Equipment defect detection method based on infrared identification

Abstract

The invention relates to the field of intelligent monitoring, and discloses an infrared identification-based equipment defect detection method, which comprises the steps of infrared thermal imaging data acquisition, noise reduction processing, nonlinear heat conduction modeling, high-order partial differential optimization and variational regularization inversion solution. An infrared thermal imager collects equipment surface temperature distribution data, after multi-scale wavelet noise reduction and boundary heat flow balance optimization, a non-linear heat conduction model is established by utilizing a temperature-related heat conduction coefficient, a heat diffusion boundary is optimized by combining a high-order Laplace operator, defect heat source distribution is inverted through a variational regularization method, and the defect heat source distribution is calculated. And geometric characteristics, thermal parameters and position coordinates of the defects are accurately extracted. The method can adapt to complex working conditions, realizes high-precision detection and characteristic analysis of equipment defects, overcomes the problems of noise interference, insufficient boundary processing, weak defect inversion capability and the like in the prior art, and is widely applied to fault diagnosis and health monitoring of industrial equipment.

Description

Technical Field

[0001] The present invention relates to the field of intelligent monitoring, and in particular to an equipment defect detection method based on infrared recognition. Background Art

[0002] With the increasing complexity and diversification of industrial equipment, defect detection and fault monitoring during equipment operation have become one of the core technologies in the field of industrial operation and maintenance. Currently, common equipment defect detection technologies mainly include visual inspection, vibration analysis, acoustic inspection, and detection methods based on infrared thermal imaging. These technologies have certain effects in specific scenarios, but they generally have limitations in applicability and detection accuracy under complex working conditions.

[0003] In recent years, with the development of infrared thermal imaging technology, its application in equipment defect detection has gradually attracted attention. Infrared thermal imaging technology can achieve a certain degree of defect location by collecting the temperature distribution on the surface of the equipment and analyzing the abnormal area. However, the existing detection methods based on infrared thermal imaging also have many shortcomings, including the following points: First, infrared thermal imaging data is usually interfered by high-frequency noise, and the existing technology mostly uses a simple filtering algorithm, which is difficult to achieve high-precision signal-to-noise ratio improvement. Secondly, the existing detection methods generally model the temperature distribution inside the equipment based on the linear heat conduction model, and fail to fully consider the temperature dependence and dynamic thermal diffusion characteristics of the material, resulting in insufficient description of the thermal diffusion behavior under complex working conditions. In addition, the boundary area data causes heat flow imbalance due to the complex geometric shape of the equipment surface. The existing methods lack effective boundary optimization technology, which affects the accuracy of the overall model. Finally, in terms of defect area inversion, the existing technology usually uses a simple threshold method to perform regional segmentation or rough fitting, which is difficult to accurately extract the geometric and thermal characteristics of the defects, especially in multi-defect scenarios or complex distribution situations. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides an equipment defect detection method based on infrared recognition, which solves the problems that infrared thermal imaging data is difficult to achieve high-precision signal-to-noise ratio improvement, has insufficient ability to describe thermal diffusion behavior under complex working conditions, and affects the accuracy of the overall model.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a device defect detection method based on infrared recognition comprises the following steps: Collect infrared thermal imaging data of the equipment surface in operation and obtain infrared thermal distribution sequence that changes with time; Perform noise reduction on infrared thermal imaging data to eliminate high-frequency noise, retain effective signals, and optimize the boundary area using heat flow balance conditions; Construct a nonlinear heat conduction model to describe the dynamic changes in temperature distribution inside the equipment; Optimize and analyze the thermal diffusion characteristics of the defect area based on the high-order partial differential expansion model; The variational regularization method is used to optimize and solve the heat conduction model and invert the heat source distribution in the defect area; According to the heat source distribution in the defect area, the defect position inside the equipment is located, and the geometric shape and thermal characteristics of the defect are extracted.

[0006] Preferably, the step of collecting infrared thermal imaging data includes: Use a high-resolution infrared thermal imager to obtain the temperature distribution on the equipment surface; Collect multiple frames of thermal imaging data at a fixed sampling frequency in the time dimension to form a thermal distribution time series; Record the equipment's operating parameters, including ambient temperature, load status, and operating time.

[0007] Preferably, the noise reduction processing step includes: Perform two-dimensional wavelet decomposition on the collected thermal imaging data to separate the data into high-frequency and low-frequency components; The soft threshold filtering method is used to reduce the noise of high-frequency components and filter random noise signals; The effective signal is retained for the low-frequency component, and the processed thermal imaging data is reconstructed.

[0008] Preferably, the optimization of the boundary region using heat flow balance conditions includes: For the complex geometric boundaries of the device surface, the boundary conditions are set as the heat flow balance conditions in the normal direction of the device surface; Neumann boundary conditions based on the device surface shape are used to ensure that the heat flux on the device surface does not flow out of the boundary.

[0009] Preferably, the nonlinear heat conduction model is: According to the thermal conductivity of the equipment material, a nonlinear heat conduction model is established to describe the dynamic behavior of temperature distribution changing with time; In the model, the thermal conductivity is temperature dependent and the heat diffusion behavior is regulated by the heat source characteristics of the device defect area.

[0010] Preferably, the high-order partial differential expansion model includes: The fourth-order partial differential term is introduced into the heat conduction model, and the heat diffusion characteristics of the defect region boundary are described by the high-order form of the Laplace operator. By introducing high-order terms, the thermal boundary clarity of the defect area is enhanced and the noise interference in the boundary area is reduced.

[0011] Preferably, the variational regularization method comprises: Define the objective function based on the residual of the heat conduction equation, and balance the fitting accuracy of the heat distribution and the data smoothness through the regularization term; The regularization term includes a second-order partial differential regularization term and a fourth-order partial differential regularization term; The objective function is solved by numerical optimization method, and the temperature distribution on the surface and inside of the equipment is optimized.

[0012] Preferably, the step of inverting the heat source distribution of the defect area includes: Using the optimized temperature distribution, calculate the heat source intensity distribution inside the equipment; According to the spatial characteristics of the heat source distribution, the location of the defect area is identified, and the geometric characteristics and thermal intensity of the defect are extracted.

[0013] Preferably, the defect detection result obtained by inverting the heat source distribution of the defect area is used to generate a three-dimensional thermal distribution map of the defect area, and is further used to provide a report on the defect location, geometry and thermal characteristics for reference in equipment operation and maintenance decision-making.

[0014] Equipment defect detection system based on infrared recognition, including: Infrared thermal imager, used to collect temperature distribution on the surface of equipment; A data processing module is used to reduce noise and optimize the boundaries of the collected infrared thermal imaging data; Model building module for building temperature distribution models based on nonlinear heat conduction equations and high-order partial differential extension models; Optimization and inversion module, which is used to optimize the temperature distribution using variational regularization method and invert the defect heat source; Output module, used to generate defect detection results, including geometric characteristics, heat intensity distribution and location coordinates of defects.

[0015] The present invention provides a device defect detection method based on infrared recognition. It has the following beneficial effects: 1. The present invention adopts an equipment defect detection method based on infrared recognition, and achieves the technical effect of accurately locating equipment defects and extracting geometric characteristics through infrared thermal imaging data acquisition, data noise reduction processing, nonlinear heat conduction model construction, high-order partial differential expansion model optimization and variational regularization method optimization solution. Compared with the methods in the prior art that rely on visual detection, vibration analysis or acoustic detection, the problems of low detection accuracy and poor adaptability of these methods under complex working conditions (such as high temperature, dynamic operation or noise interference environment) are solved.

[0016] 2. The present invention combines the nonlinear heat conduction model with the high-order partial differential expansion model, and uses the temperature-dependent thermal conductivity and the fourth-order Laplace operator to enhance the description of the heat diffusion characteristics, achieving the technical effect of accurately simulating the dynamic changes of the temperature distribution inside the equipment and clarifying the boundary of the defect area. Compared with the detection scheme based on the linear heat conduction model in the prior art, it solves the problem of being unable to characterize complex heat diffusion behavior and fuzzy defect boundaries.

[0017] 3. The present invention uses the variational regularization method to solve the optimized heat conduction model, invert the heat source distribution of the defect area, extract the geometric characteristics and thermal characteristics of the defect, and achieve the technical effect of high-precision inversion of the defect heat source and suppression of noise interference. Compared with the direct processing solution that simply relies on infrared data in the prior art, it solves the problem of unstable defect inversion caused by noise and pathological models.

[0018] 4. The present invention organically combines data processing, model building, optimization solution and result output through modular system design, generates complete detection results including defect location, geometric characteristics and thermal intensity distribution, and achieves the technical effect of efficient, comprehensive and automatic defect detection. Compared with the decentralized detection process in the prior art, it solves the problems of low work efficiency, poor detection consistency and unintuitive results. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a step diagram of the device defect detection method based on infrared recognition of the present invention; Figure 2 This is a structural diagram of the equipment defect detection system based on infrared recognition of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Please see attached Figure 1 , an embodiment of the present invention provides a device defect detection method based on infrared recognition, comprising the following steps: S1. Collect infrared thermal imaging data of the equipment surface in operation to obtain infrared thermal distribution sequence that changes with time.

[0022] Specifically, step S1 continuously monitors the infrared heat distribution during the dynamic operation of the equipment to form a heat distribution sequence that changes over time, providing basic data support for subsequent modeling analysis and defect location. This step is closely related to the subsequent noise reduction processing, nonlinear heat conduction model construction and optimization calculation, and is the basic link of the entire method, which directly determines the accuracy and reliability of subsequent detection.

[0023] In this embodiment, a high-resolution infrared thermal imager is used to collect the thermal distribution data on the equipment surface, which specifically includes the following contents: When the equipment is running, the temperature distribution at different spatial locations on the equipment surface is recorded in real time by a high-resolution infrared thermal imager. The resolution of the infrared thermal imager is set to 640×480 pixels and the frame rate is set to 30fps to ensure that the rapidly changing thermal distribution during the operation of the equipment can be captured.

[0024] The operating status of the equipment includes a variety of working conditions, such as operation under different loads, changes in external ambient temperature, dynamic processes such as equipment acceleration and deceleration. During acquisition, infrared thermal imaging data needs to be recorded for each of these working conditions in order to subsequently analyze the thermal distribution characteristics of the equipment under different working conditions.

[0025] The data collected by the infrared thermal imager is stored to form time series data ,in: Indicated in The spatial position of the device surface under the frame degree distribution, is the sampling frame number. Through the continuous time series, the dynamic change of the surface temperature of the equipment over time can be reflected.

[0026] Infrared thermal imaging time series data Every frame Indicates the time on the device surface The temperature distribution at the moment. The time-continuous heat distribution can be expressed as a function: in: , and this heat distribution sequence is used as the initial boundary condition input in subsequent modeling.

[0027] Generally, in order to reduce measurement errors, it is necessary to calibrate the environmental parameters of the infrared thermal imager, including background temperature, device surface emissivity, and the thermal imager calibration factor. The calibrated infrared thermal imager can output accurate surface temperature data, providing a reliable data basis for subsequent processing.

[0028] For the calibration of infrared thermal imagers, the emissivity correction formula is: in: is the corrected true temperature, is the measured radiation temperature, is the emissivity of the device surface, is the ambient background temperature.

[0029] Through the above formula correction and multi-angle acquisition strategy, the accuracy and comprehensiveness of the data can be effectively improved, laying the foundation for subsequent processing.

[0030] As an option, the present invention can also compensate the infrared thermal imaging data in combination with the environmental information of the equipment operation, such as external environmental temperature, humidity and airflow velocity. Specifically, when the external environmental temperature fluctuates significantly, the accuracy of the thermal distribution data can be further improved by collecting temperature data of multiple reference areas as a compensation benchmark for environmental parameters.

[0031] In some embodiments, in order to capture the possible small temperature changes in specific operating parts of the equipment (such as gearboxes, bearings, etc.), the present invention adopts a local high-resolution sampling strategy. In this case, the sampling area of ​​the infrared thermal imager can be limited to specific key parts, further improving the spatial resolution of the data.

[0032] In order to ensure the time continuity of the heat distribution sequence, the present invention also sets a time sampling step. Specifically, the time step The choice of should be determined according to the dynamic characteristics of the equipment operation. For example, for rapidly changing equipment components, the time step can be set to 0.03 seconds to avoid missing key temperature changes.

[0033] In one possible implementation, the present invention collects equipment surface thermal distribution data while also synchronously collecting equipment operating parameters (such as rotation speed, vibration signals, etc.), and synchronizes these parameters with the thermal distribution sequence to provide support for multi-parameter coupling analysis in subsequent defect modeling.

[0034] In the above data collection process, the effect of the geometric shape of the device surface on infrared imaging can also be considered. Generally speaking, for device surfaces with curved surfaces or complex geometric structures, a multi-angle collection strategy can be used to stitch the thermal distribution data from multiple perspectives into a unified surface thermal distribution to achieve more comprehensive detection.

[0035] Through the above steps, the present invention can obtain the thermal distribution sequence of the equipment surface under different operating conditions, providing a reliable data basis for subsequent noise reduction processing, modeling optimization and defect location analysis.

[0036] S2. Perform noise reduction processing on infrared thermal imaging data to eliminate high-frequency noise, retain effective signals, and optimize the boundary area using heat flow balance conditions.

[0037] Specifically, in step S2, in order to improve the reliability and accuracy of thermal imaging data during equipment defect detection, the present invention further performs noise reduction processing on the basis of the collected infrared thermal imaging data to eliminate the interference of high-frequency noise signals while retaining effective low-frequency signals to accurately reflect the true thermal distribution characteristics of the equipment during operation. In addition, since the surface of the equipment usually has a complex geometric boundary shape, it may cause distortion or imbalance of the temperature data in the boundary area. Therefore, on the basis of noise reduction processing, the present invention introduces heat flow balance condition optimization to the boundary area to ensure the smoothness and physical consistency of the data in space. This step is directly related to the subsequent modeling step, and its result directly affects the accuracy of the heat conduction model and the reliability of the subsequent defect area inversion.

[0038] In this embodiment, for the denoising of infrared thermal imaging data, a multi-scale wavelet transform method is used to denoise the collected thermal imaging sequence data. , which contains both effective temperature signals and random noise signals generated by the operation of the equipment. The high-frequency component mainly manifests as noise interference, while the low-frequency component reflects the temperature distribution characteristics of the equipment. Therefore, by effectively suppressing the high-frequency component and retaining the low-frequency component, data noise reduction can be achieved.

[0039] Specifically, firstly, each frame of thermal imaging data Perform two-dimensional wavelet decomposition and use wavelet basis functions Perform multi-scale decomposition on the data. The decomposition result can be expressed as: in: represents the scale of decomposition, Indicates direction, is the decomposed wavelet coefficient, and are the total number of decomposition scales and directions respectively. High-frequency noise is mainly concentrated in smaller scales The low-frequency information is retained in the coefficients of larger scales.

[0040] After the decomposition is completed, the wavelet coefficients of the high-frequency components The soft threshold filtering method is used to suppress high-frequency noise. The wavelet coefficients after filtering The calculation is done by the following formula: in: is the threshold parameter set to control the intensity of the filter. Less than or equal to , then the point is considered as noise and set to zero; if it exceeds the threshold, the value will be adjusted according to the soft threshold rule. The choice of threshold can be adjusted according to the device operating conditions and noise level. In one possible implementation, the threshold It can be selected by cross-validation method to balance the noise suppression effect and the retention of valid signals.

[0041] After filtering, the coefficients of the low-frequency components are kept unchanged, and the wavelet coefficients after filtering are combined , perform two-dimensional inverse wavelet transform on the data to reconstruct the denoised thermal imaging data The reconstruction formula is: As an option, to further improve the noise reduction effect, the time series data can be Perform three-dimensional wavelet transform. By incorporating the time dimension into the wavelet decomposition, time-related noise can be better suppressed, especially temperature fluctuations with strong randomness.

[0042] After completing data denoising, the present invention performs heat flow balance optimization processing on the boundary area of ​​the device surface. Generally, due to the complex geometric shape of the device surface, the infrared thermal imaging data may have abnormal gradients or unevenness in the boundary area. To this end, the present invention corrects the data by introducing the Neumann boundary condition: in: Represents the boundaries of the device surface, The normal vector represents the boundary. The introduction of boundary conditions can ensure the heat flow balance in the boundary area and avoid abnormal data that does not conform to physical laws from affecting subsequent modeling and analysis.

[0043] In a possible implementation, the de-noised thermal imaging data may be subjected to spatial interpolation processing. For areas with missing data or local anomalies, data may be supplemented by an interpolation method (eg, bicubic interpolation) to further improve data continuity.

[0044] Through the above steps, the present invention can obtain high-quality thermal imaging data after noise reduction, providing a reliable data input basis for the subsequent nonlinear heat conduction model construction and defect area inversion. Noise reduction processing can not only effectively eliminate high-frequency noise interference, but also ensure the physical consistency of data through boundary optimization, thereby significantly improving the accuracy and robustness of subsequent detection steps.

[0045] S3. Construct a nonlinear heat conduction model to describe the dynamic changes of temperature distribution inside the equipment.

[0046] Specifically, in order to accurately describe the dynamic changes of the temperature distribution inside the equipment with time and space, and further provide basic data support for the inversion of the defect area, the present invention constructs a physical modeling method based on the nonlinear heat conduction equation after completing data denoising and boundary optimization. This model can accurately simulate the heat flow transfer process inside the equipment by describing the thermal diffusion behavior inside the equipment, combining the temperature-related thermal conductivity characteristics and heat source distribution law. The construction of the model directly depends on the infrared thermal imaging data after the aforementioned denoising process, and at the same time provides theoretical support for subsequent high-order model expansion and defect inversion.

[0047] In this embodiment, the construction of the nonlinear heat conduction model is based on Fourier heat conduction theory, and specifically adopts the following physical and mathematical expressions.

[0048] In general, heat conduction behavior can be described by the following partial differential equation: in: Indicates the time inside the device and spatial location Temperature distribution; is the thermal conductivity, which indicates the heat conduction capacity of the material; is the heat source distribution function, which represents the heat generated or absorbed inside the equipment; Represents the heat diffusion term, which is used to characterize the spatial distribution characteristics of heat; Indicates the rate of change of temperature over time.

[0049] As an improvement, in order to be closer to the complex heat conduction characteristics in actual equipment operation, the present invention adopts a temperature-dependent nonlinear thermal conductivity model: in: is the thermal conductivity of the material at standard temperature; It is a nonlinear correction parameter used to reflect the dependence of thermal conductivity on temperature.

[0050] By Substituting for the dynamic thermal conductivity, the nonlinear heat conduction equation can be rewritten as: This equation can more accurately describe the changes in the thermal conductivity of materials at different temperature levels and is particularly suitable for thermal diffusion behavior in high-temperature equipment or dynamic conditions.

[0051] Specifically, in this embodiment, the infrared thermal imaging data after noise reduction is combined And boundary optimization conditions, the initial boundary conditions and initial temperature field of the model are set as follows: Initial temperature field Setting is performed based on the initial data in the infrared heat distribution sequence after noise reduction; The boundary conditions adopt Neumann boundary conditions: in: Represents the boundaries of the device surface, Represents the normal vector of the boundary.

[0052] In a possible implementation, in order to improve the stability of modeling and the accuracy of numerical solution, the present invention discretizes the above nonlinear heat conduction equation. Generally, the time dimension is discretized by implicit difference method, and the space dimension is discretized by finite element method. The discretized equation is expressed as: in: and Represent the temperature distribution vectors of the current time step and the next time step respectively; is the time step; represents the nonlinear thermal conductivity matrix, and the temperature distribution Dynamic correlation; Represents the heat source distribution vector.

[0053] As an option, the present invention uses the Newton-Raphson method to iteratively solve the discretized nonlinear equation. The Newton-Raphson method converts the nonlinear equation into a solution of a multi-order linear equation system through linear approximation, thereby improving the numerical convergence speed.

[0054] In some embodiments, in order to further meet the modeling requirements under complex working conditions, the present invention also provides a heat source distribution function Specifically, the dynamic change law of the heat source term is set according to the equipment operating parameters (such as speed and vibration amplitude), thereby enhancing the adaptability of the model to the dynamic operating environment.

[0055] By constructing the above-mentioned nonlinear heat conduction model, the present invention can achieve an accurate description of the temperature distribution inside the device, and provide a physical basis for the expansion of high-order models and defect area inversion in subsequent steps. This model is particularly suitable for equipment detection scenarios with complex heat conduction characteristics, such as defect detection in high temperature or dynamic load environments. The accuracy of the model is determined by the temperature-dependent thermal conductivity and the numerical discretization method. At the same time, the physical consistency and robustness of the results are ensured by the reasonable setting of boundary conditions and initial temperature fields.

[0056] In a possible extended application, this model can also be combined with multimodal data (such as vibration signals or acoustic signals) for joint modeling to further improve the comprehensive accuracy of defect detection. Through these improvements, the model's adaptability and detection capabilities have been significantly improved, providing strong support for equipment maintenance and operation monitoring.

[0057] S4. Optimize and analyze the thermal diffusion characteristics of the defect area based on the high-order partial differential expansion model.

[0058] Specifically, after completing the construction of the nonlinear heat conduction model in step S4, in order to further optimize the analysis accuracy of the defect area of ​​the equipment, the present invention proposes a method based on a high-order partial differential expansion model to enhance and optimize the thermal diffusion characteristics of the defect area. By introducing a high-order Laplace operator into the traditional heat conduction model, it is possible to more accurately describe the changing trend of the thermal diffusion boundary, while smoothing the impact of noise interference on the results, thereby improving the spatial resolution and analysis accuracy of defect location. This step is expanded on the basis of the nonlinear heat conduction model, and is closely connected with the previous steps in terms of model data and physical mechanism, while providing an optimized basis for the subsequent defect heat source inversion.

[0059] In this embodiment, the high-order partial differential expansion model is constructed by introducing a fourth-order Laplace operator into the nonlinear heat conduction equation. The specific equation form is as follows: in: is the temperature distribution; It is a traditional second-order heat diffusion term, which describes the spatial distribution and thermal conductivity of heat; The newly added fourth-order partial differential term is used to enhance the clarity of the thermal diffusion boundary and the noise suppression capability; is the heat source distribution function; It is the high-order smoothing coefficient, which is used to adjust the strength of the fourth-order Laplace term.

[0060] In general, the traditional heat conduction model has certain limitations in the boundary clarity of the heat diffusion region, especially under high noise or complex boundary conditions, which may cause the boundary of the heat diffusion region to appear blurred or discontinuous. , which can effectively enhance the smoothness of the boundary and adjust the coefficient Control the optimization intensity to more accurately identify the boundary characteristics of defects.

[0061] Specifically, in this embodiment, for the numerical solution of the above-mentioned high-order partial differential expansion model, the finite element method is used to discretize the spatial dimension, and the implicit difference method is used to discretize the time dimension. The discretized model equation is expressed as: in: is the mass matrix, representing the spatial integration coefficient; and Represent the temperature distribution vectors of the current time step and the next time step respectively; is the second-order Laplace operator discrete matrix, used to describe the second-order heat diffusion term; is a fourth-order Laplace operator discrete matrix, and high-order boundary enhancement is achieved by squaring the second-order operator; is the time step, which is used to control the numerical stability in the time direction; is the heat source distribution vector.

[0062] As a possible implementation method, in order to ensure the stability and calculation accuracy of the numerical solution, the present invention adopts a hybrid finite element method to solve the fourth-order partial differential term Specifically, by introducing auxiliary variables , the high-order equation is decomposed into two second-order equations and solved jointly, thereby reducing the computational complexity and improving the convergence performance.

[0063] In some embodiments, in order to further enhance the optimization capability of the high-order model for boundary characteristics, the present invention also combines a multi-scale boundary decomposition method. Specifically, by performing multi-scale decomposition on the boundary characteristics of the thermal diffusion region, the boundary region is divided into sub-domains of different scales, and different high-order partial differential term weights are introduced into each sub-domain, thereby achieving accurate optimization of complex boundary regions.

[0064] In general, the performance of high-order partial differential expansion models depends directly on the parameters and As an option, the present invention can automatically adjust the and The value of is selected to ensure that the model can obtain the best boundary clarity and noise suppression effect under different equipment and working conditions.

[0065] In this embodiment, in order to verify the effectiveness of the high-order partial differential expansion model, the fourth-order Laplace term is found through simulation analysis of real equipment operation data. It can significantly improve the smoothness of the heat diffusion boundary and effectively reduce the pseudo boundary signal caused by noise. For example, in the thermal imaging data of a certain equipment gearbox, the addition of high-order terms can clearly identify the heat distribution boundary of the gear contact area, while the traditional second-order model produces a more obvious blurring phenomenon in this area.

[0066] The high-order partial differential expansion model of the present invention can significantly improve the accuracy of defect area analysis in complex equipment and operating environments, providing a more optimized input basis for subsequent defect heat source inversion. This model is particularly suitable for equipment defect detection scenarios with complex boundaries, strong noise interference, or drastic changes in thermal diffusion gradients, such as operation monitoring of large rotating machinery, high-temperature furnace equipment, or precision machining equipment. The applicability and versatility of the model have been fully demonstrated through a variety of test verifications.

[0067] S5. Use the variational regularization method to optimize and solve the heat conduction model and invert the heat source distribution in the defect area.

[0068] Specifically, after the high-order partial differential extended model is constructed, in step S5, in order to further improve the solution accuracy and stability of the heat conduction model and realize the accurate inversion of the heat source distribution in the defect area, the present invention introduces a variational regularization method to optimize the model solution process. Variational regularization can effectively deal with ill-posed problems caused by boundary noise, incomplete data or complexity of thermal diffusion, and achieve a reasonable balance between model fitting and data smoothing. This step is based on the solution results of the nonlinear heat conduction model and the high-order partial differential model, further optimizing the calculation accuracy of the heat source distribution, laying the foundation for the subsequent positioning and feature extraction of the defect area.

[0069] In this embodiment, the variational regularization method uses the residual of the heat conduction model as the optimization target, and introduces the regularization term to enhance the smoothness and boundary clarity of the heat diffusion region, while suppressing the influence of noise on the solution process. The definition of the target optimization function is as follows: in: To optimize the objective function; the first term represents the residual of the heat conduction equation; the second term is the second-order regularization term, which is used to enhance the boundary characteristics of the thermal field; the third term is a fourth-order regularization term used to smooth the noise in the heat diffusion area; and is the regularization weight parameter, which controls the strength of the regularization term.

[0070] Specifically, in this embodiment, the solution of the objective optimization function is iteratively calculated using the conjugate gradient method. In each iteration, the gradient of the objective function is calculated and the temperature distribution is updated. , gradually approaching the optimal solution. The gradient calculation formula is: in: is the heat diffusion term; is the heat source distribution function; and are the gradients of the second-order and fourth-order regularization terms, respectively.

[0071] In general, the initial value of the conjugate gradient method is Take it as the solution of the high-order partial differential model. The iteration stop condition can be set as the Euclidean norm of the gradient is less than a certain threshold ,For example: As an option, this embodiment can also combine a multi-resolution variational method to solve the target optimization function. Specifically, by performing regularization solutions on spatial grids of different resolutions, the computational complexity can be effectively reduced while improving the ability to describe the details of the defective area. For example, after performing a preliminary solution on a low-resolution grid, the result is used as the initial value of the high-resolution grid to gradually improve the sophistication of the solution.

[0072] In some embodiments, in order to further enhance the robustness of the solution, the present invention combines a heat source dynamic update strategy. Specifically, based on the results of the previous iteration, the heat source distribution To make dynamic adjustments: in: is the update step size, which is used to control the rate at which the heat source is adjusted.

[0073] Through the above optimization solution process, the present invention can accurately invert the heat source distribution in the defect area inside the equipment. The intensity and spatial distribution of heat can directly reflect the location and characteristics of internal defects in the equipment. For example, for surface crack defects, the heat source distribution usually appears as linear local heat concentration, while for wear or fatigue defects, it may appear as scattered point or surface heat concentration.

[0074] As a possible extended application, the present invention can also further extract the geometric characteristics and thermal characteristics of defects by clustering the heat source distribution. For example, the K-Means clustering method is used to divide the heat source distribution into different sub-regions, and the specific form of the defect is analyzed based on the shape parameters of the sub-regions (such as area, length and direction). In addition, for dynamic defect areas, time series analysis methods (such as LSTM models) can be combined to predict the defect expansion trend and provide early warning for equipment maintenance.

[0075] By introducing the variational regularization method, the present invention can significantly improve the accuracy and stability of heat source inversion in the defect area. Compared with the traditional heat conduction model solution method, the present invention can not only better handle complex working conditions and noise interference, but also extract more defect characteristic information by optimizing the solution process, thereby providing strong support for accurate detection and predictive maintenance of equipment defects. The integrity of the model solution process and the accuracy of the calculation results have been fully demonstrated through the verification of multiple test scenarios.

[0076] S6. According to the heat source distribution in the defect area, locate the defect position inside the equipment and extract the geometric shape and thermal characteristics of the defect.

[0077] Specifically, after completing the optimization solution of the heat conduction model and the inversion of the heat source distribution in the defect area, the present invention further performs the inversion of the heat source distribution obtained. , accurately locate the specific location of defects inside the equipment, and extract the geometric shape and thermal characteristics of the defects in combination with the spatial distribution characteristics of the heat source. Through these analysis results, the defect characteristics inside the equipment can be fully characterized, providing key data support for subsequent equipment health assessment, maintenance and fault prediction.

[0078] This step directly relies on the defect heat source distribution data obtained by inversion in the previous step. The heat source distribution not only reflects the spatial location of the defect, but also contains relevant information about the thermal characteristics of the defect, such as heat intensity, temperature rise rate, and distribution pattern. Therefore, this step extracts multi-dimensional defect characteristic information from the heat source distribution through a multi-level data analysis method.

[0079] In this embodiment, the defect location and feature extraction process includes the following: The invention obtains the heat source distribution according to the inversion , the location of the defect is determined by judging the local high-value area of ​​the heat source. Generally, the high-value area of ​​the heat source distribution corresponds to the core part of the defect or the heat concentration area. In order to achieve accurate positioning, the present invention adopts the following positioning criteria: In the spatial heat source distribution, select the points that meet the following conditions as the defect core points: in: is the threshold coefficient, usually in the range of , which is used to screen out heat source areas that are significantly higher than the surrounding background.

[0080] For the case of multiple defects with complex distribution, the present invention further combines spatial clustering algorithms (such as K-Means clustering) to divide the high-value heat source area into multiple sub-areas. Each sub-area corresponds to the core position of a defect, and its centroid position is calculated by Determine the specific spatial coordinates.

[0081] As a possible implementation method, the present invention can also combine the time variation characteristics of multiple frames of thermal imaging data during the positioning process. For example, by calculating the time series variation rate of the heat source distribution of each frame , which can further improve the dynamic tracking capability of the defect core position.

[0082] After the defect position is located, the present invention extracts the geometric shape information of the defect based on the spatial morphology of the heat source distribution. In general, the geometric shape of the defect can be described by the following formula: in: Represents the spatial region of the defect.

[0083] Through the region The geometric parameters are calculated and the following features can be extracted: Volume: The spatial volume of the defect area , calculated by the integral formula: Boundary shape: The three-dimensional boundary shape of the defect area , extracted by boundary tracing algorithm; Maximum size: The maximum size of the defect area in each axial direction , , , calculated by the coordinate range difference: As an option, for more complex defect morphologies (such as cracks or planar defects), the present invention combines the principal component analysis (PCA) method to extract the main direction and shape characteristics of the defect. For example, by performing principal component decomposition on the spatial coordinate matrix of the defect area, the main axis direction and aspect ratio of the defect can be obtained.

[0084] The present invention also extracts the thermal characteristics of defects based on the heat source distribution data, which mainly include the following contents: Thermal intensity: Total thermal intensity of the defect area , calculated by integrating the heat source distribution function: Thermal gradient: Thermal gradient within the defect area , calculated by numerical differentiation; Temperature rise rate: Combine multiple frames of thermal imaging data to calculate the time change rate of the defect area , used to determine the dynamic expansion trend of defects.

[0085] Generally, the heat intensity can be used to measure the severity of defects, while thermal gradients This reflects the unevenness of heat diffusion in the defect area. For the dynamic defect area, the temperature rise rate It can be used as an important indicator for early fault warning.

[0086] In order to facilitate the understanding of the results and subsequent decision support, the present invention visualizes the defect location results, geometric characteristics and thermal characteristics. Specifically, by constructing a three-dimensional thermal distribution map and a defect spatial contour map, the spatial position, shape and thermal distribution characteristics of the defect can be intuitively presented.

[0087] In some embodiments, the present invention also outputs the analysis results as a structured report through a data export function. The report content includes key parameters such as the defect's location coordinates, volume, heat intensity, and expansion trend, providing a scientific basis for equipment maintenance personnel.

[0088] Through the implementation of this step, the present invention can accurately locate the defects inside the equipment from the heat source distribution data, and extract the geometric shape and thermal characteristics of the defects. This information not only provides an important reference for equipment failure analysis and maintenance, but also lays a data foundation for predictive maintenance and operating status monitoring. The positioning accuracy and feature extraction capabilities of the model have been verified through actual tests of a variety of industrial equipment, and can effectively meet the needs of equipment defect detection under complex working conditions.

[0089] Please see attached Figure 2 The present invention also provides an equipment defect detection system based on infrared recognition, comprising: Infrared thermal imager, used to collect temperature distribution on the surface of equipment; A data processing module is used to reduce noise and optimize the boundaries of the collected infrared thermal imaging data; Model building module for building temperature distribution models based on nonlinear heat conduction equations and high-order partial differential extension models; Optimization and inversion module, which is used to optimize the temperature distribution using variational regularization method and invert the defect heat source; Output module, used to generate defect detection results, including geometric characteristics, heat intensity distribution and location coordinates of defects.

[0090] Specifically, infrared thermal imager: Infrared thermal imagers are used to collect the temperature distribution of the equipment surface in real time when it is in operation. They can obtain high-resolution infrared thermal imaging data and generate dynamic temperature distribution maps in time series.

[0091] Generally, infrared thermal imagers support multi-frame sampling and dynamic adjustment of sampling frequency to adapt to the operating conditions and environmental noise of different equipment. As an option, infrared thermal imagers can also collect local high-resolution temperature distribution in combination with the monitoring needs of specific areas of specific parts of the equipment.

[0092] Data processing module: The data processing module is responsible for denoising and boundary optimization of the temperature distribution data collected by the infrared thermal imager to ensure the physical consistency and calculation accuracy of the input data.

[0093] Specifically, the data processing module can remove high-frequency noise from infrared data and optimize the heat flow balance in the boundary area of ​​the device surface to avoid data errors caused by complex geometric boundaries. As a possible implementation method, the module supports multi-dimensional noise reduction processing of time series data to improve the temporal continuity of heat distribution.

[0094] Model building modules: The model building module is used to establish a temperature distribution model of the device based on the nonlinear heat conduction equation and the high-order partial differential expansion model. The module can combine the initial temperature data provided by the infrared thermal imager and use the physical properties of the device (such as thermal conductivity and thermal diffusion characteristics) to accurately describe the dynamic temperature changes inside the device.

[0095] In some embodiments, the model building module also supports adaptive adjustment of dynamic heat conduction parameters (such as temperature-related thermal conductivity), thereby improving the adaptability of the model under complex operating conditions.

[0096] Optimization and Inversion Module: The optimization and inversion module optimizes and solves the temperature distribution model based on the variational regularization method, and uses the optimized results to invert the defect heat source distribution inside the equipment.

[0097] The module can efficiently handle ill-conditioned problems in the heat conduction model, enhance the smoothness and boundary clarity of the heat diffusion area through regularization, and thus improve the accuracy of defect area inversion. In addition, the module supports dynamic adjustment of defect heat source distribution to adapt to changes in equipment operating conditions.

[0098] Output Module: The output module is responsible for presenting the analysis results in a visual form and generating a defect detection report. The output includes the geometric characteristics of the defect (such as volume, shape, and boundary), the heat intensity distribution (such as total heat intensity and gradient), and the spatial location coordinates of the defect.

[0099] In some embodiments, the output module supports the generation of a three-dimensional thermal distribution map to intuitively display the spatial distribution of defects. At the same time, the module can export the test results in a structured data format, which is convenient for docking with the equipment monitoring system to achieve full life cycle equipment status management.

[0100] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The device defect detection method based on infrared recognition is characterized in that: The following steps are involved: Collect infrared thermal imaging data of the equipment surface in operation and obtain infrared thermal distribution sequence that changes with time; Perform noise reduction on infrared thermal imaging data to eliminate high-frequency noise, retain effective signals, and optimize the boundary area using heat flow balance conditions; Construct a nonlinear heat conduction model to describe the dynamic changes in temperature distribution inside the equipment; Optimize and analyze the thermal diffusion characteristics of the defect area based on the high-order partial differential expansion model; The variational regularization method is used to optimize and solve the heat conduction model and invert the heat source distribution in the defect area; According to the heat source distribution in the defect area, the defect position inside the equipment is located, and the defect geometry and thermal characteristics are extracted.

2. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The infrared thermal imaging data collection step comprises: Use a high-resolution infrared thermal imager to obtain the temperature distribution on the equipment surface; Collect multiple frames of thermal imaging data at a fixed sampling frequency in the time dimension to form a thermal distribution time series; Record the equipment's operating parameters, including ambient temperature, load status, and operating time.

3. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The noise reduction processing step comprises: Perform two-dimensional wavelet decomposition on the collected thermal imaging data to separate the data into high-frequency and low-frequency components; The soft threshold filtering method is used to reduce the noise of high-frequency components and filter random noise signals; The effective signal is retained for the low-frequency component, and the processed thermal imaging data is reconstructed.

4. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The optimization of the boundary region using heat flow balance conditions includes: For the complex geometric boundaries of the device surface, the boundary conditions are set as the heat flow balance conditions in the normal direction of the device surface; Neumann boundary conditions based on the device surface shape are used to ensure that the heat flux on the device surface does not flow out of the boundary.

5. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The nonlinear heat conduction model is: According to the thermal conductivity of the equipment material, a nonlinear heat conduction model is established to describe the dynamic behavior of temperature distribution changing with time; In the model, the thermal conductivity is temperature dependent and the heat diffusion behavior is regulated by the heat source characteristics of the device defect area.

6. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The high-order partial differential expansion model includes: The fourth-order partial differential term is introduced into the heat conduction model, and the heat diffusion characteristics of the defect region boundary are described by the high-order form of the Laplace operator. By introducing high-order terms, the thermal boundary clarity of the defect area is enhanced and the noise interference in the boundary area is reduced.

7. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The variational regularization method includes: Define the objective function based on the residual of the heat conduction equation, and balance the fitting accuracy of the heat distribution and the data smoothness through the regularization term; The regularization term includes a second-order partial differential regularization term and a fourth-order partial differential regularization term; The objective function is solved by numerical optimization method, and the temperature distribution on the surface and inside of the equipment is optimized.

8. The device defect detection method based on infrared recognition according to claim 1 is characterized in that: The step of inverting the heat source distribution of the defect area comprises: Using the optimized temperature distribution, calculate the heat source intensity distribution inside the equipment; According to the spatial characteristics of the heat source distribution, the location of the defect area is identified, and the geometric characteristics and thermal intensity of the defect are extracted.

9. The device defect detection method based on infrared recognition according to claim 1, characterized in that: The defect detection result obtained by inverting the heat source distribution of the defect area is used to generate a three-dimensional thermal distribution map of the defect area, and is further used to provide a report on the defect location, geometry and thermal characteristics for reference in equipment operation and maintenance decision-making.

10. An infrared recognition-based equipment defect detection system, applied to an infrared recognition-based equipment defect detection method as claimed in any one of claims 1 to 9, characterized in that: include: Infrared thermal imager, used to collect temperature distribution on the surface of equipment; A data processing module is used to reduce noise and optimize the boundaries of the collected infrared thermal imaging data; Model building module for building temperature distribution models based on nonlinear heat conduction equations and high-order partial differential extension models; Optimization and inversion module, which is used to optimize the temperature distribution using variational regularization method and invert the defect heat source; Output module, used to generate defect detection results, including geometric characteristics, heat intensity distribution and location coordinates of defects.

Images