Intelligent cleaning control method and system for engine piston ring

Abstract

The invention provides an intelligent cleaning control method and system for an engine piston ring, and the method comprises the steps: building a dynamic cleaning path through obtaining the residue phase change gradient, the laser beam incident angle, the roughness and the oxide layer thickness data of a laser thermodynamic action region on the surface of the piston ring; in the cleaning process, based on oxide layer thickness distribution, a deep learning model is used for generating target scanning frequency offset in real time, and laser power density and focusing depth parameters are further optimized. And by dynamically adjusting laser energy distribution, precise removal of residues is achieved, and meanwhile a piston ring base material is protected. The cleaning precision of the engine piston ring is improved.

Description

Technical Field

[0001] The present application relates to the field of laser precision processing and intelligent control technology, and in particular to an intelligent cleaning control method and system for an engine piston ring. Background Art

[0002] Engine piston rings, exposed to long-term high-temperature and high-pressure operating conditions, accumulate complex residues such as carbides, sludge, and oxide layers on their surfaces. Traditional mechanical or chemical cleaning methods can easily damage the underlying material. With the development of high-precision laser cleaning technology, there is an urgent need for an intelligent control method that can adapt to the piston ring surface state, ensuring thorough cleaning while avoiding thermal damage.

[0003] The current mainstream solution uses laser cleaning control based on fixed parameter rules. This uniformly cleans the piston ring using a preset laser power, scanning path, and frequency. Surface temperature is monitored using an infrared thermal imager, triggering power adjustment when the temperature exceeds a threshold. This solution achieves basic energy control through a simple temperature feedback mechanism.

[0004] This solution relies on a temperature threshold for global power regulation, which can lead to incomplete cleaning or localized overheating damage. Furthermore, a fixed scanning path struggles to adapt to changes in piston ring curvature, which can lead to uneven energy distribution. Summary of the Invention

[0005] The present application provides an intelligent cleaning control method and system for an engine piston ring, which is used to solve the problems of low cleaning accuracy and poor substrate protection effect of the engine piston ring in the prior art.

[0006] In a first aspect, the present application provides an engine piston ring intelligent cleaning control method, comprising:

[0007] Obtaining residual phase change gradient information data and laser beam incident angle value in the laser thermodynamic action area on the surface of the engine piston ring, as well as the roughness and oxide layer thickness distribution of the piston ring surface;

[0008] Constructing a dynamic cleaning path for the piston ring according to the incident angle value, the roughness and the residue phase change gradient information;

[0009] During the cleaning of the piston ring according to the dynamic cleaning path, a target scanning frequency offset is generated using a pre-trained deep learning model based on the oxide layer thickness distribution;

[0010] Laser energy distribution optimization parameters are generated according to the target scanning frequency offset. The laser energy distribution optimization parameters are used to adjust the laser power density and laser focus depth of the laser beam in real time during the cleaning process to achieve accurate removal of residues on the piston ring surface and effective protection of the base material.

[0011] Optionally, generating laser energy distribution optimization parameters according to the scanning frequency offset includes:

[0012] coupling the residue phase change gradient information data with the roughness to generate a laser power density adjustment parameter;

[0013] Generate a laser focus depth adjustment parameter based on the residual phase change gradient information data and the thermal cycle attenuation index in the piston ring operation history data, combined with the local structural morphology characteristics of the piston ring;

[0014] According to the scanning frequency offset, the laser power density adjustment parameter and the laser focus depth adjustment parameter are coupled to generate a laser energy distribution optimization parameter.

[0015] Optionally, coupling the laser power density adjustment parameter and the laser focus depth adjustment parameter according to the scanning frequency offset to generate a laser energy distribution optimization parameter includes:

[0016] According to the laser power density adjustment parameters for all areas of the piston ring surface, a numerical distribution of the laser power density adjustment parameters is constructed; according to the laser focus depth adjustment parameters for all areas of the piston ring surface, a spatial distribution of the laser focus depth adjustment parameters is constructed;

[0017] Determining a weight distribution relationship between the laser power density adjustment parameter and the laser focus depth adjustment parameter at each laser scanning point in a dynamic cleaning path based on the numerical distribution of the laser power density adjustment parameter and the spatial distribution of the laser focus depth adjustment parameter;

[0018] According to the weight distribution relationship, the laser power density adjustment parameter and the adjustment parameter are superimposed and calculated point by point to generate an initial energy distribution optimization parameter;

[0019] Based on the scanning frequency offset, the initial energy distribution optimization parameter is mapped to the laser scanning timing sequence corresponding to the dynamic cleaning path to generate an intermediate energy distribution optimization parameter associated with the timing sequence;

[0020] According to the energy gradient change of the intermediate energy distribution optimization parameter, combined with the remaining residue phase transition gradient information data, the weight distribution relationship is reversely corrected;

[0021] Based on the inversely corrected weight distribution relationship, the laser energy distribution optimization parameters are generated.

[0022] Optionally, the reverse correction of the weight distribution relationship based on the energy gradient change of the intermediate energy distribution optimization parameter and the remaining residue phase transition gradient information data includes:

[0023] Calculating an energy distribution difference based on the energy cumulative gradient difference between adjacent scanning points in the intermediate energy distribution optimization parameter;

[0024] Determine the ratio of the energy distribution difference to the remaining residue phase transition gradient information data as a weight adjustment factor;

[0025] The weight distribution relationship is modified according to the weight adjustment factor to update the superposition weight of the laser power density adjustment parameter and the laser focus depth adjustment parameter.

[0026] Optionally, the generating of the laser focus depth adjustment parameter based on the residual phase change gradient information data and the thermal cycle attenuation index in the piston ring operation history data, in combination with the local structural morphology characteristics of the piston ring, includes:

[0027] Determining phase change rate differences in different regions based on the residue phase change gradient information data;

[0028] Determine the energy absorption attenuation ratio of the thermal cycle attenuation index at each position on the piston ring surface in the piston ring operation history data based on the phase change rate difference and the thermal expansion distortion characteristics of the laser thermodynamic action area;

[0029] The energy absorption attenuation ratio is superimposed with the thermal stress accumulation at the corresponding position in the piston ring operation history data to generate a local thermodynamic influence factor;

[0030] The structural morphology weight parameter is calculated based on the product relationship between the roughness and the oxide layer thickness distribution in the local structural morphology characteristics of the piston ring;

[0031] Multiplying the local thermodynamic influence factor and the structural morphology weight parameter point by point according to the spatial coordinate position of the piston ring surface to generate a coupling correction parameter;

[0032] Based on the coupling correction parameters, the basic parameters of the laser focus depth are nonlinearly optimized to generate laser focus depth adjustment parameters.

[0033] Optionally, constructing a dynamic cleaning path for the piston ring according to the incident angle value, the roughness, and the residue phase change gradient information includes:

[0034] Based on the incident angle value, determining a main scanning direction of the laser beam on the piston ring surface, wherein the main scanning direction is parallel to a central axis of the incident angle value;

[0035] According to the roughness distribution value, the piston ring surface is divided into a high roughness area and a low roughness area;

[0036] For each divided region, searching for a sub-region in which the gradient change rate of the residual phase change gradient information data of the region exceeds a set gradient threshold, and determining a scanning point coverage range including the sub-region;

[0037] According to the main scanning direction, with the high roughness area as the starting point, the boundary vertices of the scanning point coverage range corresponding to each area are connected in sequence to generate a dynamic cleaning path.

[0038] Optionally, the generating a target scanning frequency offset based on the oxide layer thickness distribution using a pre-trained deep learning model includes:

[0039] Dividing the piston ring surface into a plurality of oxide layer regions with different thickness levels according to the oxide layer thickness distribution, extracting the average oxide layer thickness of each oxide layer region, and using the average oxide layer thickness as the thickness characteristic value of the corresponding region;

[0040] Inputting the thickness characteristic values ​​of all the regions into a pre-trained deep learning model, the deep learning model performs feature fusion on the thickness characteristic value of each region with the residual phase change gradient information data of the corresponding region in the dynamic cleaning path, and outputs a predicted value of the energy accumulation time of each region;

[0041] Based on the energy accumulation time prediction value, the optimal cleaning parameter combination for the same thickness grade area in the piston ring operation history data is called to generate an initial scanning frequency offset;

[0042] Correcting the initial scanning frequency offset according to a sudden change gradient of oxide layer thickness in adjacent areas in the dynamic cleaning path to generate an intermediate scanning frequency offset;

[0043] The initial scanning frequency offset and the intermediate scanning frequency offset are superimposed according to the spatial coordinate position in a weighted summation manner to generate a target scanning frequency offset.

[0044] In a second aspect, the present application provides an engine piston ring intelligent cleaning control system, comprising:

[0045] An acquisition module is used to obtain the residual phase change gradient information data and the incident angle value of the laser beam in the laser thermodynamic action area on the surface of the engine piston ring, as well as the roughness and oxide layer thickness distribution of the piston ring surface;

[0046] A construction module, configured to construct a dynamic cleaning path for the piston ring according to the incident angle value, the roughness, and the residue phase change gradient information;

[0047] a generation module, configured to generate a target scanning frequency offset using a pre-trained deep learning model based on the oxide layer thickness distribution during cleaning of the piston ring according to the dynamic cleaning path;

[0048] An adjustment module is used to generate laser energy distribution optimization parameters based on the target scanning frequency offset. The laser energy distribution optimization parameters are used to adjust the laser power density and laser focus depth of the laser beam in real time during the cleaning process to achieve accurate removal of residues on the piston ring surface and effective protection of the base material.

[0049] In a third aspect, the present application provides a computing device comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute an engine piston ring intelligent cleaning control method as described in any one of the first aspects.

[0050] In a fourth aspect, the present application provides a computer storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement an engine piston ring intelligent cleaning control method as described in any one of the first aspects.

[0051] In the present application, a method for intelligent cleaning control of an engine piston ring is provided, which includes: obtaining residual phase change gradient information data and an incident angle value of a laser beam in a laser thermodynamic action area on the surface of an engine piston ring, as well as the roughness of the piston ring surface and the oxide layer thickness distribution; constructing a dynamic cleaning path for the piston ring based on the incident angle value, the roughness and the residual phase change gradient information; in the process of cleaning the piston ring according to the dynamic cleaning path, generating a target scanning frequency offset based on the oxide layer thickness distribution using a pre-trained deep learning model; generating laser energy distribution optimization parameters based on the target scanning frequency offset, the laser energy distribution optimization parameters being used to adjust the laser power density and laser focusing depth of the laser beam in real time during the cleaning process, so as to achieve precise removal of residues on the piston ring surface and effective protection of the base material.

[0052] The technical solution of this application has the following beneficial effects:

[0053] This application provides comprehensive surface state characteristics for intelligent cleaning by collecting data on the phase change gradient of the residue, the laser incident angle, the roughness and the thickness of the oxide layer in the area where the laser thermodynamics is applied, thereby ensuring the data basis for parameter optimization. By combining the incident angle, roughness and the phase change gradient information of the residue, an adaptive cleaning path that matches the surface morphology and contamination state of the piston ring is generated to solve the problem of uneven energy distribution caused by the traditional fixed path. Based on the oxide layer thickness distribution and the deep learning model, the laser scanning frequency is dynamically adjusted to achieve precise control of the cleaning speed in different areas and balance the cleaning efficiency and effect. By adjusting the laser power density and the focus depth in real time, the residue is ensured to be completely removed while avoiding thermal damage to the base material, thereby improving the cleaning accuracy and safety.

[0054] Furthermore, the present application also generates power density adjustment parameters by coupling the residual phase change gradient and roughness, generates focus depth adjustment parameters by combining the thermal cycle attenuation index and local morphology characteristics, and finally merges the two types of parameters into laser energy distribution optimization parameters based on the scanning frequency offset to achieve dynamic coordinated control of energy distribution.

[0055] In addition, this technology solves the problem of matching laser energy distribution with the complex surface state of the piston ring through multi-parameter coupling and dynamic optimization, while improving the residue removal rate and reducing the risk of thermal damage to the substrate, achieving dual optimization of cleaning accuracy and material protection.

[0056] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0058] Figure 1 A flow chart of an engine piston ring intelligent cleaning control method provided in an embodiment of the present application;

[0059] Figure 2 A schematic diagram of the structure of an engine piston ring intelligent cleaning control system provided in an embodiment of the present application;

[0060] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0062] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0063] Researchers have found that existing engine piston ring laser cleaning technology has problems with fixed cleaning parameters and inability to adapt to complex surface conditions, resulting in incomplete cleaning or damage to the substrate material. Based on this, an embodiment of the present application provides an intelligent cleaning control method for engine piston rings. This method constructs a dynamic cleaning path by acquiring multi-dimensional data such as the phase change gradient of surface residues, laser incident angle, roughness, and oxide layer thickness in real time, and generates a scanning frequency offset based on a deep learning model. Ultimately, it optimizes the coordinated control of laser power density and focal depth to achieve precise removal of residues and effective protection of the substrate material. The technical solution of this application can be applied to industrial scenarios such as high-precision engine piston ring maintenance and remanufacturing.

[0064] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0065] Figure 1 This is a flow chart of an engine piston ring intelligent cleaning control method provided in an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0066] Step 101: Obtaining residual phase change gradient information data and incident angle value of the laser beam in the laser thermodynamic action area on the surface of the engine piston ring, as well as the roughness and oxide layer thickness distribution of the piston ring surface.

[0067] In this step, the laser thermodynamic action area refers to the heat-affected zone generated by the interaction between the laser beam and the piston ring surface, and its range is determined by the laser parameters (power, spot size) and the thermal conductivity of the material. The residual phase change gradient information data represents the difference in phase change rate of residues in different regions under the action of the laser, and is obtained by inverting the temperature field data collected by the infrared thermal imager. The incident angle value represents the angle between the axis of the laser beam and the normal of the piston ring surface, which affects the energy absorption rate. Roughness refers to the microscopic unevenness of the surface, which is measured by a white light interferometer. The oxide layer thickness distribution refers to the spatial variation in the thickness of the surface oxide film, which is detected by laser-induced breakdown spectroscopy.

[0068] In this embodiment, a probe is first used to scan the piston ring circumference to obtain oxide layer thickness distribution data, while a confocal microscope is used to simultaneously collect surface roughness data. Simultaneously, an infrared thermal imager is used to monitor temperature changes during the laser pre-scan process, and a heat conduction model is used to invert the residual phase transition gradient. A laser tracking system records the incident angle in real time. All data is then spatiotemporally aligned by edge computing nodes and uploaded to a central controller to construct a surface state database.

[0069] For example, using a diesel engine piston ring as an example, laser-induced breakdown spectroscopy revealed an oxide layer thickness of 50 μm in the groove area, while only 10 μm in the land area. Infrared thermal imaging revealed a slow phase transition rate in the groove area (0.2 K / μs). The laser incident angle varied between 60° and 85° depending on the ring curvature. These data provide a basis for subsequent path planning.

[0070] Step 102: constructing a dynamic cleaning path for the piston ring according to the incident angle value, the roughness and the residue phase change gradient information.

[0071] In this step, the dynamic cleaning path represents the laser scanning trajectory that is adaptively adjusted according to the real-time surface status and contains spatial coordinates and timing information.

[0072] In this embodiment, a modified Dijkstra algorithm is used for path planning. An energy transfer efficiency matrix is ​​constructed based on incident angle values, an obstacle weight matrix is ​​generated based on roughness data, and a priority matrix is ​​formed based on phase gradient information. Through multi-objective optimization, an optimal path is obtained that balances cleaning efficiency and safety. Specifically, this involves 1. discretizing the piston ring surface into a grid; 2. calculating a comprehensive cost function for each grid point; and 3. applying dynamic programming to solve the global optimal path.

[0073] For example, continuing with the previous example, for the high oxide layer thickness in the annular groove area, the algorithm automatically generates a spiral progressive path, increasing the path density in the annular groove area by 50%; for the 60° incident angle area, the path spacing is reduced by 30% to compensate for the decrease in energy absorption efficiency.

[0074] Step 103: During the cleaning of the piston ring according to the dynamic cleaning path, a target scanning frequency offset is generated based on the oxide layer thickness distribution using a pre-trained deep learning model.

[0075] In this step, the deep learning model is an intelligent decision-making system trained based on the correspondence between a large amount of piston ring surface state data and the optimal cleaning parameters. The model analyzes the spatial characteristics of the oxide layer thickness distribution and combines it with the material thermophysical property database to predict the laser action time required for each area, and then converts it into a scanning frequency adjustment amount. The training process uses 100,000 samples collected from the actual production line, covering piston rings with different degrees of wear, oil types and oxidation states, to ensure that the model has strong generalization capabilities. When applied online, the model updates the frequency offset output every 50ms to achieve millisecond-level response. The target scanning frequency offset represents the laser scanning rate correction value dynamically adjusted based on real-time working conditions, and the unit is Hz / mm.

[0076] In this embodiment, a deep learning model is deployed that receives an oxide layer thickness distribution map and outputs a frequency adjustment coefficient. The model is trained using a transfer learning strategy, pre-training on 100,000 sets of simulated data before fine-tuning using 5,000 sets of measured data. During online inference, laser power feedback is combined with real-time correction to generate a spatially varying frequency offset field.

[0077] For example, when it is detected that the oxide layer at a certain point in the ring groove area suddenly increases to 80μm, the model output frequency is reduced by 40% (from 200Hz to 120Hz), and the adjacent areas are transitioned according to the gradient to avoid energy mutation.

[0078] Step 104: Generate laser energy distribution optimization parameters based on the target scanning frequency offset. The laser energy distribution optimization parameters are used to adjust the laser power density and laser focus depth of the laser beam in real time during the cleaning process to achieve accurate removal of residues on the piston ring surface and effective protection of the base material.

[0079] In this step, the laser energy distribution optimization parameters include a power density correction coefficient (0.8-1.2 times the reference value) and a focus depth adjustment amount (±0.2 mm).

[0080] In this embodiment, a scanning frequency-energy coupling model is established, and the frequency offset is decomposed into a spatial spectrum using Fourier transform. The power density adjustment is calculated based on the impedance matching principle, and the depth of focus compensation value is derived using the depth of focus formula. The resulting optimized parameters are transmitted to the laser controller in real time via a fieldbus.

[0081] For example, for a 120Hz scanning frequency in the annular groove area, the system automatically increases the power density to 1.15 times and reduces the focus depth by 0.15mm to ensure that no heat accumulation occurs at low scanning speeds.

[0082] This solution improves piston ring cleaning quality through multi-dimensional surface condition perception, intelligent path planning, and dynamic parameter optimization. The system automatically identifies contamination characteristics and material conditions in different areas and precisely controls laser energy distribution, thoroughly removing stubborn residues while effectively protecting the integrity of the underlying material. Practical applications have demonstrated that this solution achieves industrial-grade surface cleanliness for cleaned piston rings while completely avoiding the localized overheating and incomplete cleaning common with traditional methods, thereby enhancing the reliability of remanufactured engine products.

[0083] In addition, it should be noted that the above steps can be implemented when the engine is not disassembled.

[0084] In order to solve the problem of mismatch between energy distribution and surface condition during laser cleaning of piston rings and further improve cleaning accuracy and substrate protection, in some embodiments, step 104: generating laser energy distribution optimization parameters based on the scanning frequency offset includes:

[0085] Step 201: coupling the residue phase change gradient information data with the roughness to generate a laser power density adjustment parameter.

[0086] In step 201, the laser power density adjustment parameter refers to a laser power correction factor dynamically adjusted according to the surface state, ranging from 0.7 to 1.3 times the reference value. This parameter comprehensively reflects the difficulty of residue removal and the heat resistance threshold of the substrate.

[0087] In the embodiment of the present application, this is achieved by establishing a coupling model of the residue phase change gradient and roughness. First, the phase change gradient data is normalized to a cleaning difficulty coefficient of 0-1, and the roughness data is converted into a surface absorption rate correction coefficient (0.9-1.1). The two are combined through a weighted fusion algorithm to generate a power density adjustment parameter.

[0088] Step 202: Generate a laser focus depth adjustment parameter based on the residue phase change gradient information data and the thermal cycle attenuation index in the piston ring operation history data, combined with the local structural morphology characteristics of the piston ring.

[0089] In step 202, local structural morphological features refer to the microscopic physical structural characteristics of the piston ring surface, including: microcrack distribution (such as fatigue cracks and thermal stress cracks), surface pits or protrusions (morphology caused by machining marks or wear), and local curvature changes (such as the transition zone at the piston ring edge). The laser thermodynamic action zone refers to the macroscopic area on the piston ring surface where the laser beam actually acts. Its range is determined by factors such as the laser beam spot size and power density, the thermal conductivity characteristics of the piston ring surface, and the distribution of residues. The difference between the two is the spatial scale: local structural morphological features focus on microscopic morphological details (μm level); the laser thermodynamic action zone focuses on the energy coverage at the macroscopic scale (mm level). The energy distribution in the laser thermodynamic action zone must be adapted to the local structural morphological features. For example, in microcrack areas (micromorphology), the power density must be reduced within the thermodynamic action zone (macroscopic range); in areas with sudden changes in curvature (micromorphology), the focus depth of the thermodynamic action zone must be adjusted. The thermal cycle attenuation index (TCI) is a characteristic parameter that quantifies the degree of performance degradation of piston ring materials under repeated thermal cycling loads and reflects the material's resistance to thermal fatigue. The laser focus depth adjustment parameter refers to the amount of focus position adjustment determined based on the material's thermal history and micromorphology, with an adjustment range of ±0.3mm.

[0090] In this application, a three-stage calculation is implemented: 1. Calculation of the material's thermal fatigue coefficient based on the thermal cycle attenuation index; 2. Extraction of microcrack density and curvature characteristics from the local structural morphology; and 3. Establishment of a thermal-mechanical coupling equation to output a focus depth compensation value. This coupling equation is solved in real time using a proxy model trained with finite element simulation data.

[0091] Step 203: according to the scanning frequency offset, the laser power density adjustment parameter and the laser focus depth adjustment parameter are coupled to generate a laser energy distribution optimization parameter.

[0092] In step 203 , the laser energy distribution optimization parameter is a comprehensive control instruction that integrates power density and focus depth, and includes a spatial coordinate-energy value mapping table and a time series control instruction.

[0093] In an embodiment of the present application, parameter fusion is achieved through space-time coding technology, the scanning frequency offset is converted into a time reference axis, the power and focus parameters are interpolated and aligned on the time axis, and parameter conflicts are eliminated through an energy conservation verification algorithm, and finally a control instruction stream including a 200Hz refresh rate is generated.

[0094] Here's a specific example:

[0095] During the laser cleaning process of a certain diesel engine piston ring, the system detected a residual phase change gradient of 0.85 (calculated by inverting the temperature field measured with an infrared thermal imager) and a surface roughness of Ra3.2μm (measured with a white light interferometer). Combining these two parameters, a weighted fusion algorithm was used to generate a laser power density adjustment parameter for this area of ​​1.2 times the baseline value (with the weight coefficients taking the phase change gradient as 70% and the roughness as 30%). Furthermore, a query of the piston ring's operational history database revealed a thermal cycle attenuation index of 0.65 (calculated based on 2000 300-20°C temperature cycles and a 25% drop in hardness). Combined with the local structural morphology detection, which revealed a microcrack density of 5 cracks / mm (observed with a confocal microscope), a thermal-mechanical coupling model was used to calculate the need to reduce the laser focus depth by 0.15mm. Finally, based on the scanning frequency offset of 150Hz (the base frequency of 200Hz was reduced by 50Hz) output by the deep learning model for the oxide layer thickness of 80μm in this area (measured by laser-induced breakdown spectroscopy), the two parameters of power density 1.2 times and focusing depth -0.15mm were fused through space-time coding technology to generate the final laser energy distribution optimization parameters for this area, thereby achieving effective removal of carbides in this area (no residue was confirmed by electron microscopy after cleaning) while protecting the base material (microhardness testing showed no hardness drop in the heat-affected zone).

[0096] In the embodiments of the present application, this solution achieves precise matching of laser energy and surface state, safe cleaning of complex morphology areas, and stable control under dynamic working conditions through multi-parameter dynamic coupling and real-time optimization, providing a reliable solution for the remanufacturing of high-value components.

[0097] In order to solve the problem of insufficient spatiotemporal matching accuracy in the laser energy parameter coupling process and further improve the accuracy of coordinated control of cleaning parameters, in some embodiments, step 203: coupling the laser power density adjustment parameter and the laser focus depth adjustment parameter according to the scanning frequency offset to generate the laser energy distribution optimization parameter includes:

[0098] Step 301: constructing a numerical distribution of laser power density adjustment parameters based on laser power density adjustment parameters for all areas of the piston ring surface, and constructing a spatial distribution of laser focus depth adjustment parameters based on laser focus depth adjustment parameters for all areas of the piston ring surface.

[0099] In step 301, the numerical distribution of the laser power density adjustment parameter refers to the spatial arrangement matrix of the power correction coefficients for each surface area, with matrix unit values ​​ranging from 0.7 to 1.3 times the reference power ratio. The spatial distribution of the laser focus depth adjustment parameter refers to the three-dimensional coordinate set of the focus position adjustment amount for each area, including the axial offset (±0.3mm) and radial compensation.

[0100] In an embodiment of the present application, parameter distribution construction is achieved by establishing a surface grid coordinate system. The specific process is to divide the piston ring surface into 1mm×1mm grid units, fill the power density adjustment parameters into the numerical matrix according to the grid coordinates, and convert the focus depth parameters into a three-dimensional vector and associate them with the corresponding grid.

[0101] Step 302: Based on the numerical distribution of the laser power density adjustment parameter and the spatial distribution of the laser focus depth adjustment parameter, determine the weight distribution relationship between the laser power density adjustment parameter and the laser focus depth adjustment parameter at each laser scanning point in the dynamic cleaning path.

[0102] In step 302, the weight distribution relationship refers to the synergistic proportional coefficient between the power density and the focal depth parameter in the spatiotemporal dimension, including a spatial weight (0.4-0.6) and a temporal weight (0.3-0.7).

[0103] In this embodiment, a dynamic programming algorithm is used to determine the weights. The specific process is to establish a spatiotemporal constraint equation based on the topological relationship of the scanning path, use energy distribution uniformity as the objective function, and solve the optimal weights using the Lagrange multiplier method. The power density weight is set to be greater than 0.5 in the high temperature zone, and the focus depth weight is set to be 0.6 in the deformation zone.

[0104] Step 303: performing point-by-point superposition calculation on the laser power density adjustment parameter and the adjustment parameter according to the weight distribution relationship to generate initial energy distribution optimization parameters.

[0105] In step 303, the initial energy distribution optimization parameters are the original energy control instruction set without time sequence calibration, including the energy intensity value (unit J / mm) of each grid point. 2 ) and action time (μs level).

[0106] In the embodiment of the present application, it is achieved through tensor superposition operation. The specific process is to perform Kronecker product operation on the power density matrix and the focus depth tensor, introduce the material absorption rate coefficient for correction, and output the initial energy distribution optimization parameters containing 100,000 control nodes.

[0107] Step 304: Based on the scanning frequency offset, the initial energy distribution optimization parameter is mapped to the laser scanning timing sequence corresponding to the dynamic cleaning path to generate an intermediate energy distribution optimization parameter associated with the timing sequence.

[0108] In step 304, the laser scanning timing sequence refers to the time-space coordinated instruction set that controls the laser parameters as the laser beam moves along the dynamic cleaning path. The timing-associated intermediate energy distribution optimization parameters refer to a parameter sequence synchronized with the laser scanning timing, with a time resolution of 50μs.

[0109] In an embodiment of the present application, space-time mapping is achieved through cubic spline interpolation. The specific process includes converting the scanning frequency offset into non-uniform sampling points on the time axis, performing sliding window interpolation on the initial parameters, and generating a timing instruction stream synchronized with the laser galvanometer control signal.

[0110] Step 305: According to the energy gradient change of the intermediate energy distribution optimization parameter and in combination with the remaining residue phase transition gradient information data, the weight distribution relationship is reversely corrected.

[0111] In step 305, the energy gradient change refers to the cumulative difference in laser energy between adjacent laser scanning points during the dynamic cleaning process, reflecting the continuity and mutation of energy distribution. Its essence is the energy change rate along the cleaning path after the laser power density adjustment parameter and the focus depth adjustment parameter are spatially superimposed. The expression is as follows: the spatial gradient is the difference between the energy value of a certain scanning point and the energy value of its adjacent points (for example, the energy of point A is 100J / cm 2 , the adjacent point B is 80J / cm 2 , then the gradient change is 20J / cm 2 The time gradient is the energy change of the same scanning point in the continuous cleaning cycle (such as the first scanning energy is 100J / cm 2 , the second scan was adjusted to 110J / cm 2 For example, the scenario is: there is a sudden change in the thickness of the oxide layer in a certain area on the piston ring surface (for example, the oxide layer in area 1 is 100 μm thick, and the thickness in the adjacent area 2 is 30 μm), which results in: Laser power density adjustment parameters: Area 1 requires high power (120 J / cm 2 ), area 2 requires low power (60J / cm 2). Laser focus depth adjustment parameters: Area 1 requires shallow focus (0.5mm), and Area 2 requires deep focus (1.0mm). Energy gradient change generation process: Initial energy distribution optimization parameter generation: Area 1 superposition result: high power + shallow focus is an energy value of 150 (unit: normalized value); Area 2 superposition result: low power + deep focus is an energy value of 50. Energy gradient change: The energy mutation value from area 1 to area 2 is 100 (150-50). Reverse correction is a feedback process that adjusts the weight distribution according to the gradient abnormality area.

[0112] In the embodiment of the present application, a control strategy is adopted for correction. The specific process is to detect the gradient mutation point (change rate > 15%), query the remaining residue data of the corresponding area, and adjust the weight coefficient according to the deviation ratio (such as if the gradient exceeds the standard by 10%, the weight is adjusted by 5%).

[0113] Step 306: Generate laser energy distribution optimization parameters based on the inversely corrected weight distribution relationship.

[0114] In the embodiment of the present application, it is generated through an iterative convergence algorithm, and the specific process is to execute 3-5 weight correction cycles, terminate when the energy gradient change rate is stable within ±5%, and output the final parameters including spatial coding and timestamp.

[0115] Here's a specific example:

[0116] During the laser cleaning process of a diesel engine piston ring, the system performed energy parameter coupling optimization for the top dead center region (coordinates X20-Y45 to X25-Y50). Based on the obtained parameters for this region, the following were applied: 1) a power density adjustment parameter of 1.2x (derived from a fusion of a residue phase transition gradient of 0.85 and a roughness of 3.2μm); 2) a focus depth adjustment parameter of -0.15mm (calculated from a thermal cycling attenuation exponent of 0.65 and a microcrack density of 5 / mm); and 3) a scanning frequency offset of 150Hz (corresponding to an oxide layer thickness of 80μm). The system then divided the region into a fine 0.5mm×0.5mm grid. A power density numerical distribution matrix was constructed (1.2x in the center, gradient-shifted to 1.1x at the edges). A focus depth spatial distribution vector field (-0.15mm at the center, -0.12mm at the edges) was generated. A dynamic weighting algorithm was used to determine a power weight of 0.55 for the center region and a focus weight of 0.45 for the edges. During the point-by-point superposition calculation, a 20% energy gradient change was observed at point X23-Y48 (caused by a power jump at an adjacent point). Immediately, a reverse correction was performed based on the residual phase transition gradient data of 0.82 at this point, reducing the power weight to 0.5 and increasing the focus weight to 0.5. The resulting optimized parameters enabled this region to be scanned at a 150Hz frequency.

[0117] In the embodiment of the present application, the scheme achieves millimeter-level matching accuracy of laser energy and surface state, microsecond-level parameter synchronization during dynamic scanning, and adaptive energy adjustment in complex morphology areas through multi-level optimization and closed-loop control of spatiotemporal parameters, providing a reliable intelligent control solution for laser cleaning of precision parts.

[0118] In order to solve the problem of weight distribution inaccuracy during the dynamic adjustment of laser energy parameters and further improve the accuracy of coordinated control of cleaning parameters, in some embodiments, step 305: according to the energy gradient change of the intermediate energy distribution optimization parameter, combined with the residual residue phase transition gradient information data, the weight distribution relationship is reversely corrected, including:

[0119] Step 401: Calculate the energy distribution difference based on the energy cumulative gradient difference between adjacent scanning points in the intermediate energy distribution optimization parameters.

[0120] In step 401, adjacent scanning points refer to spatially adjacent surface points continuously illuminated by the laser beam on the dynamic cleaning path. In the cleaning path, there are two points before and after the current scanning point (a total of five points constitute the calculation window), and the spacing is determined by the scanning frequency and laser movement speed (e.g., at a scanning frequency of 200Hz, the point spacing is 0.1mm). For example: if the current point is A, the adjacent points are the first two points (A-2, A-1) and the last two points (A+1, A+2). The energy cumulative gradient difference is an indicator that quantifies the degree of sudden change in laser energy distribution between adjacent points. The energy distribution difference refers to the relative rate of change of energy values ​​between adjacent laser scanning points. It is obtained by calculating the percentage difference between the energy value of the current point and the average energy value of the two points before and after it, and is used to quantify the degree of sudden change in energy distribution.

[0121] In the embodiment of the present application, the five-point difference method is used to calculate the gradient difference. The specific process is to select five positions, namely the target point and two adjacent points before and after it, and calculate the percentage difference between the energy value of the target point and the average energy value of the adjacent four points.

[0122] Step 402: Determine the ratio of the energy distribution difference to the remaining residue phase change gradient information data as a weight adjustment factor.

[0123] In step 402, the residual residue phase transition gradient information data refers to the state characteristics of the remaining residue that has not been removed, updated in real time during the laser cleaning process. The weight adjustment factor is a proportional coefficient used to dynamically adjust the weights of the power density and focus depth parameters. It ranges from 0.8 to 1.2 and is determined by the ratio of the energy difference to the residue phase transition rate.

[0124] In the embodiment of the present application, an adjustment factor is generated through normalization processing. The specific process is to convert the energy difference into a standardized value of 0-1, take the inverse of the residue phase change gradient data as the sensitivity coefficient, and output the final adjustment factor after multiplying the two and constraining them.

[0125] Step 403: Modify the weight distribution relationship according to the weight adjustment factor to update the superposition weight of the laser power density adjustment parameter and the laser focus depth adjustment parameter.

[0126] In step 403, weight correction refers to dynamically optimizing the coordinated ratio of power density and focus depth parameters according to real-time working conditions to ensure that the energy distribution matches the surface state.

[0127] In the embodiment of the present application, an incremental adjustment strategy is adopted, the weight history data of the first three iterations is saved, and a linear or nonlinear correction mode is selected according to the size of the adjustment factor. The power density weight and the focus depth weight are adjusted in the same proportion and inversely, so that the total weight sum is kept at 1. The minimum adjustment step size of 0.1 is set during the correction process.

[0128] Here's a specific example:

[0129] During the laser cleaning process of a certain diesel engine piston ring, the system performs weight correction optimization on the top dead center area (coordinates X23-Y48 and four points within 0.5mm around it). The specific process is: first, the energy cumulative gradient difference of point X23-Y48 is calculated by the five-point difference method, which is 20% (the energy value at this point is 180J / cm 2 The average energy value of the four surrounding points X22-Y47, X22-Y49, X24-Y47, and X24-Y49 is 150 J / cm 2 ), and at the same time obtain the phase change gradient data of the remaining residue at this point 0.82 (obtained by inversion calculation of the temperature field change detected by a real-time infrared thermal imager). Multiply the ratio of the difference of 20% and the phase change gradient of 0.82 by the normalization coefficient of 0.5 to obtain the weight adjustment factor 1.22 (20 / 0.82×0.5). According to this factor, the original weight distribution (power weight 0.55, focus weight 0.45) is corrected: the power weight is adjusted to 0.55 / 1.22≈0.45, and the focus weight is adjusted to 1-0.45=0.55. After correction and recalculation, the energy value of this point is reduced to 165J / cm 2 , the difference with the surrounding average is reduced to 10%, and the phase change gradient is reduced to 0.35 (confirmed by secondary infrared detection), realizing the simultaneous optimization of energy distribution and cleaning effect.

[0130] In the embodiment of the present application, the solution achieves millisecond-level identification and correction of abnormal energy mutations, adaptive optimization and adjustment of weight parameters, and precise balance between cleaning effect and material protection by establishing a dynamic feedback mechanism of energy gradient-residue state, so that the laser cleaning process has stronger adaptability to working conditions.

[0131] In order to solve the problem of mismatch between the focus depth parameter and the material thermal damage risk during laser cleaning of piston rings and further improve the accuracy of laser energy control, in some embodiments, step 202: generating the laser focus depth adjustment parameter based on the residue phase transition gradient information data and the thermal cycle attenuation index in the piston ring operation history data, combined with the local structural morphology characteristics of the piston ring, includes:

[0132] Step 501: Determine the phase change rate differences between different regions based on the residue phase change gradient information data.

[0133] In step 501, "different regions" refer to sub-regions of the piston ring surface divided based on the residual phase transition gradient information data. Different regions are further subdivisions of the laser thermodynamic action area. For example, a laser thermodynamic action area (e.g., the coverage area of ​​a 2mm diameter circular spot) may contain multiple "different regions" (e.g., a central region with a fast phase transition rate and an edge region with a slow phase transition rate). Phase transition rate differences refer to the differences in the thermal response speed of residues in different regions of the piston ring surface under laser action. This difference is calculated using the temperature change rate captured by an infrared thermal imager and has a value range of 0.1-1.0, with larger values ​​indicating faster phase transition.

[0134] In the embodiment of the present application, a spatiotemporal difference algorithm is used to calculate the phase change rate. The specific process is to obtain the temperature-time curve of each region in the laser pre-scanning stage, perform the first-order derivative of the curve to obtain the instantaneous phase change rate, and generate the phase change rate after normalization.

[0135] Step 502: Determine the energy absorption attenuation ratio of the thermal cycle attenuation index at each position on the piston ring surface in the piston ring operation history data based on the phase change rate difference and the thermal expansion distortion characteristics of the laser thermodynamic action area.

[0136] In step 502, thermal expansion distortion refers to the geometric deformation of the material caused by uneven local heating when the laser irradiates the piston ring surface. The energy absorption attenuation ratio quantifies the degree of thermal aging of the material, reflecting the effective utilization of laser energy in that area. Its value ranges from 0.5 to 1.2 and is calculated by multiplying the thermal cycle attenuation index by the real-time thermal expansion data.

[0137] In the embodiment of the present application, the thermodynamic coupling model is used for calculation. The specific process is to extract the temperature cycle number and peak temperature data in the thermal cycle attenuation index, and combine the thermal expansion strain value during the laser action (measured by laser speckle interferometer) to calculate the energy absorption efficiency loss ratio.

[0138] Step 503: superimpose the energy absorption attenuation ratio and the thermal stress accumulation at the corresponding position in the piston ring operation history data to generate a local thermodynamic influence factor.

[0139] In step 503, the accumulated thermal stress is used to quantify the degree of internal stress damage to the piston ring material caused by repeated thermal cycling during service. The local thermodynamic impact factor is a risk assessment indicator that integrates thermal history and real-time thermal response. Its value ranges from 0 to 2, with larger values ​​indicating a higher risk of thermal damage.

[0140] In the embodiment of the present application, it is generated by weighted superposition. The specific process is to perform logarithmic conversion on the energy absorption attenuation ratio, divide the accumulated thermal stress (unit: MPa·h) by the yield strength of the material, add the two at a weight of 6:4, and output the range.

[0141] Step 504: Calculate a structural morphology weight parameter based on the product relationship between the roughness and the oxide layer thickness distribution in the local structural morphology characteristics of the piston ring.

[0142] In step 504, the structural morphology weight parameter is a coefficient that characterizes the degree of influence of surface geometric features on laser focusing, with a value range of 0.8-1.5, and is obtained by normalizing the product of the roughness arithmetic mean (Ra) and the oxide layer thickness.

[0143] In the embodiment of the present application, the Ra value (white light interferometer) and the oxide layer thickness (laser induced breakdown spectroscopy) at each point are measured, and both are dimensionlessly processed. The weight is calculated according to the formula (Ra×thickness) / maximum value, and an additional compensation coefficient of 0.2 is added to the microcrack area to obtain the structural morphology weight parameter.

[0144] Step 505: multiply the local thermodynamic influence factor and the structural morphology weight parameter point by point according to the spatial coordinate position of the piston ring surface to generate a coupling correction parameter.

[0145] In step 505, the spatial coordinate position of the piston ring surface refers to the precise positioning data of each detection point on the piston ring surface in a three-dimensional coordinate system. The coordinate system is defined as follows: the axial coordinate X represents the direction along the piston ring axis (equally divided between 0 and 360 degrees), the radial coordinate Y represents the radial distance from the inner diameter to the outer diameter, and the height coordinate Z represents the surface topography fluctuation height (±0.5mm). The coupling correction parameter is a combined correction factor for thermodynamic and topographic characteristics, used to guide the spatial variation of the focus depth, with a value range of 0.5-2.0.

[0146] In the embodiment of the present application, a point-by-point multiplication fusion algorithm is adopted. The specific process is to establish a surface grid coordinate system, spatially align the local thermodynamic influencing factors and structural morphology weight parameters, multiply them element by element according to the grid points, and generate coupling correction parameters.

[0147] Step 506: Based on the coupling correction parameters, perform nonlinear optimization on the basic parameters of the laser focus depth to generate laser focus depth adjustment parameters.

[0148] In step 506 , the basic parameter of the laser focal depth refers to the initial focus position setting value before correction.

[0149] In the embodiment of the present application, nonlinear mapping is used for generation, and the specific process is to set a basic focusing depth (such as 0.5 mm), input the coupling correction parameter into a pre-calibrated cubic polynomial function, and output the focusing depth compensation value of each point.

[0150] Here's a specific example:

[0151] During the laser cleaning process of a certain diesel engine piston ring, the system optimized the depth of focus parameters for the top dead center region (coordinates X20-Y45). The following data were measured for this region: 1. Residue phase transition gradient of 0.85 (calculated by infrared thermal imaging inversion); 2. Cyclic attenuation index of 0.65 (calculated based on 2000 cycles of 300-20°C temperature and a 25% hardness drop); and 3. Real-time thermal expansion distortion of 0.12 mm (measured by laser speckle pattern interferometry). The energy absorption attenuation ratio was first calculated as 0.72 (derived by multiplying the thermal cycling index and the thermal expansion coefficient). This was then combined with the accumulated thermal stress of 80 MPa·h at this point (derived from a historical database) to generate a local thermodynamic influence factor of 1.25. Furthermore, the product of the roughness of 3.2 μm (white light interferometry) and the oxide layer thickness of 80 μm (laser-induced breakdown spectroscopy) at this point (256, dimensionless) was converted into a structural morphology weighting parameter of 1.28 (normalized). Multiplying the thermodynamic factor 1.25 by the topography parameter 1.28 at each coordinate point X20-Y45 yields the coupling correction parameter 1.6. Based on this parameter, a nonlinear optimization (using a cubic polynomial mapping) is performed on the base focus depth of 0 mm, ultimately generating a focus depth adjustment parameter of -0.18 mm.

[0152] In the embodiment of the present application, the solution achieves millimeter-level matching of focus depth and material thermal damage threshold, adaptive compensation of complex surface morphology, and fusion application of historical working conditions and real-time monitoring data through thermal-mechanical-morphological multi-parameter coupling and nonlinear optimization, making the laser cleaning process both efficient and safe.

[0153] In order to solve the problem of uneven energy distribution in piston ring laser cleaning path planning and further improve cleaning efficiency and quality, in some embodiments, step 102: constructing a dynamic cleaning path for the piston ring based on the incident angle value, the roughness, and the residue phase change gradient information includes:

[0154] Step 601: Based on the incident angle value, determine the main scanning direction of the laser beam on the piston ring surface, where the main scanning direction is parallel to the central axis of the incident angle value.

[0155] In step 601, the main scanning direction refers to the reference trajectory direction of the laser beam moving on the piston ring surface, which is determined by projecting the central axis of the laser incident angle value onto the piston ring surface.

[0156] In an embodiment of the present application, the incident angle (e.g., 60°) is acquired in real time by a laser tracking system, and the central axis corresponding to the angle is projected onto the surface of the piston ring three-dimensional model to generate a reference spiral line as the main direction trajectory, and its pitch is automatically adjusted according to the height of the piston ring (e.g., an increase of 2 mm per turn).

[0157] Step 602: Divide the piston ring surface into a high-roughness area and a low-roughness area according to the roughness distribution value.

[0158] In step 602 , the high roughness region refers to a sub-region where the surface roughness value exceeds a set threshold (eg, Ra>3.0 μm), and the low roughness region refers to a region where the roughness value is lower than the threshold.

[0159] In an embodiment of the present application, an adaptive threshold segmentation algorithm is used to perform Gaussian filtering on the roughness data measured by the white light interferometer, and the global roughness mean (such as Ra2.5μm) is calculated as a benchmark. The area that exceeds 20% of the benchmark value (i.e., Ra>3.0μm) is marked as a high roughness area, and the rest is a low roughness area.

[0160] Step 603: for each divided region, search for a sub-region in which the gradient change rate of the residue phase change gradient information data of the region exceeds a set gradient threshold, and determine a scanning point coverage range including the sub-region.

[0161] In step 603 , the scanning point coverage refers to the minimum convex polygon boundary containing the sub-region with high gradient change rate, and the gradient change rate threshold is set to 0.5 (normalized value of phase change gradient data).

[0162] In the embodiment of the present application, image processing technology is used to binarize the phase change gradient data (>0.5 is 1, otherwise it is 0), extract the connected domain, calculate the minimum circumscribed rectangle of each connected domain, and expand the 10% safety boundary as the coverage range.

[0163] Step 604: According to the main scanning direction, with the high-roughness area as the starting point, the boundary vertices of the scanning point coverage corresponding to each area are sequentially connected to generate a dynamic cleaning path.

[0164] In step 604, the boundary vertices refer to the coordinates of the corner points of the polygonal outline of the coverage range of each scanning point, which are connected in the order of the main direction to form a complete path.

[0165] In the embodiment of the present application, a greedy algorithm is used for path optimization, and dense scanning lines (spacing 0.2 mm) are preferentially arranged in high-roughness areas. The vertices of each area are connected in sequence according to the main direction, and arc transitions (radius 0.5 mm) are used at the corners. The paths of overlapping areas are fused to avoid repeated scanning.

[0166] Here's a specific example:

[0167] During the laser cleaning process of a certain diesel engine piston ring, the system performed dynamic path planning for the top dead center area. First, based on a laser incident angle of 62 degrees (measured by a laser tracker), the main scanning direction was determined to be a spiral at a 28-degree angle with the piston ring axis. White-light interferometry measured the surface roughness distribution in this area: the ring groove had a Ra of 3.5 μm (high roughness zone) and the land had a Ra of 2.1 μm (low roughness zone). Infrared thermal imaging revealed two subregions within the ring groove (1.8 mm × 2.5 mm and 2.2 mm × 3.0 mm, respectively, calculated through temperature field inversion) with phase change gradients exceeding a threshold of 0.6. The system then expanded these subregions by a 15% safety margin to define the scanning coverage area (2.1 mm × 2.9 mm and 2.5 mm × 3.5 mm, respectively). Starting from the ring groove, the boundary vertices of each coverage area were connected sequentially along the main spiral direction, generating a dynamic cleaning path with a total length of 0.9 meters.

[0168] In the embodiment of the present application, the solution achieves precise matching of path planning and surface morphology, targeted cleaning of highly contaminated areas, and smooth optimization of motion trajectories by integrating geometric features and material state data, thereby improving cleaning uniformity and efficiency.

[0169] In order to solve the problem of inaccurate matching between the scanning frequency and the oxide layer thickness during laser cleaning and further improve the adaptability of cleaning parameters, in some embodiments, step 103: generating a target scanning frequency offset based on the oxide layer thickness distribution using a pre-trained deep learning model includes:

[0170] Step 701: Divide the piston ring surface into multiple oxide layer regions with different thickness levels according to the oxide layer thickness distribution, extract the average oxide layer thickness of each oxide layer region, and use the average oxide layer thickness as the thickness characteristic value of the corresponding region.

[0171] In step 701, the average oxide layer thickness refers to the arithmetic mean of the thickness data of all detection points in each oxide layer region, and is used to characterize the overall oxidation degree of the region.

[0172] In an embodiment of the present application, a laser-induced breakdown spectrometer is used to collect thickness data at a spacing of 0.5 mm along the cleaning path. The surface is divided into 3-5 thickness grade areas based on a clustering algorithm, and the average thickness of all measuring points within each area is calculated. For example, if the thickness value measured in a certain area is [75, 82, 79] μm, the average value is 78.7 μm.

[0173] Step 702: The thickness characteristic values ​​of all the regions are input into a pre-trained deep learning model. The deep learning model fuses the thickness characteristic values ​​of each region with the residual phase change gradient information data of the corresponding region in the dynamic cleaning path, and outputs the energy accumulation time prediction value of each region.

[0174] In step 702, the energy accumulation duration prediction value refers to the optimal action time of the laser in a specific area recommended by the deep learning model, in milliseconds (ms).

[0175] In the embodiment of the present application, a deep learning model is used, in which the input layer receives thickness feature values ​​and phase change gradient data of the corresponding area, extracts multi-scale features through a three-level encoding-decoding structure, and the output layer uses an activation function to generate a 0-10ms prediction value.

[0176] Step 703: Based on the energy accumulation time prediction value, the optimal cleaning parameter combination of the same thickness grade area in the piston ring operation history data is called to generate an initial scanning frequency offset.

[0177] In step 703, the optimal cleaning parameter combination refers to the frequency-power parameter pair with the best cleaning effect for the same thickness level area in the historical database. The initial scanning frequency offset refers to the laser scanning frequency adjustment value preliminarily determined based on the oxide layer thickness level and the historical optimal parameters. It represents the adjustment amount of the laser scanning frequency relative to the reference value (e.g., 200Hz), and the unit is Hz. A positive value indicates an increase in frequency (accelerated scanning), while a negative value indicates a decrease in frequency (decelerated scanning). For example, a value of -50Hz means adjusting the scanning frequency from 200Hz to 150Hz to increase the laser dwell time.

[0178] In an embodiment of the present application, a thickness level-parameter lookup table is established (for example, a thickness of 80-100 μm corresponds to a frequency of 150 Hz and a power of 1.2 times), and the Euclidean distance is used to screen the 10 most similar historical cases. The top 3 parameter combinations with the best cleaning effects are averaged to generate an initial offset.

[0179] Step 704: Correct the initial scanning frequency offset according to the sudden change gradient of the oxide layer thickness in adjacent areas in the dynamic cleaning path to generate an intermediate scanning frequency offset.

[0180] In step 704, adjacent areas refer to sub-areas with continuous spatial positions and different oxide layer thicknesses on the dynamic cleaning path. The division is determined by the gradient detection of the oxide layer thickness distribution map, and is usually based on a thickness mutation gradient greater than 10μm / mm as the boundary. For example: if the center distance between area A (thickness 80μm) and area B (thickness 50μm) is 2mm, the gradient is 15μm / mm, and they are determined to be adjacent mutation areas. The oxide layer thickness mutation gradient refers to the ratio of the thickness difference between adjacent areas to the distance between the areas, and is used to identify boundary areas with sharp changes in thickness. The intermediate scanning frequency offset is a transition frequency adjustment value generated to resolve the thickness mutation of adjacent areas.

[0181] In an embodiment of the present application, the thickness difference between adjacent areas is calculated and divided by the center distance (for example, area A 80 μm and area B 50 μm are 2 mm apart, and the gradient is 15 μm / mm). When the gradient is greater than 10 μm / mm, a transition frequency is inserted in the boundary area (for example, a linear gradient from 150 Hz to 200 Hz) to generate a smooth intermediate offset curve.

[0182] Step 705: superimpose the initial scanning frequency offset and the intermediate scanning frequency offset according to the spatial coordinate position by weighted summation to generate a target scanning frequency offset.

[0183] In step 705 , weighted summation refers to a parameter fusion method that assigns weights according to regional importance (weight 0.7 for the high oxide layer region and weight 0.3 for the transition region).

[0184] In the embodiment of the present application, a position-related weight is assigned to the initial offset (0.6 in the high oxidation area and 0.4 in other areas), and a gradient-related weight is assigned to the intermediate offset (0.5 in the mutation area and 0.2 in the flat area), and point-by-point weighted fusion is performed according to the weight matrix.

[0185] Here's a specific example:

[0186] During the laser cleaning process of a certain diesel engine piston ring, the system optimized the scanning frequency for the top dead center region (coordinates X20-Y45 to X25-Y50). Laser-induced breakdown spectroscopy was first used to measure the oxide layer thickness distribution in this region and categorize it into three levels: 1) 85μm thickness in the center of the groove (level 3); 2) 65μm thickness in the transition region (level 2); and 3) 45μm thickness in the land region (level 1). A deep learning model was trained to evaluate the thickness characteristics of each region (85μm, 65μm, and 45μm) and the corresponding residual phase transition gradient data (0.85, 0.60, and 0.35). The model output predicted energy accumulation durations of 7.2ms, 5.5ms, and 3.8ms, respectively. A historical database query was used to match the optimal frequency parameters for each thickness level: level 3 corresponds to 140Hz (initial offset -60Hz), level 2 corresponds to 170Hz (-30Hz), and level 1 corresponds to 220Hz (+20Hz). The thickness gradient from the annular groove to the transition zone was detected to be 10 μm / mm (85 μm to 65 μm, with a center distance of 2 mm). An intermediate offset with a linear transition from -60 Hz to -30 Hz was generated within the 2 mm boundary. Finally, through weighted fusion (annular groove weight of 0.7, transition zone weight of 0.3), the target scanning frequency offset was generated: -55 Hz in the annular groove (actual frequency 145 Hz), and -40 Hz in the transition zone (160 Hz).

[0187] In the embodiment of the present application, the solution realizes adaptive matching of scanning frequency and oxide layer thickness, smooth transition processing of thickness mutation area, balanced control of cleaning efficiency and quality through collaborative optimization of deep learning and historical data, and improves the process stability of laser cleaning.

[0188] Figure 2 This is a schematic diagram of the structure of an engine piston ring intelligent cleaning control system provided in an embodiment of the present application, as shown in FIG. Figure 2 As shown, the system includes:

[0189] An acquisition module 21 is used to obtain residual phase change gradient information data and the incident angle value of the laser beam in the laser thermodynamic action area on the surface of the engine piston ring, as well as the roughness and oxide layer thickness distribution of the piston ring surface;

[0190] A construction module 22 is configured to construct a dynamic cleaning path for the piston ring according to the incident angle value, the roughness, and the residue phase change gradient information;

[0191] a generating module 23 for generating a target scanning frequency offset using a pre-trained deep learning model based on the oxide layer thickness distribution during the cleaning of the piston ring according to the dynamic cleaning path;

[0192] The adjustment module 24 is used to generate laser energy distribution optimization parameters based on the target scanning frequency offset. The laser energy distribution optimization parameters are used to adjust the laser power density and laser focus depth of the laser beam in real time during the cleaning process to achieve accurate removal of residues on the piston ring surface and effective protection of the base material.

[0193] Figure 2 The intelligent cleaning control system for engine piston rings can be executed Figure 1 The implementation principles and technical effects of the engine piston ring intelligent cleaning control method described in the illustrated embodiment are not further elaborated. The specific manner in which the various modules and units perform operations in the engine piston ring intelligent cleaning control system described in the aforementioned embodiment have been described in detail in the relevant embodiments of the method and will not be further elaborated here.

[0194] In one possible design, Figure 2 An engine piston ring intelligent cleaning control system of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0195] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0196] The processing component 32 is as follows Figure 1 The embodiment provides an intelligent cleaning control method for an engine piston ring.

[0197] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0198] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random-access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0199] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0200] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0201] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0202] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0203] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 An engine piston ring intelligent cleaning control method according to the illustrated embodiment.

[0204] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0205] The device embodiments described above are merely illustrative. 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, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0206] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0207] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent cleaning control method for an engine piston ring, characterized in that: include: Obtaining residual phase change gradient information data and laser beam incident angle value in the laser thermodynamic action area on the surface of the engine piston ring, as well as the roughness and oxide layer thickness distribution of the piston ring surface; Constructing a dynamic cleaning path for the piston ring according to the incident angle value, the roughness and the residue phase change gradient information; During the cleaning of the piston ring according to the dynamic cleaning path, a target scanning frequency offset is generated using a pre-trained deep learning model based on the oxide layer thickness distribution; Laser energy distribution optimization parameters are generated according to the target scanning frequency offset. The laser energy distribution optimization parameters are used to adjust the laser power density and laser focus depth of the laser beam in real time during the cleaning process to achieve accurate removal of residues on the piston ring surface and effective protection of the base material.

2. The method according to claim 1, characterized in that Generating the laser energy distribution optimization parameter according to the scanning frequency offset includes: coupling the residue phase change gradient information data with the roughness to generate a laser power density adjustment parameter; Generate a laser focus depth adjustment parameter based on the residual phase change gradient information data and the thermal cycle attenuation index in the piston ring operation history data, combined with the local structural morphology characteristics of the piston ring; According to the scanning frequency offset, the laser power density adjustment parameter and the laser focus depth adjustment parameter are coupled to generate a laser energy distribution optimization parameter.

3. The method according to claim 2, characterized in that The step of coupling the laser power density adjustment parameter and the laser focus depth adjustment parameter according to the scanning frequency offset to generate the laser energy distribution optimization parameter includes: According to the laser power density adjustment parameters for all areas of the piston ring surface, a numerical distribution of the laser power density adjustment parameters is constructed; according to the laser focus depth adjustment parameters for all areas of the piston ring surface, a spatial distribution of the laser focus depth adjustment parameters is constructed; Determining a weight distribution relationship between the laser power density adjustment parameter and the laser focus depth adjustment parameter at each laser scanning point in a dynamic cleaning path based on the numerical distribution of the laser power density adjustment parameter and the spatial distribution of the laser focus depth adjustment parameter; According to the weight distribution relationship, the laser power density adjustment parameter and the adjustment parameter are superimposed and calculated point by point to generate an initial energy distribution optimization parameter; Based on the scanning frequency offset, the initial energy distribution optimization parameter is mapped to the laser scanning timing sequence corresponding to the dynamic cleaning path to generate an intermediate energy distribution optimization parameter associated with the timing sequence; According to the energy gradient change of the intermediate energy distribution optimization parameter, combined with the remaining residue phase transition gradient information data, the weight distribution relationship is reversely corrected; Based on the inversely corrected weight distribution relationship, the laser energy distribution optimization parameters are generated.

4. The method according to claim 3, characterized in that The reverse correction of the weight distribution relationship based on the energy gradient change of the intermediate energy distribution optimization parameter and the remaining residue phase transition gradient information data includes: Calculating an energy distribution difference based on the energy cumulative gradient difference between adjacent scanning points in the intermediate energy distribution optimization parameter; Determine the ratio of the energy distribution difference to the remaining residue phase transition gradient information data as a weight adjustment factor; The weight distribution relationship is modified according to the weight adjustment factor to update the superposition weight of the laser power density adjustment parameter and the laser focus depth adjustment parameter.

5. The method according to claim 2, characterized in that The laser focus depth adjustment parameter is generated based on the residual phase change gradient information data and the thermal cycle attenuation index in the piston ring operation history data, combined with the local structural morphology characteristics of the piston ring, including: Determining phase change rate differences in different regions based on the residue phase change gradient information data; Determine the energy absorption attenuation ratio of the thermal cycle attenuation index at each position on the piston ring surface in the piston ring operation history data based on the phase change rate difference and the thermal expansion distortion characteristics of the laser thermodynamic action area; The energy absorption attenuation ratio is superimposed with the thermal stress accumulation at the corresponding position in the piston ring operation history data to generate a local thermodynamic influence factor; The structural morphology weight parameter is calculated based on the product relationship between the roughness and the oxide layer thickness distribution in the local structural morphology characteristics of the piston ring; Multiplying the local thermodynamic influence factor and the structural morphology weight parameter point by point according to the spatial coordinate position of the piston ring surface to generate a coupling correction parameter; Based on the coupling correction parameters, the basic parameters of the laser focus depth are nonlinearly optimized to generate laser focus depth adjustment parameters.

6. The method according to claim 1, characterized in that The step of constructing a dynamic cleaning path for the piston ring according to the incident angle value, the roughness, and the residue phase change gradient information includes: Based on the incident angle value, determining a main scanning direction of the laser beam on the piston ring surface, wherein the main scanning direction is parallel to a central axis of the incident angle value; According to the roughness distribution value, the piston ring surface is divided into a high roughness area and a low roughness area; For each divided region, searching for a sub-region in which the gradient change rate of the residual phase change gradient information data of the region exceeds a set gradient threshold, and determining a scanning point coverage range including the sub-region; According to the main scanning direction, with the high-roughness area as the starting point, the boundary vertices of the scanning point coverage range corresponding to each area are connected in sequence to generate a dynamic cleaning path.

7. The method according to claim 1, characterized in that The method of generating a target scanning frequency offset based on the oxide layer thickness distribution using a pre-trained deep learning model includes: Dividing the piston ring surface into a plurality of oxide layer regions with different thickness levels according to the oxide layer thickness distribution, extracting the average oxide layer thickness of each oxide layer region, and using the average oxide layer thickness as a thickness characteristic value of the corresponding region; Inputting the thickness characteristic values ​​of all the regions into a pre-trained deep learning model, the deep learning model performs feature fusion on the thickness characteristic value of each region with the residual phase change gradient information data of the corresponding region in the dynamic cleaning path, and outputs a predicted value of the energy accumulation time of each region; Based on the energy accumulation time prediction value, the optimal cleaning parameter combination for the same thickness grade area in the piston ring operation history data is called to generate an initial scanning frequency offset; Correcting the initial scanning frequency offset according to a sudden change gradient of oxide layer thickness in adjacent areas in the dynamic cleaning path to generate an intermediate scanning frequency offset; The initial scanning frequency offset and the intermediate scanning frequency offset are superimposed according to the spatial coordinate position in a weighted summation manner to generate a target scanning frequency offset.

8. An intelligent cleaning control system for an engine piston ring, characterized in that: include: An acquisition module is used to obtain the residual phase change gradient information data and the incident angle value of the laser beam in the laser thermodynamic action area on the surface of the engine piston ring, as well as the roughness and oxide layer thickness distribution of the piston ring surface; A construction module, configured to construct a dynamic cleaning path for the piston ring according to the incident angle value, the roughness, and the residue phase change gradient information; a generation module, configured to generate a target scanning frequency offset using a pre-trained deep learning model based on the oxide layer thickness distribution during cleaning of the piston ring according to the dynamic cleaning path; An adjustment module is used to generate laser energy distribution optimization parameters based on the target scanning frequency offset. The laser energy distribution optimization parameters are used to adjust the laser power density and laser focus depth of the laser beam in real time during the cleaning process to achieve accurate removal of residues on the piston ring surface and effective protection of the base material.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an engine piston ring intelligent cleaning control method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, an engine piston ring intelligent cleaning control method according to any one of claims 1 to 7 is implemented.

Images