Dynamic torque compensation method for high-precision tightening operation and application

Abstract

The invention discloses a dynamic torque compensation method for high-precision tightening operation and application, and relates to the technical field of automation control, and the method comprises the steps: building a multi-modal sensing module array to collect multi-modal physical signals in the tightening operation process; extracting cross-modal spatial-temporal features of the multi-modal physical signal by using a spatial-temporal convolutional network, and performing feature fusion on the cross-modal spatial-temporal features to generate a fusion feature vector; calculating a dynamic torque compensation value based on a hierarchical near-end strategy optimization algorithm according to the fusion feature vector; and performing compensation adjustment on the tightening operation according to the dynamic torque compensation value. The method can solve the problems that an existing compensation method lacks multi-modal information fusion capability, the real-time performance of edge calculation is poor, and the control precision is poor due to dependence on a fixed threshold value control algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automation control, in particular to a dynamic torque compensation method and device for high-precision tightening operation, an electronic device and a computer readable storage medium. BACKGROUND

[0002] Traditional industrial tightening equipment relies on fixed threshold control (such as PID algorithm), which is difficult to cope with dynamic disturbances such as material deformation and thread wear, resulting in over-tightening or under-tightening risks; and existing AI compensation methods (such as CN117698442A) are mostly based on single sensor data, lack multi-modal information fusion capability, and do not solve the real-time problem of edge computing; CN117195996A discloses a torque prediction method based on deep neural network, but its application scenario is limited to vehicle line control steering, and does not cover the micro-mechanical properties of industrial tightening; CN119017317A discloses a tightening system that improves safety through fastener length detection, but lacks adaptive adjustment of process parameters. SUMMARY

[0003] In order to overcome the defects of the above-mentioned prior art, the embodiments of the present application provide a dynamic torque compensation method and application for high-precision tightening operation, aiming to combine multi-modal sensing, reinforcement learning and edge computing dynamic compensation scheme to achieve sub-microsecond response and cross-scene adaptation.

[0004] In one aspect, the embodiments of the present application provide a dynamic torque compensation method for high-precision tightening operation, comprising: assembling a multi-modal sensing module array to collect multi-modal physical signals during the tightening operation; using a spatio-temporal convolution network to extract cross-modal spatio-temporal features of the multi-modal physical signals, performing feature fusion on the cross-modal spatio-temporal features to generate a fusion feature vector; calculating a dynamic torque compensation value based on a hierarchical proximal policy optimization algorithm according to the fusion feature vector; and compensating and adjusting the tightening operation according to the dynamic torque compensation value.

[0005] In one embodiment of the present application, the use of a spatio-temporal convolution network to extract cross-modal spatio-temporal features of the multi-modal physical signals, perform feature fusion on the cross-modal spatio-temporal features, and generate a fusion feature vector comprises: independently inputting vibration signals, thermal imaging data and strain data using a spatio-temporal separation convolution method, independently inputting and pre-processing the three signals to obtain spatio-temporal feature maps of each modality; generating a trainable weight matrix for each modality through a linear layer based on the spatio-temporal feature maps; calculating a similarity matrix between modalities based on the trainable weight matrix, and calculating attention interaction between modalities according to the similarity matrix; outputting and concatenating each attention interaction and generating the fusion feature vector through a fully connected layer.

[0006] In an embodiment of the present application, the calculation formula of the attention interaction is represented as: ; wherein, represents a query vector, represents a key vector, represents a value vector, represents a feature dimension of the key vector; the query vector of the vibration modality and the key vector of the thermal imaging modality are calculated to obtain a space-time correlation weight, and the value vector is weighted and fused to obtain the attention interaction between the vibration modality and the thermal imaging modality.

[0007] In an embodiment of the present application, the space-time convolution network includes a time dimension convolution kernel and a space / spectrum dimension convolution kernel, the time dimension convolution kernel slides along the time axis to extract dynamic evolution features; and the space / spectrum dimension convolution kernel processes the spatial distribution of the thermal imaging or the energy aggregation area of the vibration spectrum.

[0008] In an embodiment of the present application, the dynamic torque compensation method for high-precision tightening operation further includes: introducing a dynamic convolution kernel generator to adjust the convolution kernel parameters in real time according to the input data characteristics, and the expression of the dynamic convolution kernel generator is: ; wherein, is a lightweight LSTM network, and the input is the current sensor state and historical context.

[0009] In an embodiment of the present application, the expression for calculating the dynamic torque compensation value based on the hierarchical proximal policy optimization algorithm according to the fusion feature vector is: ; wherein, is a temperature term parameter, is a vibration term parameter, is a cross-coupling term parameter; In the temperature term parameter, , represents an actual measured temperature, wherein, represents a reference temperature, represents a temperature-torque coupling coefficient, represents a temperature change rate sensitive coefficient; In the vibration term parameter, , wherein, represents a deformation strain ε correction factor for vibration-temperature coupling, is a basic vibration-temperature gain coefficient, is a spectrum energy weight factor, is a nonlinear attenuation coefficient, is a vibration acceleration; , wherein, a deformation-torque conversion coefficient, a dynamic strain rate correction factor; in the cross-coupling term parameter, wherein, a temperature gradient and mechanical coefficient, a dynamic change rate coupling factor, second derivative a strain acceleration.

[0010] In an embodiment of the present application, the dynamic torque compensation method for high-precision tightening operation further comprises: For metal materials with obvious yield phenomenon, the yield strength calculation formula is constructed as: ; wherein, Re represents the yield strength, Fe represents the force maintained constant in the yield stage, represents the original cross-sectional area of the sample; For metal materials without obvious yield phenomenon, the specified non-proportional elongation strength is used as the yield strength, which is represented as: ; wherein, represents the force corresponding to 0.2% plastic deformation of the material.

[0011] On the other hand, the embodiment of the present application also proposes a dynamic torque compensation device for high-precision tightening operation, comprising: a multi-modal signal acquisition module for assembling a multi-modal sensing module array to acquire multi-modal physical signals in the tightening operation process; a cross-modal feature fusion module for extracting cross-modal spatio-temporal features of the multi-modal physical signals using a spatio-temporal convolution network, performing feature fusion on the cross-modal spatio-temporal features, and generating a fusion feature vector; a dynamic torque compensation value calculation module for calculating a dynamic torque compensation value based on a hierarchical proximal policy optimization algorithm based on the fusion feature vector; a compensation adjustment module for compensating and adjusting the tightening operation according to the dynamic torque compensation value.

[0012] In still another aspect, the embodiment of the present application also proposes an electronic device, comprising: a memory and one or more processors connected to the memory, the memory storing a computer program, and the processor being configured to execute the computer program to implement the dynamic torque compensation method for high-precision tightening operation as described in any one of the above embodiments.

[0013] In still another aspect, the embodiment of the present application also proposes a computer readable storage medium, the computer readable storage medium storing computer executable instructions, the computer executable instructions being configured to execute the dynamic torque compensation method for high-precision tightening operation as described in any one of the above embodiments.

[0014] As can be seen from the above, the above embodiments of the present invention can have at least one or more of the following beneficial effects compared with the prior art: The dynamic torque compensation method for high-precision tightening operations proposed in the present invention collects multimodal physical signals of the tightening operation process through a multimodal sensing module, uses a spatiotemporal convolutional network to extract cross-modal spatiotemporal features, generates a high-dimensional state vector, and dynamically calculates the dynamic torque compensation value online based on the hierarchical proximal policy optimization (PPO) algorithm, thereby realizing multimodal information fusion and solving the spatiotemporal asynchrony problem of multimodal sensor data. Compared with the method using a fixed threshold, the proposed dynamic compensation algorithm improves the control accuracy of industrial tightening equipment and the real-time performance of edge computing. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 A flow chart of a dynamic torque compensation method for high-precision tightening operations provided by an embodiment of the present invention; Figure 2 A schematic structural diagram of a dynamic torque compensation system provided by an embodiment of the present invention; Figure 3 A multimodal feature fusion flow chart provided in an embodiment of the present invention; Figure 4 A schematic diagram of the PPO model training provided by an embodiment of the present invention; Figure 5 A schematic structural diagram of a dynamic torque compensation device for high-precision tightening operations provided by an embodiment of the present invention; Figure 6 A schematic structural diagram of an electronic device provided by an embodiment of the present invention; Figure 7 A schematic structural diagram of a computer-readable storage medium provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described below with reference to the accompanying drawings and in combination with the embodiments.

[0017] In order to enable ordinary technicians in this field to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments, and all should fall within the scope of protection of the present invention.

[0018] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and in the above drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the terms thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to the process, method, product or device.

[0019] It should also be noted that the division of the various embodiments in the present application is only for the convenience of description, and should not constitute a special limitation. The features in the various embodiments can be combined with each other under the condition of no contradiction, and can be mutually referenced.

[0020] As shown in Figure 1 The first embodiment of the present application proposes a dynamic torque compensation method for high-precision tightening operation, for example, comprising: step S1, assembling a multi-modal sensing module array to collect multi-modal physical signals in the tightening operation process; step S2, using a space-time convolution network to extract cross-modal space-time features of the multi-modal physical signals, performing feature fusion on the cross-modal space-time features to generate a fusion feature vector; step S3, calculating a dynamic torque compensation value based on a hierarchical near-end strategy optimization algorithm according to the fusion feature vector; and step S4, compensating and adjusting the tightening operation according to the dynamic torque compensation value.

[0021] Specifically, in combination with Figure 2 As shown in step S1, the multi-modal sensing module is integrated with a high-frequency vibration sensor, an infrared thermal imager and a micro quantum strain gauge, for example. Among them, the high-frequency vibration sensor detects the thread friction frequency and measures the vibration acceleration, the infrared thermal imager monitors the material deformation heat distribution, and the micro quantum strain gauge detects the nanoscale deformation.

[0022] In step S2, a space-time convolution network (STCN) is used to extract multi-modal features and generate a high-dimensional state vector. The following is a detailed explanation of the space-time convolution network: The space-time convolution network adopts a space-time separation convolution branch, in which the input layer independently inputs vibration signals, thermal imaging data and strain data, and respectively performs preprocessing. The vibration signal is converted into a time-frequency spectrogram through short-time Fourier transform (STFT), including (time x frequency) information; the thermal imaging data is converted into a time series slice, including (space x time x temperature value) information; and the strain data is converted into an original time series signal, including (time x strain value) information.

[0023] The spatio-temporal convolution network includes a time dimension convolution kernel and a space / spectrum dimension convolution kernel, the time dimension convolution kernel slides along the time axis to extract dynamic evolution features (such as vibration frequency trend); the space / spectrum dimension convolution kernel processes the spatial distribution of thermal imaging or the energy aggregation area of vibration spectrum. The spatio-temporal convolution network outputs the spatio-temporal feature map of each mode (the dimension is unified to Batch x Time x Channel).

[0024] The spatio-temporal convolution network adopts a cross-modal attention mechanism, for example, the three-mode feature maps are respectively passed through linear layers to generate query vectors Query (Q), key vectors Key (K) and value vectors Value (V). The formula is expressed as: ; wherein, is the feature map of the mth mode, , , is a trainable weight matrix.

[0025] Further, cross-attention calculation between different modes is performed, for example, through a query-key-value attention mechanism, the query vector of the vibration signal and the key vector of the thermal imaging data are calculated to obtain the spatio-temporal correlation weight, and the value vector is weighted and fused. The formula is expressed as: ; The query vector and the key vector are respectively from sensors with different sampling rates, and feature alignment is realized through a dynamic dimension projection layer.

[0026] In the above formula, the query vector represents the “question” features extracted from the vibration signal, representing the target that needs to be focused on at the moment (such as abnormal high-frequency vibration), for example, an abnormal frequency component of 12 kHz is detected in the vibration signal at a certain moment, and the Q vector encodes the spatio-temporal distribution of the frequency. The key vector represents the “index” features extracted from the thermal imaging data, representing the potential associated information that can be queried (such as temperature gradient change), for example, the coordinates (X, Y) of the hot spot area of the thread contact surface temperature field and the temperature rise rate (°C / s), and the K vector encodes these space-temperature associations. The value vector represents the “content” features extracted from the thermal imaging data, including the original information (such as temperature absolute value) that needs to be weighted and fused, for example, the original temperature value (unnormalized gray value) of the thermal imaging pixel at the same moment, and the V vector retains these original data for subsequent fusion.

[0027] Similarly, the attention interaction between other modes (thermal→strain, strain→vibration, etc.) is calculated. Finally, the attention outputs of each mode are spliced and passed through a fully connected layer to generate a fused feature vector.

[0028] Further, the space-time convolution network adopts a dynamic convolution kernel design. Since the traditional scheme adopts a fixed-size convolution kernel, it is difficult to adapt to the space-time scale difference of multi-modal data (such as high-frequency vibration requiring a small convolution kernel, and thermal imaging requiring a large kernel to capture spatial correlation). The embodiment introduces a dynamic convolution kernel generator to adjust the convolution kernel parameters in real time according to the characteristics of the input data, which is expressed by the formula: ; wherein, is a lightweight LSTM network, and the input is the current sensor state (such as sampling rate, signal-to-noise ratio) and historical context (such as feature distribution in the previous 10 seconds).

[0029] In step S3, as shown in Figure 3 , for example, deploy an edge computing unit, based on the PPO (Proximal Policy Optimization) algorithm principle, output the dynamic torque compensation value, and here introduce the multi-modal parameter-torque coupling compensation algorithm expression: ; wherein, is a temperature term parameter, is a vibration term parameter, is a cross-coupling term parameter.

[0030] The temperature term parameter can be further expanded as: , that is, ; wherein, represents the actual temperature measured by the sensor, which is used to reflect the actual temperature state of the current system, and is the direct input value of dynamic compensation; represents the reference temperature, which is usually the ambient temperature or the initial temperature of the system (such as the temperature at the beginning of tightening), and is used as the zero point of temperature compensation to quantify the deformation deviation caused by thermal expansion or contraction; represents the temperature change rate, which represents the influence of temperature instantaneous fluctuation on the system, and high-frequency temperature change may cause the accumulation of internal thermal stress of the material, which needs to be adjusted dynamically compensation amount; represents the temperature-torque coupling coefficient, which is related to the thermal expansion coefficient and stiffness of the material, and may change nonlinearly with temperature, and engineering calibration needs to be fitted through experiments, such as measuring the relationship between torque deviation and temperature difference in temperature step test; represents the temperature change rate sensitive coefficient, which represents the response speed of the system to the transient change of temperature, and the accumulation of internal thermal stress of the material needs to be dynamically adjusted compensation amount.

[0031] Further, when > , the material deforms due to thermal expansion (such as bolt elongation), which needs to be corrected torque output, and vice versa. As an ambient temperature benchmark, it can distinguish the effects of system self-heating and ambient temperature fluctuations. For example, if the ambient temperature drops suddenly (such as outdoor work in winter), the system needs to compensate for external interference first; the temperature change rate and For example, when the temperature rises rapidly (such as when the motor starts), the high temperature change rate needs to be Amplify the compensation amplitude to prevent overshoot.

[0032] Among the vibration item parameters, ;in, It represents the correction factor of deformation strain ε to vibration-temperature coupling. The specific form can be calibrated according to material property experiments. The basic vibration-temperature gain coefficient reflects the efficiency of the material's damping properties in converting vibration energy into heat energy and is related to the material's internal friction and microstructure; is the spectrum energy weight factor, which is the contribution ratio of the main frequency band vibration energy to the thermal effect, and is determined by extracting the frequency domain energy distribution through STFT (short-time Fourier transform); The nonlinear attenuation coefficient characterizes the attenuation characteristics of the high-frequency vibration energy transfer efficiency. It is related to the viscoelasticity of the material (for example, the λ value of rubber materials is relatively low). In experimental calibration, λ is often taken as 0.1 to 1.0 s² / m. Vibration acceleration is collected in real time by MEMS accelerometer or Laser Doppler Vibrometer (LDV). It needs to be combined with wavelet packet decomposition for denoising. High acceleration (such as |a|>10g) will cause a significant increase in frictional heat generation inside the material. . ;in, is the deformation-torque conversion coefficient, which is positively correlated with the yield strength of the material (e.g., η≈0.8 for steel and η≈0.6 for aluminum alloy); is the dynamic strain rate correction factor, which is similar to the dynamic adjustment strategy of the thermal deformation model mentioned above.

[0033] Among the cross-coupling parameters, the thermal-strain coupling coefficient ;in, is the temperature gradient and mechanical stress coefficient, reflecting the non-uniformity of thermal expansion of the material (such as the thermal elongation effect of the screw); is the dynamic change rate coupling factor, which characterizes the nonlinear superposition of temperature change rate and strain rate (such as high temperature rise leading to increased lattice distortion); the second-order derivative It is strain acceleration, used to warn of material fatigue or fracture risks (e.g., triggering shutdown protection when the value exceeds a threshold).

[0034] Further, the AI output is corrected by loading a material physics equation constraint (material yield strength formula) to prevent overfitting; for metal materials with obvious yield phenomenon (such as low carbon steel), a yield strength calculation formula is constructed as follows: ; Wherein, Re represents the yield strength (unit: MPa); Fe represents the force (N) that remains constant in the yield stage, that is, the minimum stable load when the material yields; represents the original cross-sectional area of the sample (mm²).

[0035] For metal materials without obvious yield phenomenon, the specified non-proportional elongation strength is used as the yield strength, which is represented as: ; Wherein, represents the force (N) corresponding to the 0.2% plastic deformation of the material; the meanings of other parameters are the same as above.

[0036] Temperature rise will cause the yield strength to decrease, and the formula is corrected as follows: ; Wherein, is the temperature correction coefficient.

[0037] In addition, the increase of strain rate will cause the yield strength to increase, which can be corrected by a dynamic hardening model.

[0038] In step S4, the model is trained by collecting experimental data, as shown in Figure 4 for example, the model parameters are adjusted and optimized according to the experimental results, and the parameter optimization results are verified in the digital twin platform. The optimized model transmits the compensation parameters to the dynamic compensator for compensation adjustment, realizes sub-microsecond level dynamic adjustment of the tightening operation, and improves the yield rate of the tightening operation.

[0039] In summary, the first embodiment of the present application proposes a dynamic torque compensation method for high-precision tightening operation, which collects multi-modal physical signals in the tightening operation process through a multi-modal sensing module, extracts cross-modal spatio-temporal features using a spatio-temporal convolution network, generates a high-dimensional state vector, and dynamically calculates a dynamic torque compensation value based on a hierarchical proximal policy optimization (PPO) algorithm, thereby realizing multi-modal information fusion, solving the spatio-temporal asynchronous problem of multi-modal sensor data, and improving the control accuracy of industrial tightening equipment and the real-time performance of edge computing compared with the method using a fixed threshold.

[0040] In addition, as Figure 5As shown, the second embodiment of the present invention further proposes a dynamic torque compensation device for high-precision tightening operations, for example, including: a multimodal signal acquisition module 201, a cross-modal feature fusion module 202, a dynamic torque compensation value calculation module 203 and a compensation adjustment module 204.

[0041] Among them, the multimodal signal acquisition module 201 is used to form a multimodal sensing module array to collect multimodal physical signals during the tightening operation; the cross-modal feature fusion module 202 is used to use a spatiotemporal convolutional network to extract the cross-modal spatiotemporal features of the multimodal physical signals, perform feature fusion on the cross-modal spatiotemporal features, and generate a fusion feature vector; the dynamic torque compensation value calculation module 203 is used to calculate the dynamic torque compensation value based on the hierarchical proximal strategy optimization algorithm according to the fusion feature vector; the compensation adjustment module 204 is used to compensate and adjust the tightening operation according to the dynamic torque compensation value.

[0042] The dynamic torque compensation method for high-precision tightening operations implemented by the dynamic torque compensation device for high-precision tightening operations disclosed in the second embodiment of the present invention is as described in the first embodiment above, and therefore will not be described in detail here. Optionally, the various modules and other operations or functions described above in the second embodiment are respectively for implementing the method described in the first embodiment. The beneficial effects of the dynamic torque compensation device for high-precision tightening operations provided by this embodiment are the same as the beneficial effects of the dynamic torque compensation method for high-precision tightening operations provided by the first embodiment above, and for the sake of brevity, they are not further described here.

[0043] like Figure 6 As shown, the third embodiment of the present invention further proposes an electronic device, for example, comprising: at least one processing unit, and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the method as described in the first embodiment, and the beneficial effects of the electronic device provided by this embodiment are the same as the beneficial effects of the dynamic torque compensation method for high-precision tightening operations provided by the first embodiment.

[0044] like Figure 7 As shown, the fourth embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above method when executed by a processor, and the beneficial effects of the computer-readable storage medium provided by this embodiment are the same as the beneficial effects of the dynamic torque compensation method for high-precision tightening operations provided by the first embodiment.

[0045] The computer readable storage medium can include any type of disk including floppy disks, optical disks, DVDs, CD-ROMs, micro-drives, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nano-systems (including molecular memory ICs), or any type of media or device suitable for storing instructions and / or data.

[0046] It should be noted that, for the foregoing method embodiments, the purposes of brief description, the foregoing methods are all described as a series of action combinations, but those skilled in the art should know that the present application is not limited to the order of the actions described, because according to the present application, some steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0047] In the above embodiments, the description of each embodiment is focused on, and the part not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0048] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic. The division of the units is only a logical function division. There can be another division manner in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some services interfaces, devices or units, and can be electrical or other forms.

[0049] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0050] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0051] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable memory. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned memory includes: a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0052] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be instructed by a program to be completed by relevant hardware, and the program can be stored in a computer readable memory, which can include a flash disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0053] The above is only exemplary embodiments of the present disclosure, and cannot limit the scope of the present disclosure. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure are still within the scope of the present disclosure. Those skilled in the art will easily think of embodiments of the present disclosure after considering the specification and practicing the disclosure herein. The present application is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or conventional technical means in the art that are not described in the present disclosure. The specification and examples are only considered as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

[0054] The technical features of the above embodiments can be combined in any way. To make the description concise, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not contradict, they should be considered within the scope of the present disclosure.

[0055] Those skilled in the art readily understand that the above only describes preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A dynamic torque compensation method for high-precision tightening operations, characterized in that: include: Establish a multimodal sensor module array to collect multimodal physical signals during the tightening operation; Extracting cross-modal spatiotemporal features of the multimodal physical signal using a spatiotemporal convolutional network, performing feature fusion on the cross-modal spatiotemporal features, and generating a fused feature vector; Calculating a dynamic torque compensation value based on a hierarchical proximal strategy optimization algorithm according to the fused feature vector; Compensation adjustment is performed on the tightening operation according to the dynamic torque compensation value.

2. The dynamic torque compensation method for high-precision tightening operation according to claim 1, characterized in that: The cross-modal spatiotemporal features of the multimodal physical signal are extracted using a spatiotemporal convolutional network, and feature fusion is performed on the cross-modal spatiotemporal features. Generate fused feature vector, including: The vibration signal, thermal imaging data, and strain data are independently input using a spatiotemporal separation convolution method. The three signals are independently input and preprocessed separately to obtain the spatiotemporal feature maps of each mode. Passing the spatiotemporal feature map through a linear layer to generate a trainable weight matrix for each modality; Calculating a similarity matrix between modalities based on the trainable weight matrix, and calculating attention interactions between the modalities according to the similarity matrix; The attention interaction outputs are concatenated and passed through a fully connected layer to generate the fused feature vector.

3. The dynamic torque compensation method for high-precision tightening operation according to claim 2, characterized in that: The calculation formula of the attention interaction is expressed as: ； in, represents the query vector, represents the key vector, represents a value vector, Represents the feature dimension of the key vector; The query vector of the vibration modality and the key vector of the thermal imaging modality are used to calculate the spatiotemporal association weight, and the value vectors are weightedly fused to obtain the attention interaction between the vibration modality and the thermal imaging modality.

4. The dynamic torque compensation method for high-precision tightening operation according to claim 1, characterized in that: The spatiotemporal convolution network includes a time dimension convolution kernel and a space / spectrum dimension convolution kernel. The time dimension convolution kernel slides along the time axis to extract dynamic evolution features; the space / spectrum dimension convolution kernel processes the spatial distribution of thermal imaging or the energy concentration area of ​​the vibration spectrum.

5. The dynamic torque compensation method for high-precision tightening operation according to claim 4, characterized in that: Also includes: A dynamic convolution kernel generator is introduced to adjust the convolution kernel parameters in real time according to the characteristics of the input data. The expression of the dynamic convolution kernel generator is: ； in, It is a lightweight LSTM network whose input is the current sensor state and historical context.

6. The dynamic torque compensation method for high-precision tightening operation according to claim 1, characterized in that: The expression for calculating the dynamic torque compensation value based on the hierarchical proximal strategy optimization algorithm according to the fused feature vector is: ； in, is the temperature parameter, is the vibration parameter, is the cross-coupling parameter; Among the temperature parameters, , represents the actual measured temperature, where Indicates the base reference temperature, represents the temperature-torque coupling coefficient, It represents the sensitivity coefficient of temperature change rate; Among the vibration item parameters, ,in, represents the correction factor of deformation strain ε to vibration-temperature coupling, is the basic vibration-temperature gain coefficient, is the spectrum energy weighting factor, is the nonlinear attenuation coefficient, is the vibration acceleration; ,in, is the deformation-torque conversion coefficient, is the dynamic strain rate correction factor; Among the cross-coupling parameters, ,in, is the temperature gradient and mechanical stress coefficient, is the dynamic rate of change coupling factor, the second-order derivative is the strain acceleration.

7. The dynamic torque compensation method for high-precision tightening operation according to claim 6, characterized in that: Also includes: For metal materials with obvious yield phenomenon, the yield strength calculation formula is constructed as follows: ； Where Re represents the yield strength, Fe represents the force that remains constant during the yield stage, represents the original cross-sectional area of ​​the specimen; For metal materials without obvious yield phenomenon, the specified non-proportional extension strength is used as the yield strength, which is expressed as: ； in, It indicates the force when the material produces 0.2% plastic deformation.

8. A dynamic torque compensation device for high-precision tightening operations, characterized in that: include: A multimodal signal acquisition module is used to form a multimodal sensor module array to collect multimodal physical signals during the tightening operation; A cross-modal feature fusion module is used to extract cross-modal spatiotemporal features of the multimodal physical signal using a spatiotemporal convolutional network, perform feature fusion on the cross-modal spatiotemporal features, and generate a fused feature vector; A dynamic torque compensation value calculation module, configured to calculate a dynamic torque compensation value based on the fused feature vector and a hierarchical proximal strategy optimization algorithm; A compensation adjustment module is used to perform compensation adjustment on the tightening operation according to the dynamic torque compensation value.

9. An electronic device, characterized in that: include: A memory and one or more processors connected to the memory, the memory storing a computer program, the processor being configured to execute the computer program to implement the dynamic torque compensation method for high-precision tightening operations according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable commands, and the computer-executable commands are used to execute the dynamic torque compensation method for high-precision tightening operations according to any one of claims 1 to 7.

Images