating a self-adaptive time domain model prediction controller which is used for adjusting the prediction time domain and the control time domain of the tested vehicle under different working conditions in real time. The adaptability and trajectory tracking effect of the system are effectively improved, and the vehicle out-of-control risk is reduced.

The invention relates to the technical field of image recognition, in particular to an intelligent image quality detection and classification method based on an ultrasonic image. The method comprises the following steps: acquiring standard ultrasonic image data; performing ultrasonic image visual feature distribution analysis based on the standard ultrasonic image data to generate ultrasonic image visual feature distribution data; performing ultrasonic image effective information quantitative intelligent analysis based on the ultrasonic image visual feature distribution data to generate ultrasonic image effective information quantitative data; performing effective image screening on the standard ultrasonic image data according to the ultrasonic image effective information quantitative data to obtain effective ultrasonic image data; performing image quality intelligent analysis on the effective ultrasonic image data to generate quality identification ultrasonic image data; and performing ultrasonic image intelligent classification operation based on the quality identification ultrasonic image data. Intelligent ultrasonic image quality detection and classification are realized by analyzing the effective image information and the image blurring degree of the ultrasonic image.

The invention is suitable for the technical field of oil-gas exploration, and provides a seismic data convolution cross-correlation intelligent FWI method, which integrates a proposed seismic data convolution cross-correlation FWI target equation into a deep learning framework, and combines an automatic encoder technology to significantly reduce the nonlinearity of the traditional FWI, effectively avoid the cycle jump problem, and improve the accuracy of the FWI. Therefore, the speed model inversion precision of the FWI is improved. Finally, the FWI technology which can still carry out stable and accurate high-resolution speed modeling under the conditions that seismic source wavelets are unknown, the initial speed is unknown and low-frequency data are missing is realized, and the method has important significance on seismic exploration.

The invention discloses a robot rigidity identification method and system based on parallel rope drive loading multidirectional force, and belongs to the technical field of robot rigidity identification. Calculating to obtain an adaptability evaluation index of the initial load group and a condition number of an observation matrix corresponding to the initial load group, and further establishing a system disturbance-oriented identification load evaluation index; wherein different external loads are different in magnitude and direction; by taking identification load evaluation index minimization facing system disturbance as a target and taking combined robot bearing load constraint force limit as a constraint condition, a system disturbance-oriented variable load evaluation model is constructed, and an optimal load group is selected according to a model solving result; and based on the optimal load group, a parallel rope-driven measurement loading system is used for carrying out a rigidity identification test on the robot, coordinate values of a target ball at the tail end of the robot under different loads are measured, a deformation vector is obtained through calculation, and then the accurate joint rigidity of the robot is obtained through calculation.

The invention belongs to the technical field of eddy current nondestructive testing, and discloses a variable-diameter pipeline pulsed eddy current testing device which comprises a main body support and a lift-off and testing structure, the main body support comprises a support left arm and a support right arm, the upper end of the main body support is connected with two variable-diameter guide frames, and the tail ends of the variable-diameter guide frames are inserted into the lift-off and testing structure. The lift-off and detection structure comprises a lift-off box, the lift-off box is connected with the body support, a lift-off motor is arranged on the upper portion of the lift-off box, the lift-off motor penetrates into a lift-off rod and is connected with the lift-off box, the terminal of the lift-off rod is connected with a probe box, an excitation coil and a detection coil are placed in the probe box, a wheel support is arranged at the tail end of the body support, and a magnetic wheel is connected to the tail end of the wheel support. The left side of the magnetic wheel is connected with an encoder, and the right side is connected with a driving motor. Corrosion defects of pipelines with different pipe diameters can be detected in real time, personal errors caused by manually moving the probe are reduced, and the detection accuracy is improved.

The invention relates to a quadruped robot single-leg sliding mode control method based on a disturbance observer, and belongs to the technical field of robot control. Firstly, a foot end track of a single leg of the quadruped robot in a swinging stage is established; then, establishing a kinematics model of a single leg of the quadruped robot to realize conversion between a foot end target position and a corresponding joint angle; on the basis, a nonsingular terminal sliding mode controller is designed to cope with dynamic control under the non-interference condition; finally, a final composite control strategy is designed in combination with a sliding mode controller and interference estimation, stable tracking of the joint angle of the robot is achieved, and therefore the anti-interference performance of the single-leg system is improved. According to the method, a composite controller is designed by combining a finite time disturbance observer and a nonsingular terminal sliding mode controller, so that the foot end trajectory tracking precision and the system robustness of the quadruped robot are improved, and the uncertainty caused by external disturbance and modeling errors of the robot is eliminated.

According to the vector power synchronization control method without the phase-locked loop, self-synchronization with a power grid is achieved through power synchronization control, and active and reactive decoupling and conjugate pole elimination are achieved through a cross decoupling control structure controlled by vector current; in an adopted vector power synchronous controller, an open-loop control part is used for keeping the stability of a system and tracking a reference signal, and a closed-loop control part is used for eliminating the influence of small signal disturbance in the open-loop control part.

The invention discloses a dynamic multi-objective collaborative optimization method for a climbing operation robot. The method comprises the following steps: determining kinematic characteristic and dynamic characteristic optimization variables of the climbing operation robot according to an operation process and operation indexes of the climbing operation robot; according to the operation process of the climbing operation robot, a kinematical equation and a kinetic equation are obtained; establishing an operation efficiency maximization optimization objective function and a stability maximization optimization objective function according to operation requirements of the climbing operation robot; determining constraint conditions of kinematic characteristic and dynamic characteristic optimization variables by using a mechanism analysis method; obtaining a climbing operation robot key index regression analysis data model by using a depth data driving algorithm, describing a constraint relationship between an optimization variable and an uncertainty factor, and obtaining an optimization variable uncertainty constraint condition; determining a system global collaborative optimization target, and establishing a system global collaborative optimization model; and solving the global collaborative optimization model of the system to obtain an optimal solution set of optimization variables.

The invention relates to the technical field of data processing, and discloses an engine mold size obtaining method and system based on a scanner. According to the method and system, design parameters and initial data of the mold are accurately recorded through an input module, normalization and traceability of the design process are ensured, the mold is comprehensively scanned through a high-precision full-size scanner, actual size data of the mold can be rapidly obtained, errors and uncertainty caused by manual measurement are eliminated, and the accuracy of the mold design is improved. Through an advanced contrastive analysis algorithm, the system can quickly identify the deviation between the actual size and the design size, and timely highlights the part exceeding the tolerance range, so that design engineers can conveniently carry out targeted improvement and correction, and the process not only improves the accuracy of mold size acquisition, but also improves the production efficiency. And the detection and correction time is greatly shortened, and the production efficiency is improved, so that the overall production quality of the mold is effectively improved, and the manufacturing cost is reduced.

The utility model discloses an aluminum sheet polishing device, which belongs to the technical field of aluminum sheet processing and comprises a polishing device, supporting legs are connected below the polishing device, a control panel is connected to the front side of the polishing device, a conveying belt is arranged in the polishing device, a connecting rod is connected to one side of the polishing device, and the other side of the polishing device is connected with the control panel. And an aluminum sheet box is arranged in the connecting rod, one side of the polishing equipment is connected with a pushing device, and one side of the aluminum sheet box is connected with a cleaning device. According to the aluminum sheet polishing device, the supporting rod, the motor, the half gear, the tooth block and the push rod are arranged, the half gear drives the tooth block and the push rod to move, the push rod drives the push plate to move, the push plate drives an aluminum sheet to move through the sliding groove, then the aluminum sheet is pushed onto the conveying belt, and pushing of the aluminum sheet is completed at the moment; the uncertainty possibly caused by manual placement is eliminated, and the risks of potential inconsistency and quality fluctuation caused by manual operation are reduced.

The invention discloses a method for rapidly realizing crystallographic orientation imaging in a transmission electron microscope. The method comprises the following steps: collecting conical scanning dark field images of front 8-10 diffraction rings under inclination angles of 3-5 samples through the transmission electron microscope; carrying out drift correction on the dark field images under different sample inclination angles and then carrying out binaryzation on all the dark field images; building an incident electron beam conical scanning model, simulating electron diffraction patterns of any orientation under different sample inclination angles, and building an electron diffraction pattern database; reconstructing a diffraction pattern of each pixel in the dark field image under the same sample inclination angle, and calibrating orientations of all the pixels; and completing orientation calibration of all pixels in combination with data of multiple sample inclination angles. The rapid characterization of the grain orientation is realized based on the transmission electron microscope conical scanning dark field imaging technology and the crystallographic orientation calibration algorithm, and the method is suitable for the crystallographic orientation characterization of the crystal material under the transmission electron microscope.

The utility model discloses a non-woven fabric needling device with an adjustable hole diameter, and relates to the technical field of non-woven fabric production auxiliary equipment, in particular to the non-woven fabric needling device with the adjustable hole diameter. Comprising a main body, an anti-skid handle, an aperture generator, a pinhole storage groove, input equipment, a display, an electric adjusting device, a rotary button, a laser range finder, pinholes, hot press equipment, a positioning system module, an integrated control module and output equipment. According to the utility model, the operation is stable through the anti-skid handle, the aperture parameters are set by the input device, and the distance between the needle holes is accurately adjusted through the electric adjusting device. A laser range finder and a positioning system ensure accurate needling, and the hot press accurately controls the hot pressing process. The display displays the working state in real time, and the rotary button provides man-machine interaction. The pinhole storage groove realizes flexible production, and the output equipment prints a production report. And the pinhole storage groove and the integrated control module store the working state, so that quick recovery is facilitated.

The present invention provides a hierarchical multi-loop control method for a manipulator for compliant assembly of electronic components. The design reference trajectory, impedance force and the environmental impedance force of the previous cycle are input into the impedance feature knowledge base of the electronic component assembly scene to obtain the design impedance trajectory of the manipulator joint, and the difference is calculated with the position information of the manipulator joint. The control variable is predicted by the MPC cycle output model, and the auxiliary control variable is obtained by summing with the IHOSMC cycle output. The control torque of the manipulator system is obtained by the inverse dynamics control cycle equation, and then the joint angle variable and the output trajectory of the end effector are obtained. The environmental impedance force is obtained in combination with the environmental information; when the output trajectory of the end effector is equal to the design impedance trajectory and the environmental impedance force is equal to the design impedance force, the manipulator system is stable, and the assembly of electronic components is performed according to the output trajectory and environmental impedance force at this time. The present invention has the characteristics of high assembly accuracy and fast response speed.

An automatic device for measuring run-out and perpendicularity of a turbocharger middle shell comprises an automatic measuring clamp jig installed above a plurality of layers of shockproof tables, a start button connected to the front of the automatic measuring clamp jig, a protective guard installed on the plurality of layers of shockproof tables, and a controller installed on the protective guard. The safety light curtain, the touch screen controller and the three-color indicator lamp are installed at the front end of the protective fence, and a PLC control cabinet and a displayer of a PC main control computer are fixed to the multiple layers of shockproof tables. Safety door locks are mounted on the double doors on the left and right sides of the protective fence; and a compressed air three-stage filter, a laser coding host and a PC (Personal Computer) main control computer are also arranged on the plurality of layers of shockproof tables. According to the utility model, the jumping and perpendicularity dimensions of the turbocharger intermediate housing product can be automatically measured, and the outflow of such defective products is prevented; moreover, the product information also has a binding storage function; the traceability of the product is realized; and meanwhile, the labor cost is saved, and the working efficiency is greatly improved.

The invention is suitable for the technical field of piezoelectric ceramic wafers, and provides a method for judging positions through Hall elements, which comprises the following steps: S01, setting a first Hall element and a second Hall element; s02, driving a motor to operate at a certain speed; s03, continuously monitoring output signals of the first Hall element and the second Hall element; s04, when the first Hall element detects a first falling edge, recording the position of the motor as x1; s05, when the second Hall element detects a first falling edge, recording the position of the motor as x2; s06, calculating the position of the original point, namely (x1 + x2) / 2; s07, setting the calculated original point position as a mechanical original point of the motor; by recording the position of the motor when the first Hall element and the second Hall element detect the first falling edge, the initial position can be captured in the starting stage of the motor, an accurate reference point is provided for subsequent position control, and the average value of output signals is calculated to determine the position of the original point. And the uncertainty caused by delay of a single Hall sensor is eliminated.

The application provides a graphite ore flotation intelligent optimization method and system based on AI visual identification, and relates to the technical field of mineral processing. The method comprises the following steps: acquiring a flotation foam image in real time through an image acquisition module, extracting multi-dimensional features such as foam size distribution, stability, color depth, texture and flow rate after improved bilateral filtering denoising and adaptive histogram equalization enhancement, predicting the graphite grade, recovery rate and foam state category according to the features by an attention-enhanced ResNet50 model, dynamically generating an optimal reagent scheme by combining a safety constraint and a reward function through a reinforcement learning agent based on a deep deterministic policy gradient, and finally automatically adjusting a reagent feeding device by a control execution module to realize closed-loop optimization control. The application can accurately predict the indicators of the foam and dynamically optimize the reagent feeding scheme according to the indicators, and significantly improves the production efficiency.

The present application relates to an automatic pattern making method and system for personalized clothing customization; the method includes: obtaining the net body size data and demand data of the user; inputting the net body size data and demand data into a pre-trained automatic pattern making algorithm model for calculation, and outputting the garment size data; obtaining the human body image data of the user; inputting the human body image data into a pre-trained body type classification algorithm model for processing, and outputting the body type category label; determining the corresponding body type modification rule according to the body type category label, combining the body type modification rule with the initial pattern adjustment rule to output the final pattern adjustment rule; outputting the garment size data, change items and final pattern adjustment rules as a file in a specified format. The scheme of the present application is based on an artificial intelligence machine learning algorithm, which makes the entire clothing customization process intelligent and automated, can eliminate the uncertainty caused by human factors in the actual production process, and improve the accuracy and comfort of the final clothing size.

The invention discloses a space vector measurement auxiliary ball for radome assembly, comprising a spherical body and a lock core connection body, the spherical body comprising a hemisphere one and a hemisphere two vertically arranged on the bottom surface, the hemisphere one and the hemisphere two being fixedly connected to form an integrated structure, the outer spherical surfaces of the hemisphere one and the hemisphere two are on the same spherical surface, a U-shaped groove is arranged on the side of the hemisphere one away from the hemisphere two, the lock core connection body is a rectangular column with four corners arranged to be equilaterally chamfered, one end of the column is fixedly connected to the center of the bottom surface of the hemisphere two to form an integrated structure, and the lock core connection body is provided with a square hole for fixedly connecting a lock. The invention can realize the transfer of a theoretical connection intersection of an assembly lock that is invisible and cannot be physically measured to a physical surface that can be directly collected, and then the theoretical intersection position can be inferred, so as to predict the installation accuracy of the assembly lock, eliminate uncertainty, and ensure the quality of part delivery.

The application provides a spatial light modulation method and system of Gray code coding rearrangement, and relates to the technical field of computational imaging and spatial light modulation. The method combines a Gray code sequence and a high-performance basic modulation matrix, a basic modulation matrix is constructed according to a target image resolution and a sampling rate, a Gray code index sequence with only one different adjacent index is generated, a measurement matrix is obtained through a mapping function rearrangement matrix, the measurement matrix is loaded to a spatial light modulator to code light signals, a double-channel detector collects transmission and reflection light intensity responses and performs differential processing, and a compressed sensing reconstruction algorithm is input to recover an image. The application realizes stable and efficient information capture at a low sampling rate, improves the anti-interference ability and repeatability of the system, has strong hardware compatibility, low implementation cost, and is suitable for the fields of single-pixel imaging, three-dimensional measurement and the like.

This invention belongs to the field of high-temperature heat pipe performance testing technology, specifically relating to a fully autonomous adaptive heat pipe heat transfer performance testing method. It includes the following steps: Step S1: Generating a heat flux density benchmark based on geometric features; Step S2: Using P... base Based on this, differentiated power loading is performed using a stage-dependent loading coefficient β. During step S2, steps S3 to S5 are executed in parallel: Step S3: Parallel execution of adaptive physical stage switching based on a dual criterion of "temperature-anomaly"; Step S4: Parallel execution of autonomous steady-state determination for each power loading at key temperature points in the linear heat transfer stage and during each power loading in the limit approximation stage; Step S5: Parallel execution of full-cycle automated safety protection based on monitoring the second derivative of the evaporation section temperature. The beneficial effects are: improved adaptability of the testing method, compatible with heat pipes of different geometric dimensions; and effective improvement of testing efficiency through the implementation of a staged control strategy.

The utility model discloses a kind of battery plastic support, it includes upper support main body (1) and lower support main body (2);The upper support main body (1) and lower support main body (2) are engaged by shaft pin structure and using screw lock fixed connection, the shaft pin structure includes the multiple upper lock screw column (3) being set on the upper support main body (1), and the multiple lower lock screw column (4) being set on the lower support main body (2), the upper lock screw column (3) can be nestedly matched and inserted into corresponding lower lock screw column (4) and by screw lock, to realize the stable connection of the upper support main body (1) with the lower support main body (2).The utility model discloses a kind of battery plastic support reduces the switching lock link, simplifies assembly process, reduce cost, and more comply with the integrity principle.

The invention discloses a robot system for laboratory pretreatment autonomous operation, which relates to the technical field of intelligent robots and is formed by organically fusing four core parts, namely a moving and executing platform, a sensing module, an operation and end tool module and a control and calculation module. Autonomous operation is realized through a set of layered progressive closed-loop feedback intelligent control logic: firstly, a macroscopic experiment task is gradually decomposed into executable'action-nodes' from top to bottom; secondly, real-time closed-loop feedback is formed through multi-source 3D vision and tail end force perception, the track of the mechanical arm is dynamically corrected, and the operation force is accurately controlled; and finally, two arms are driven to work cooperatively through parameterized mapping and accurate time sequence control, so that unmanned, standardized and high-reliability operation of a sample pretreatment process is realized in a variable experimental environment, and the limitation of manual operation and an existing single-function automation scheme is effectively overcome.

The application discloses a power magnetic core common mode inductor and a preparation method thereof, and belongs to the technical field of electronic component manufacturing, and comprises a winding execution unit, a magnetic circuit assembly unit and a material flow transfer module, wherein the winding execution unit comprises a main shaft rotating motor and a wire arranging servo module, the magnetic circuit assembly unit comprises a servo press and a glue injection nozzle, and the material flow transfer module is connected with the winding execution unit and the magnetic circuit assembly unit. The scheme physically increases the average distance of interlayer conductors through staggered span actions, changes the distribution direction of electric field vectors, effectively reduces the parasitic capacitance between windings, directly improves the high-frequency characteristics of the common mode inductor, expands the high-frequency impedance bandwidth, and enables the common mode inductor to more effectively inhibit high-frequency noise transmission.

This invention discloses a system for measuring electrical parameters along a lightning path based on spatiotemporal clustering analysis, belonging to the field of lightning path electrical parameter measurement technology. It includes a data acquisition module for collecting lightning data, including the time, longitude, and latitude of the lightning event; an electrical parameter measurement module for measuring electrical parameters along the lightning path, including electric field strength, current intensity, and voltage; and sensors for real-time monitoring of areas along the lightning path to obtain accurate electrical parameter data. The system uses sensors to measure electric field strength, current transformers to measure current intensity, and voltage sensors to measure voltage. This invention significantly improves the accuracy of electrical parameter measurement, addressing the problem of poor accuracy. When measuring the instantaneous large current of a lightning strike, the current transformer, with its suitable range and fast response, can accurately record changes in current intensity, providing reliable data for studying lightning current characteristics.

This invention relates to the field of chromatographic analysis technology and discloses a data collaborative processing method and system for distributed chromatographic analysis equipment. The method includes: constructing a global time-slot pre-allocation matrix by combining gradient elution commands and real-time calibration data of solvent peaks; performing finite difference operations on discrete digital signals to extract physical feature quantities and generate local signal status flags; performing dynamic arbitration of the bus through logical comparison, and triggering an absolute timestamp deep latching mode using direct memory access redirection when bandwidth is saturated; performing time-domain reconstruction on down-frequency data, and performing asynchronous backfilling and splicing on latched data to output a globally time-axis aligned chromatographic matrix. This invention solves the problems of data packet loss and bus congestion caused by multi-node concurrency through dynamic time-slot allocation and hardware-level local latching mechanisms, ensuring the integrity of chromatographic peak shapes under extreme bandwidth occupancy, and improving the data collaborative efficiency and quantitative analysis accuracy of the distributed system.

The present invention discloses a simulation method and device for predicting the influence of SLM laser wavelength on absorption behavior, the method comprising the following steps: establishing a mesoscopic three-dimensional powder bed random stacking model; changing light source parameters and optical parameters defining powder properties to simulate laser absorption behavior at different wavelengths; obtaining a GO‑RT model consisting of a powder bed, a substrate, a light source and a spherical analysis surface in Fred software for predicting the influence of laser wavelength on SLM absorption behavior. The present invention takes into account the influence of laser energy and SLM powder layer thickness, which are not easy to accurately measure, on the laser absorption rate and laser‑powder bed interaction in the SLM forming process, and establishes a three-dimensional random stacking model by balancing the computational complexity of numerical simulation and the authenticity of powder layer stacking from a mesoscopic scale; by observing the influence of laser wavelength changes on the laser absorption behavior of the model SLM process, the problem of the influence of laser wavelength on the formability of SLM components, which is not fully considered in existing experiments and simulations, can be predicted to a certain extent.

The application discloses an AC solid-state circuit breaker adaptive turn-off method and system, the core of the method is to construct a two-dimensional phase space geometric evolution manifold according to the current amplitude sequence of the fault current and the hardware differential rate signal, realize geometric extrapolation of the short-circuit current peak value containing the non-periodic component through the trajectory curvature characteristics; combined with the cumulative equivalent operation times, the radiator temperature and the standing time and other physical state parameters, input them into the endogenous physical information neural network model, define the safe operation interval with physical certainty by using the hard-coded embedded Cauer thermal network operator; and in the interval, an analytical projection algorithm based on stress cost manifold is adopted to execute constant level optimization of the optimal turn-off time by using the control barrier function. The application can effectively solve the distortion problem of the traditional model under the magnetic saturation working condition and the uncertainty of the soft constraint prediction, while ensuring the breaking reliability, significantly reducing the online calculation complexity.