Method and system for predicting wear resistance of nickel-based composite coating based on machine learning

Abstract

The invention relates to the technical field of performance prediction, in particular to a nickel-based composite coating wear resistance prediction method and system based on machine learning. Multi-source data such as workpiece surface features, coating microstructures, process parameters and future working conditions are integrated, and an intelligent prediction model based on a multi-layer perceptron is constructed; according to the method, accurate prediction and early warning of the wear-resistant life of the nickel-based composite coating are realized, a complex nonlinear relation can be learned from historical data, weak links of the coating performance can be recognized in advance, the predictability and accuracy of quality control are improved, the equipment maintenance cost and the shutdown risk caused by coating failure are remarkably reduced, and the service life of the nickel-based composite coating is prolonged. And a reliable decision basis is provided for optimizing a spraying process and prolonging the service life of a workpiece.

Description

Technical Field

[0001] This invention relates to the field of performance prediction technology, and in particular to a method and system for predicting the wear resistance of nickel-based composite coatings based on machine learning. Background Technology

[0002] In modern industry, critical equipment such as aircraft engine blades, heavy-duty gas turbine components, and oil drilling tools operate under extreme conditions of high temperature, high pressure, and intense wear. Their surface properties directly determine the service life and operational reliability of the entire equipment. To improve the wear and corrosion resistance of these critical components, thermal spraying technology, especially high-speed oxygen fuel and atmospheric plasma spraying technology, is widely used to prepare high-performance nickel-based composite coatings. These coatings, by introducing high-hardness carbide phases, can significantly improve the wear resistance of the workpiece surface.

[0003] However, the wear life of nickel-based composite coatings is not a constant value; it is influenced by multiple factors within a complex system. Specifically, its lifespan mainly depends on the following aspects: Workpiece substrate characteristics: The geometry, surface roughness, preheating temperature, and material properties of the workpiece to be coated directly affect the bonding strength between the coating and the substrate, forming the basis of the coating system's durability; Coating microstructure: The microstructure characteristics of the coating, such as porosity, crack density, oxide content, proportion of unmelted particles, and the distribution and degradation degree of carbides, are the fundamental reasons determining its macroscopic mechanical properties and wear behavior; Spraying process parameters: During the spraying process, even small fluctuations in dozens of process parameters, such as spraying distance, powder feed rate, fuel flow rate, and power, can significantly alter the droplet flight state, flattening behavior, and solidification process, ultimately leading to significant differences in the coating's microstructure and performance. Currently, the industry faces the following major technical bottlenecks in evaluating the wear resistance and lifespan management of coatings: Traditional methods mainly rely on bench tests or destructive metallographic analysis of samples after spraying. This method is costly, time-consuming, and cannot test every actual workpiece. Lifespan prediction largely depends on the personal experience of engineers, lacking quantitative and scientific basis, resulting in poor accuracy and consistency. Due to the extremely complex and highly nonlinear mapping relationship between process parameters, microstructure, and final performance, existing technologies struggle to accurately reveal its inherent laws. Process optimization often requires numerous trial-and-error experiments, which is inefficient and makes it difficult to guarantee the stability and reproducibility of coating performance. Existing technologies cannot accurately predict the wear resistance life of coatings under specific operating conditions before they are put into service. This leads to equipment maintenance typically being performed at fixed intervals or after significant failure, failing to achieve condition-based predictive maintenance. Unplanned downtime due to unexpected coating failure can cause significant economic losses and safety risks.

[0004] Therefore, there is an urgent need for an intelligent method that can systematically integrate multi-source data and achieve high-precision forward-looking prediction. Summary of the Invention

[0005] To overcome the shortcomings and deficiencies of existing technologies, this invention provides a method and system for predicting the wear resistance of nickel-based composite coatings based on machine learning. This application integrates multi-source data such as workpiece surface features, coating microstructure, process parameters, and future operating conditions, and constructs an intelligent prediction model based on a multilayer perceptron. This enables accurate prediction and early warning of the wear life of nickel-based composite coatings. This method can learn complex nonlinear relationships from historical data and identify weak links in coating performance in advance. It not only improves the predictability and accuracy of quality control, but also significantly reduces equipment maintenance costs and downtime risks caused by coating failure. This provides a reliable decision-making basis for optimizing the spraying process and extending the service life of workpieces.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for predicting the wear resistance of nickel-based composite coatings based on machine learning, comprising the following steps: S100: Acquire the surface condition of the workpiece and the condition of the nickel-based composite coating, and at the same time acquire the environmental conditions during the coating process and the future use of the corresponding workpiece; S200. Workpiece coating adhesion analysis is performed by examining the workpiece surface flatness and the distribution of hard phases in the nickel-based composite coating. S300. Construct a machine learning model to predict the wear resistance of the workpiece coating by analyzing the workpiece coating connection results, the average microhardness in the nickel-based composite coating, and the environmental conditions during the coating process. S400: Obtain the future usage of the corresponding workpiece and the prediction results of the workpiece coating wear resistance effect to predict the workpiece coating wear life. S500 provides early warning of coating performance based on the predicted wear life of the workpiece coating.

[0007] In one implementation of the present invention, step S100 includes the following specific contents: Step 110: Acquire a three-dimensional image of the workpiece surface and the condition of the nickel-based composite coating using the corresponding sensor. The condition of the nickel-based composite coating includes: Hard phase distribution: The wear resistance of nickel-based composite coatings mainly comes from the uniformly dispersed hard phase; the hard phase plays the role of bearing load and resisting abrasive wear; its uniformity of distribution is the core indicator of coating quality; average microhardness: This is a direct measure of the coating's resistance to plastic deformation and indentation. Step 120: Obtain the temperature, humidity, and particle flow rate during the thermal spraying process using sensors. Temperature affects the flattening behavior of particles and their thermal interaction with the substrate. Particle flow rate determines the kinetic energy of particles when they collide with the substrate, affecting the density and bonding strength of the coating. Step 130: Obtain the corresponding workpiece's friction load, usage frequency, and friction speed from the workpiece's usage log.

[0008] In one implementation of the present invention, the workpiece coating adhesion analysis in S200 includes the following specific steps: Step 210: Obtain a three-dimensional image of the workpiece surface, and then obtain the distribution of pits and the corresponding surface roughness. Simultaneously, obtain the distribution of the hard phase in the coating. The surface roughness is evaluated as follows: the roughness is obtained by the ratio of the average absolute value of the height of each point on the workpiece surface relative to the reference surface to the safe height. Roughness anomalies are obtained by the standard deviation of the roughness from the safe roughness range. The average absolute value of the height of each point relative to the reference surface is equivalent to the arithmetic mean height in the roughness parameters, providing an intuitive and stable three-dimensional roughness measure that comprehensively reflects the overall surface undulations, rather than a single contour line. Normalization using the ratio to the safe height allows the evaluation standard to adapt to different process requirements or different workpiece models, enhancing the method's universality and comparability. Instead of obtaining a single roughness value, the method quantifies the degree of anomaly by calculating the standard deviation from the safe roughness range. Step 220: Obtain the size of the pits on the workpiece surface and the size deviation of the hard phase particles to obtain the size matching anomaly. Obtain the hard phase distribution uniformity anomaly by analyzing the distribution of the hard phase in the coating. The hard phase distribution uniformity anomaly is obtained by: obtaining the hard phase distribution density at each location in the coating, obtaining the standard deviation of the hard phase distribution density at each location, and obtaining the hard phase distribution uniformity anomaly. If the hard phase particles are much larger than the surface pits, they will not be able to embed; if they are much smaller, the embedding effect will be poor. Ideally, the particle size and the pit size have an optimal matching range. Step 230: Obtain the distribution of pits on the workpiece surface, set the standard deviation of pit size as pit size anomaly, and at the same time obtain the spacing between pits, and obtain the standard deviation of the spacing to obtain pit distribution anomaly. Step 240: Construct a data sequence in order of the acquired roughness anomalies, size matching anomalies, hard phase distribution uniformity anomalies, pit size anomalies, and pit distribution anomalies, and set the constructed sequence as the workpiece coating connection analysis sequence; integrate multiple indicators of different dimensions that describe the surface and coating state into a unified, structured data sequence.

[0009] In one implementation of the present invention, the prediction of the wear resistance effect of the workpiece coating in S300 specifically includes the following: Step 310: Obtain the temperature, humidity, and particle flow rate during historical thermal spraying processes; simultaneously, obtain the average microhardness of historical workpiece coating adhesion analysis sequences, the average microhardness of nickel-based composite coatings, and the wear life of historical workpiece coatings under standard friction frequency, standard friction speed, and standard friction force. Construct a multilayer perceptron model with the input of temperature, humidity, particle flow rate, average microhardness of nickel-based composite coatings, and workpiece coating adhesion analysis sequences during historical thermal spraying processes, and the output of wear life of workpiece coatings under standard friction frequency, standard friction speed, and standard friction force. Here, standard friction frequency, standard friction speed, and standard friction force are the standard output features of the testing equipment. The thermal spraying process is a complex physicochemical process involving droplet spreading, solidification, and phase transitions, which is difficult to accurately describe using a purely physical model. The multilayer perceptron model is very suitable for learning such complex nonlinear mapping relationships from a large amount of historical data. Step 320: Import the temperature, humidity, particle flow rate, average microhardness of the nickel-based composite coating, and workpiece coating adhesion analysis sequence obtained in the real-time thermal spraying process into the constructed multilayer perceptron model, and output the predicted wear life of the workpiece coating under standard friction frequency, standard friction speed, and standard friction force.

[0010] In one implementation of the present invention, the prediction of the wear resistance life of the workpiece coating in S400 includes the following specific contents: Step 410: Obtain the future usage friction load, usage frequency, and friction speed of the corresponding workpiece. Divide the obtained usage friction load, usage frequency, and friction speed by the corresponding standard values ​​and then perform a weighted sum to obtain the usage anomaly prediction result. The corresponding standard values ​​are the output standard characteristics of the test equipment. Multiply the usage anomaly prediction result by the wear resistance influence coefficient to obtain the wear resistance anomaly prediction result. Step 420: Divide the wear life prediction result of the workpiece coating under standard friction frequency, standard friction speed and standard friction force by the wear anomaly prediction result.

[0011] In one implementation of the present invention, the coating performance warning in S500 includes the following specific contents: The predicted wear life of the workpiece coating is obtained and compared with the corresponding required wear life value. If the predicted wear life is greater than or equal to the required wear life value, it indicates that the coating performance meets the workpiece's usage requirements. If the predicted wear life is less than the required wear life value, it indicates that the coating performance does not meet the workpiece's usage requirements, and a coating warning needs to be issued to remind maintenance personnel to replace it.

[0012] Secondly, the present invention also provides a machine learning-based system for predicting the wear resistance of nickel-based composite coatings, comprising: The data acquisition module acquires information about the workpiece surface condition and the nickel-based composite coating condition, as well as the environmental conditions during the coating process and the future use of the corresponding workpiece. The coating adhesion analysis module analyzes the coating adhesion of the workpiece by considering the workpiece surface flatness and the distribution of hard phases in the nickel-based composite coating. The machine learning analysis module constructs a machine learning model to predict the wear resistance of the workpiece coating by analyzing the workpiece coating connection results, the average microhardness in the case of nickel-based composite coating, and the environmental conditions during the coating process. The life prediction module obtains the future usage of the corresponding workpiece and the prediction results of the workpiece coating wear resistance effect to predict the wear resistance life of the workpiece coating. The performance early warning module provides early warnings about coating performance based on the predicted wear life of the workpiece coating.

[0013] Thirdly, the present invention provides an electronic device comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes a machine learning-based method for predicting the wear resistance of nickel-based composite coatings by calling the computer program stored in the memory.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform a machine learning-based method for predicting the wear resistance of nickel-based composite coatings.

[0015] Compared with the prior art, the present invention has the following advantages and beneficial effects: By integrating multi-source data such as workpiece surface features, coating microstructure, process parameters, and future operating conditions, and constructing an intelligent prediction model based on a multilayer perceptron, this method achieves accurate prediction and early warning of the wear resistance life of nickel-based composite coatings. This method can learn complex nonlinear relationships from historical data and identify weak links in coating performance in advance. It not only improves the predictability and accuracy of quality control, but also significantly reduces equipment maintenance costs and downtime risks caused by coating failure, providing a reliable decision-making basis for optimizing the spraying process and extending the service life of workpieces. Attached Figure Description

[0016] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the overall process structure of an embodiment of the method of the present invention; Figure 2 This is a schematic diagram of the process structure of step S200 in the embodiment of the method of the present invention; Figure 3 This is the construction result of the machine learning model in the embodiment of the method of the present invention; Figure 4 This is a schematic diagram of the module composition structure of an embodiment of the system of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0018] Please see Figures 1 to 3 This invention provides a method for predicting the wear resistance of nickel-based composite coatings based on machine learning, specifically including the following steps: S100: Acquire the surface condition of the workpiece and the condition of the nickel-based composite coating, and at the same time acquire the environmental conditions during the coating process and the future use of the corresponding workpiece; In this embodiment, S100 includes the following specific contents: Step 110: Acquire a three-dimensional image of the workpiece surface and the condition of the nickel-based composite coating using the corresponding sensor. The condition of the nickel-based composite coating includes: Hard phase distribution: The wear resistance of nickel-based composite coatings mainly comes from the uniformly dispersed hard phase (such as tungsten carbide, chromium carbide, etc.). The hard phase plays the role of bearing load and resisting abrasive wear. Its uniformity of distribution is the core indicator of coating quality. Uneven distribution will lead to premature wear in some areas. Average microhardness: This is a direct measure of the coating's resistance to plastic deformation and indentation. Higher average hardness is usually associated with better wear resistance, but it must be combined with toughness. Purely high hardness and brittle coatings are prone to peeling. Step 120: Obtain the temperature, humidity, and particle flow rate during the thermal spraying process using sensors. Temperature affects the flattening behavior of particles and their thermal interaction with the substrate. Particle flow rate determines the kinetic energy of particles when they impact the substrate, affecting the density and bonding strength of the coating. These parameters together determine the microstructure of the coating, such as porosity, oxide content, and residual stress, thus directly affecting wear resistance. Step 130: Obtain the corresponding workpiece's frictional load, usage frequency, and friction speed from the workpiece's usage log, and then analyze the workpiece's wear load. S200. Workpiece coating adhesion analysis is performed by examining the workpiece surface flatness and the distribution of hard phases in the nickel-based composite coating. In this embodiment, the workpiece coating adhesion analysis in S200 includes the following specific steps: Step 210: Obtain a 3D image of the workpiece surface, thereby obtaining the distribution of pits and the corresponding surface roughness. Simultaneously, obtain the distribution of the hard phase in the coating. The surface roughness is evaluated as follows: the roughness is obtained by the ratio of the average absolute value of the height of each point on the workpiece surface relative to the reference surface to the safe height. Roughness anomalies are obtained by the standard deviation of the roughness from the safe roughness range. The average absolute value of the height of each point relative to the reference surface is equivalent to the arithmetic mean height in the roughness parameters, providing an intuitive and stable 3D roughness measure. The roughness measure comprehensively reflects the overall undulation of the surface, rather than a single contour line. By normalizing the ratio to the safety height, the evaluation standard can be adapted to different process requirements or different workpiece models, enhancing the universality and comparability of the method. It is not satisfied with obtaining a roughness value, but quantifies its degree of abnormality by calculating the standard deviation from the safety roughness range. This step is based on surface metrology and industrial quality control standards. In thermal spraying, the roughness of the substrate surface directly affects the bonding strength of the coating. Moderate roughness can provide mechanical interlocking, but excessive or insufficient roughness will lead to a decrease in bonding strength or coating defects. Step 220: Obtain the dimensions of the pits on the workpiece surface and the size deviation of the hard phase particles to obtain dimensional mismatch anomalies. Obtain the uniformity of hard phase distribution anomalies by analyzing the distribution of the hard phase in the coating. The method for obtaining the uniformity of hard phase distribution anomalies is as follows: obtain the hard phase distribution density at each location in the coating, and obtain the standard deviation of the hard phase distribution density at each location to obtain the uniformity of hard phase distribution anomalies. If the hard phase particles are much larger than the surface pits, they will not be able to embed; if they are much smaller, the embedding effect will be poor. Ideally, the particle size and pit size have an optimal matching range. Correlating the two independent measurements (pit size, particle size) forms a quality index (matching degree) with clear physical meaning, which can predict in advance whether the coating is prone to detachment due to poor mechanical embedding. The coating performance (such as hardness, wear resistance) requires its composition distribution to be as uniform as possible. Too little local hard phase will lead to soft spots and easy wear; too much may cause stress concentration or peeling.

[0019] Step 230: Obtain the distribution of pits on the workpiece surface, set the standard deviation of pit size as pit size anomaly, and at the same time obtain the spacing between pits, and obtain the standard deviation of the spacing to obtain pit distribution anomaly. Step 240: Construct a data sequence from the acquired roughness anomalies, size matching anomalies, hard phase distribution uniformity anomalies, pit size anomalies, and pit distribution anomalies in order, and set the constructed sequence as the workpiece coating adhesion analysis sequence; integrate multiple indicators of different dimensions describing the surface and coating state into a unified, structured data sequence (or feature vector); this sequence highly condenses the key pre-process features that affect the coating adhesion strength and final performance, providing high-quality input data for subsequent prediction models; S300. Construct a machine learning model to predict the wear resistance of the workpiece coating by analyzing the workpiece coating connection results, the average microhardness in the nickel-based composite coating, and the environmental conditions during the coating process. In this embodiment, the prediction of the wear resistance effect of the workpiece coating in S300 specifically includes the following: Step 310: Obtain the temperature, humidity, and particle flow rate during historical thermal spraying processes; simultaneously, obtain the average microhardness of historical workpiece coating adhesion analysis sequences, the average microhardness of nickel-based composite coatings, and the wear life of historical workpiece coatings under standard friction frequency, standard friction speed, and standard friction force. Construct a multilayer perceptron model with the input of temperature, humidity, particle flow rate, average microhardness of nickel-based composite coatings, and workpiece coating adhesion analysis sequences during historical thermal spraying processes, and the output of wear life of workpiece coatings under standard friction frequency, standard friction speed, and standard friction force. Here, standard friction frequency, standard friction speed, and standard friction force are the standard output features of the testing equipment. The thermal spraying process is a complex physicochemical process involving droplet spreading, solidification, and phase transitions, which is difficult to accurately describe using a purely physical model. The multilayer perceptron model is very suitable for learning such complex nonlinear mapping relationships from a large amount of historical data. The specific steps of the multilayer perceptron model are as follows: Input layer construction: directly corresponding to 9 input features, this layer does not perform any calculations, but only receives the data and passes it to the first hidden layer. Therefore, this layer has 9 neurons. The hidden layers are preferably set to three layers. The first hidden layer, as the information receiving and preliminary processing layer, should have a relatively large number of neurons to capture enough initial feature combinations; it can be set to a large number, such as 128 neurons. The second hidden layer further refines and compresses the output of the first layer; the number of neurons is halved, set to 64 neurons. The third hidden layer performs a final high-precision feature extraction before the final decision; the number of neurons is halved again, set to 32 neurons. Through multi-layer nonlinear transformation, the input features are abstracted and combined layer by layer to learn the extremely complex mapping relationship between "process parameters" and "coating characteristics" and "wear resistance life". All three hidden layers use the ReLU (Modified Linear Unit) activation function. The output layer has only one neuron, corresponding to the predicted wear life; activation function: since it is a regression prediction task (outputting continuous values), a linear activation function (i.e., without non-linear transformation) is usually used, or to ensure that the output is positive, functions such as Softplus can be used; Loss Function: MSE is the standard loss function for regression tasks. It amplifies the penalty for large prediction errors, making the model training process focus on reducing predictions that deviate significantly from the true value. This is crucial for industrial quality prediction. Optimizer: Adam combines the advantages of other optimizers, automatically adjusting the learning rate. In most cases, it converges quickly and stably, making it the preferred optimizer in deep learning. Key Hyperparameter Settings: Learning Rate: This is the most important tuning knob. The initial value can be set to 0.001. If the loss decreases too slowly during training, it can be increased appropriately. If the loss oscillates without decreasing, it can be decreased appropriately. Batch Size: Choose a suitable value based on the amount of data and computing resources, such as 32 or 64. Number of Iteration Cycles: Set a large value (such as 1000) and use it in conjunction with early stopping techniques. That is, when the model's performance on the validation set no longer improves for several consecutive cycles, training is automatically stopped to prevent overfitting. Training process: Initialize all weights of the hidden and output layers to small random numbers, and initialize biases to 0 or small constants; take a batch (e.g., 32) of historical workpiece data; the data flows into the input layer (6 neurons), passes through the first hidden layer (128 ReLU neurons) for calculation, then flows into the second hidden layer (64 ReLU neurons), then the third hidden layer (32 ReLU neurons), and finally reaches the output layer (1 linear neuron) to obtain the predicted wear life of this batch of workpieces; compare the model's predicted value with the actual, known wear life of these 32 workpieces, and calculate a total error value using the MSE formula; using the chain rule of calculus, starting from the output layer, calculate the gradient of the loss function with respect to each weight and each bias in the network in reverse; the Adam optimizer intelligently updates all weights and biases in the network based on the calculated gradients and the preset learning rate (0.001), while also considering historical gradient information; repeat the training steps for the next batch of data, and training stops when the model performs best on the validation dataset; the final multilayer perceptron model is obtained; Step 320: Import the temperature, humidity, particle flow rate, average microhardness of the nickel-based composite coating, and workpiece coating connection analysis sequence obtained in the real-time thermal spraying process into the constructed multilayer perceptron model, and output the predicted wear life of the workpiece coating under standard friction frequency, standard friction speed, and standard friction force. This can help analyze whether the coating is qualified for production. Even if it cannot be used by this workpiece after passing S400-S500, it can be used by other workpieces if it is qualified. S400: Obtain the future usage of the corresponding workpiece and the prediction results of the workpiece coating wear resistance effect to predict the workpiece coating wear life. In this embodiment, the prediction of the wear resistance life of the workpiece coating in S400 includes the following specific contents: Step 410: Obtain the future usage friction load, usage frequency, and friction speed of the corresponding workpiece. Divide the obtained usage friction load, usage frequency, and friction speed by the corresponding standard values ​​and then perform a weighted sum to obtain the usage anomaly prediction result. The corresponding standard values ​​are the output standard characteristics of the test equipment. Multiply the usage anomaly prediction result by the wear resistance influence coefficient to obtain the wear resistance anomaly prediction result. Step 420: Divide the wear life of the workpiece coating under standard friction frequency, standard friction speed and standard friction force predicted by the multilayer perceptron model into the wear anomaly prediction result to obtain the wear life prediction result of the workpiece coating. The weighting weight and wear influence coefficient here are obtained by fitting historical data. S500: Early warning of coating performance based on the predicted wear life of workpiece coating; In this embodiment, the coating performance warning in S500 includes the following specific contents: The predicted wear life of the workpiece coating is obtained and compared with the corresponding required wear life value. If the predicted wear life is greater than or equal to the required wear life value, it indicates that the coating performance meets the workpiece's usage requirements. If the predicted wear life is less than the required wear life value, it indicates that the coating performance does not meet the workpiece's usage requirements, and a coating warning needs to be issued to remind maintenance personnel to replace it.

[0020] The above embodiments have the following advantages: by integrating multi-source data such as workpiece surface features, coating microstructure, process parameters and future operating conditions, and constructing an intelligent prediction model based on a multilayer perceptron, accurate prediction and early warning of the wear resistance life of nickel-based composite coatings are achieved. This method can learn complex nonlinear relationships from historical data and identify weak links in coating performance in advance. It not only improves the predictability and accuracy of quality control, but also significantly reduces equipment maintenance costs and downtime risks caused by coating failure, providing a reliable decision-making basis for optimizing the spraying process and extending the service life of workpieces.

[0021] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of the machine learning-based nickel-based composite coating wear resistance prediction system provided in an embodiment of the present invention, including: The data acquisition module acquires information about the workpiece surface condition and the nickel-based composite coating condition, as well as the environmental conditions during the coating process and the future use of the corresponding workpiece. The coating adhesion analysis module analyzes the coating adhesion of the workpiece by considering the workpiece surface flatness and the distribution of hard phases in the nickel-based composite coating. The machine learning analysis module constructs a machine learning model to predict the wear resistance of the workpiece coating by analyzing the workpiece coating connection results, the average microhardness in the case of nickel-based composite coating, and the environmental conditions during the coating process. The life prediction module obtains the future usage of the corresponding workpiece and the prediction results of the workpiece coating wear resistance effect to predict the wear resistance life of the workpiece coating. The performance early warning module provides early warnings about coating performance based on the predicted wear life of the workpiece coating.

[0022] The parameters and steps for each unit module to achieve the corresponding functions in the machine learning-based nickel-based composite coating wear resistance prediction system of the present invention can be referred to the parameters and steps in the embodiments of the machine learning-based nickel-based composite coating wear resistance prediction method, and will not be repeated here.

[0023] Embodiments of the present invention also provide an electronic device, including a memory, a processor, and a communication bus; the memory and the processor are connected via the communication bus. The memory stores a machine learning-based method for predicting the wear resistance of nickel-based composite coatings, which can be loaded and executed by the processor 320 as provided in the above embodiments.

[0024] The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory 310 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing the machine learning-based nickel-based composite coating wear resistance prediction method provided in the above embodiments. The data storage area may store data involved in the machine learning-based nickel-based composite coating wear resistance prediction method provided in the above embodiments.

[0025] The processor may include one or more processing cores. The processor executes instructions, programs, code sets, or instruction sets stored in memory, and calls data stored in memory to perform various functions and process data according to the present invention. The processor may be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), controller, microcontroller, and microprocessor. It is understood that, for different devices, the electronic devices used to implement the functions of the processor 320 described above may also be other types, and the embodiments of the present invention do not specifically limit this.

[0026] A communication bus may include a pathway for transmitting information between the aforementioned components. The communication bus 330 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Communication buses can be categorized as address buses, data buses, control buses, etc.

[0027] This invention provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described in the above embodiments, a machine learning-based method for predicting the wear resistance of nickel-based composite coatings.

[0028] In this embodiment of the invention, the computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. The computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, the computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), spoofing random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory stick, floppy disk, optical disk, magnetic disk, mechanical encoding device, or any combination thereof.

[0029] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0030] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to the technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this invention.

Claims

1. A method for predicting the wear resistance of nickel-based composite coatings based on machine learning, characterized in that, Includes the following steps: S100: Acquire the surface condition of the workpiece and the condition of the nickel-based composite coating, and at the same time acquire the environmental conditions during the coating process and the future use of the corresponding workpiece; S200. Workpiece coating adhesion analysis is performed by examining the workpiece surface flatness and the distribution of hard phases in the nickel-based composite coating. S300. Construct a machine learning model to predict the wear resistance of the workpiece coating by analyzing the workpiece coating connection results, the average microhardness in the nickel-based composite coating, and the environmental conditions during the coating process. S400: Obtain the future usage of the corresponding workpiece and the prediction results of the workpiece coating wear resistance effect to predict the workpiece coating wear life. S500 provides early warning of coating performance based on the predicted wear life of the workpiece coating.

2. The method for predicting the wear resistance of nickel-based composite coatings based on machine learning according to claim 1, characterized in that, S100 includes the following specific contents: Step 110: Obtain a three-dimensional image of the workpiece surface and the condition of the nickel-based composite coating through the corresponding sensor. The condition of the nickel-based composite coating includes the distribution of the hard phase and the average microhardness. Step 120: Obtain the temperature, humidity, and particle flow rate during the thermal spraying process using sensors; Step 130: Obtain the corresponding workpiece's friction load, usage frequency, and friction speed from the workpiece's usage log.

3. The method for predicting the wear resistance of nickel-based composite coatings based on machine learning according to claim 1, characterized in that, The workpiece coating adhesion analysis in S200 includes the following specific steps: Step 210: Obtain a three-dimensional image of the workpiece surface, and then obtain the distribution of pits on the workpiece surface and the corresponding roughness of the workpiece surface. At the same time, obtain the distribution of hard phase in the coating, and obtain the roughness anomaly by the standard deviation of the roughness from the safe roughness range. Step 220: Obtain the size of the pits on the workpiece surface and the size deviation of the hard phase particles to obtain the size matching anomaly. Obtain the hard phase distribution uniformity anomaly by the distribution of the hard phase in the coating. The hard phase distribution uniformity anomaly is obtained by: obtaining the hard phase distribution density at each position in the coating, obtaining the standard deviation of the hard phase distribution density at each position, and obtaining the hard phase distribution uniformity anomaly. Step 230: Obtain the distribution of pits on the workpiece surface, set the standard deviation of pit size as pit size anomaly, and at the same time obtain the spacing between pits, and obtain the standard deviation of the spacing to obtain pit distribution anomaly. Step 240: Construct a data sequence from the obtained roughness anomalies, size matching anomalies, hard phase distribution uniformity anomalies, pit size anomalies, and pit distribution anomalies in order, and set the constructed sequence as the workpiece coating connection analysis sequence.

4. The method for predicting the wear resistance of nickel-based composite coatings based on machine learning according to claim 1, characterized in that, The prediction of the wear resistance effect of the workpiece coating in S300 specifically includes the following: Step 310: Obtain the temperature, humidity, and particle flow rate during the historical thermal spraying process; simultaneously obtain the historical workpiece coating adhesion analysis sequence, the average microhardness of the nickel-based composite coating, and the wear life of the historical workpiece coating under standard friction frequency, standard friction speed, and standard friction force. Construct a multilayer perceptron model with the input of the temperature, humidity, particle flow rate, average microhardness of the nickel-based composite coating, and workpiece coating adhesion analysis sequence during the historical thermal spraying process, and the output of the wear life of the workpiece coating under standard friction frequency, standard friction speed, and standard friction force. Step 320: Import the temperature, humidity, particle flow rate, average microhardness of the nickel-based composite coating, and workpiece coating adhesion analysis sequence obtained in the real-time thermal spraying process into the constructed multilayer perceptron model, and output the predicted wear life of the workpiece coating under standard friction frequency, standard friction speed, and standard friction force.

5. The method for predicting the wear resistance of nickel-based composite coatings based on machine learning according to claim 1, characterized in that, The prediction of workpiece coating wear resistance life in S400 includes the following specific contents: Step 410: Obtain the future usage friction load, usage frequency, and friction speed of the corresponding workpiece. Divide the obtained usage friction load, usage frequency, and friction speed by the corresponding standard values ​​and then perform a weighted sum to obtain the usage anomaly prediction result. The corresponding standard values ​​are the output standard characteristics of the test equipment. Multiply the usage anomaly prediction result by the wear resistance influence coefficient to obtain the wear resistance anomaly prediction result. Step 420: Divide the wear life prediction result of the workpiece coating under standard friction frequency, standard friction speed and standard friction force by the wear anomaly prediction result.

6. The method for predicting the wear resistance of nickel-based composite coatings based on machine learning according to claim 2, characterized in that, The coating performance warning in the S500 includes the following specific contents: The predicted wear life of the workpiece coating is obtained and compared with the corresponding required wear life value. If the predicted wear life is greater than or equal to the required wear life value, it indicates that the coating performance meets the workpiece's usage requirements. If the predicted wear life is less than the required wear life value, it indicates that the coating performance does not meet the workpiece's usage requirements, and a coating warning needs to be issued to remind maintenance personnel to replace it.

7. The method for predicting the wear resistance of nickel-based composite coatings based on machine learning according to claim 2, characterized in that, The roughness of the workpiece surface is evaluated by using the average absolute value of the height of each point on the workpiece surface relative to the reference surface, and then using the ratio of the average absolute value to the safety height to obtain the roughness.

8. A machine learning-based system for predicting the wear resistance of nickel-based composite coatings, used to implement the machine learning-based method for predicting the wear resistance of nickel-based composite coatings as described in any one of claims 1-7, characterized in that, Specifically, it includes: The data acquisition module acquires information about the workpiece surface condition and the nickel-based composite coating condition, as well as the environmental conditions during the coating process and the future use of the corresponding workpiece. The coating adhesion analysis module analyzes the coating adhesion of the workpiece by considering the workpiece surface flatness and the distribution of hard phases in the nickel-based composite coating. The machine learning analysis module constructs a machine learning model to predict the wear resistance of the workpiece coating by analyzing the workpiece coating connection results, the average microhardness in the case of nickel-based composite coating, and the environmental conditions during the coating process. The life prediction module obtains the future usage of the corresponding workpiece and the prediction results of the workpiece coating wear resistance effect to predict the wear resistance life of the workpiece coating. The performance early warning module provides early warnings about coating performance based on the predicted wear life of the workpiece coating.

9. An electronic device, comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; characterized in that the processor executes the machine learning-based method for predicting the wear resistance of nickel-based composite coatings as described in any one of claims 1-7 by calling the computer program stored in the memory.

Images