A construction machinery production optimization method and system based on deep learning

Abstract

The present invention relates to the technical field of mechanical production optimization, and specifically to a construction machinery production optimization method and system based on deep learning. First, production data of an excavator crawler production process and working data of a sample crawler during operation are obtained; secondly, a production quality scoring model and a working quality scoring model are constructed to identify the production data and working data respectively, obtain a production quality score and a working quality score, and generate a first optimization parameter and a second optimization parameter; then, a crawler association analysis model is used to obtain the association between the first optimization parameter and the second optimization parameter, and generate a crawler production optimization parameter; the crawler production optimization parameter is used to optimize the crawler production data and working data to obtain a production optimization quality score and a working optimization quality score; finally, the production quality score is compared with the production optimization quality score to obtain a first optimization rate; the working quality score is compared with the working optimization quality score to obtain a second optimization rate, and the crawler production optimization rate is further generated.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical production optimization, and specifically to a construction machinery production optimization method and system based on deep learning. Background Art

[0002] With the rapid development of the construction machinery industry, excavators are important construction equipment, and the quality and performance of their key components, the crawler tracks, directly affect the operating efficiency and service life of the entire machine. However, there are many deficiencies in the existing technology for optimizing the crawler production process. Quality monitoring during the crawler production process mostly relies on manual inspection or post-inspection, which cannot achieve real-time data collection and dynamic monitoring, and can easily lead to potential defects being overlooked during the production process; at the same time, improvements to the crawler production process are usually based on empirical judgment or limited experiments, lacking systematic, data-driven analysis methods, making it difficult to accurately optimize the crawler wear characteristics under actual working conditions. In the existing technology, production quality data and work performance data are often analyzed independently, and there is a lack of effective correlation modeling methods, resulting in the inability to make targeted adjustments to the production process according to actual usage needs, thereby reducing the market adaptability and competitiveness of the product.

[0003] To this end, a construction machinery production optimization method and system based on deep learning is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a construction machinery production optimization method and system based on deep learning. First, the production data of the excavator crawler production process and the working data of the sample crawler when working are obtained; secondly, a production quality scoring model and a working quality scoring model are constructed to identify the production data and working data respectively, obtain the production quality score and the working quality score, and generate a first optimization parameter and a second optimization parameter; then, the association between the first optimization parameter and the second optimization parameter is obtained through the crawler association analysis model to generate the crawler production optimization parameter; the crawler production optimization parameter optimizes the crawler production data and the working data to obtain a production optimization quality score and a working optimization quality score; finally, the production quality score is compared with the production optimization quality score to obtain a first optimization rate; the working quality score is compared with the working optimization quality score to obtain a second optimization rate, and the crawler production optimization rate is further generated.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A deep learning-based construction machinery production optimization method comprising:

[0007] Acquire production data of excavator tracks during the production process through sensors and monitoring equipment; obtain working data by conducting experiments on sample tracks in a simulated real environment;

[0008] Constructing a production quality scoring model to identify the production data, obtain a production quality score, and generate a first optimization parameter;

[0009] Constructing a work quality scoring model to analyze the work data, obtain a work quality score, and generate a second optimization parameter;

[0010] Analyzing the first optimization parameter and the second optimization parameter using a crawler correlation analysis model, obtaining a correlation between the first optimization parameter and the second optimization parameter, and generating crawler production optimization parameters;

[0011] Optimizing crawler production data and working data using the crawler production optimization parameters to obtain a production optimization quality score and a working optimization quality score;

[0012] By comparing the production quality score with the production optimization quality score, the first optimization rate is obtained; by comparing the work quality score with the work optimization quality score, the second optimization rate is obtained;

[0013] The first optimization rate and the second optimization rate are weightedly integrated to obtain the crawler production optimization rate; the specific calculation formula of the crawler production optimization rate is:

[0014] ;

[0015] in, To optimize the production rate of crawler tracks, is the first optimization rate weight, is the first optimization rate, is the second optimization rate weight, is the second optimization rate.

[0016] Preferably, the production data includes track material characteristic data, processing parameter data and quality inspection data;

[0017] The track material characteristic data includes material parameter data, material physical properties and material coating parameters; the processing parameter data includes processing temperature, processing time, molding pressure and welding parameters; the quality inspection data includes geometric dimension parameters, surface defect parameters, hardness and stress distribution; the working data includes mechanical performance data, actual working condition data and extreme condition data; the mechanical performance data includes wear rate, stress distribution and fatigue life; the actual working condition data includes force, vibration frequency and displacement response in wetlands, gravel roads and steep slopes; the extreme condition data is the performance changes of the track under extreme environments of high temperature and high humidity.

[0018] Preferably, the production quality scoring model includes a first data preprocessing layer, a first feature extraction layer, a production quality scoring output layer and an optimization parameter acquisition layer;

[0019] The first data processing layer obtains first standard production data by normalizing the production data;

[0020] The first feature extraction layer extracts corresponding high-dimensional features by establishing fully connected layers for the track material characteristic data, processing parameter data, and quality inspection data in the first standard production data;

[0021] The production quality score output layer obtains the production quality score by fusing high-dimensional features. The specific calculation formula of the production quality score is:

[0022] ;

[0023] in, Rate production quality, Score weights for material properties, Score material properties, Processing technology scoring weight, Score the processing technology. Quality inspection score weight, Score the quality inspection. is the interaction function weight, is the interaction function;

[0024] The optimization parameter acquisition layer obtains the first optimization parameter by identifying the production quality score.

[0025] Preferably, the work quality scoring model includes a data processing and analysis layer, a second feature extraction layer, a work quality scoring layer and a second parameter optimization layer;

[0026] The data processing and analysis layer obtains standard working data by performing denoising and outlier detection on the working data;

[0027] The second feature extraction layer obtains working data features by extracting features from the standard working data; the working data features include wear rate, fatigue life, force distribution and vibration frequency;

[0028] The work quality scoring layer obtains a work quality score by performing regression analysis on work data features. The specific calculation formula for the work quality score is:

[0029] ;

[0030] in, Rate the quality of the work, is the mechanical performance scoring weight, Score the mechanical properties, is the actual working condition scoring weight, Score the actual working conditions. Score weights for extreme conditions, score for extreme conditions;

[0031] The second parameter optimization layer obtains the second optimization parameter by identifying the work quality score.

[0032] Preferably, the crawler association analysis model includes a feature association layer, an association analysis layer and a production optimization parameter generation layer;

[0033] The feature association layer obtains a feature association data set by performing feature extraction on the first optimization parameter and the second optimization parameter;

[0034] The correlation analysis layer performs correlation analysis on the feature correlation data set using the Pearson correlation coefficient to obtain the correlation between the first optimization parameter and the second optimization parameter;

[0035] The production optimization parameter generation layer obtains the crawler production optimization parameter by analyzing the relationship between the first optimization parameter and the second optimization parameter using an optimization algorithm.

[0036] Preferably, the specific calculation formula of the first optimization rate is:

[0037] ;

[0038] in, is the first optimization rate, Optimize quality scores for production, rate production quality;

[0039] The specific calculation formula of the second optimization rate is:

[0040] ;

[0041] in, is the second optimization rate, Optimize quality ratings for your work, Rate the quality of the work.

[0042] A construction machinery production optimization system based on deep learning, comprising:

[0043] The data acquisition unit uses sensors and monitoring equipment to obtain production data of excavator crawlers during the production process; and obtains working data by conducting experiments on sample crawlers in a simulated real environment;

[0044] A first optimization parameter unit is configured to construct a production quality scoring model to identify the production data, obtain a production quality score, and generate a first optimization parameter;

[0045] A second optimization parameter unit constructs work quality score data, analyzes the work data, obtains a work quality score, and generates a second optimization parameter;

[0046] an optimization parameter unit, which analyzes the first optimization parameter and the second optimization parameter using a crawler correlation analysis model, obtains a correlation between the first optimization parameter and the second optimization parameter, and generates crawler production optimization parameters; optimizes crawler production data and work data using the crawler production optimization parameters to obtain a production optimization quality score and a work optimization quality score;

[0047] The production optimization rate acquisition unit obtains a first optimization rate by comparing the production quality score with the production optimization quality score; obtains a second optimization rate by comparing the work quality score with the work optimization quality score; and obtains a crawler production optimization rate by weighted fusion of the first optimization rate and the second optimization rate. The specific calculation formula of the crawler production optimization rate is:

[0048] ;

[0049] in, To optimize the production rate of crawler tracks, is the first optimization rate weight, is the first optimization rate, is the second optimization rate weight, is the second optimization rate.

[0050] Preferably, the step of training the production quality scoring model using the production data includes:

[0051] Data preprocessing: normalize all data;

[0052] Processing to obtain a first preprocessing set;

[0053] Dataset division: The first preprocessed dataset is divided into a training set and a test set in a ratio of 7:3;

[0054] Training the model: inputting the training set into the production quality scoring model until the model training stopping condition is met, thereby obtaining an initial production quality scoring model; the model training stopping condition is when the loss of the production quality scoring model reaches convergence;

[0055] Testing model: inputting the test set into the production quality scoring model for testing, optimizing the production quality scoring model according to the test results to obtain the optimal production quality scoring model;

[0056] The production quality scoring model includes a first data preprocessing layer, a first feature extraction layer, a quality score output layer and an optimization parameter acquisition layer;

[0057] The first data processing layer obtains first standard production data by normalizing the production data;

[0058] The first feature extraction layer extracts corresponding high-dimensional features by establishing fully connected layers for the track material characteristic data, processing parameter data, and quality inspection data in the first standard production data;

[0059] The quality score output layer obtains a production quality score by fusing high-dimensional features;

[0060] The optimization parameter acquisition layer obtains the first optimization parameter by identifying the production quality score.

[0061] Preferably, the crawler association analysis model includes a feature association layer, an association analysis layer and a production optimization parameter generation layer;

[0062] The feature association layer obtains a feature association data set by performing feature extraction on the first optimization parameter and the second optimization parameter;

[0063] The correlation analysis layer performs correlation analysis on the feature correlation data set using the Pearson correlation coefficient to obtain the correlation between the first optimization parameter and the second optimization parameter;

[0064] The production optimization parameter generation layer analyzes the relationship between the first optimization parameter and the second optimization parameter by using an optimization algorithm to obtain the crawler production optimization parameters.

[0065] Preferably, the specific calculation formula of the first optimization rate is:

[0066] ;

[0067] in, is the first optimization rate, Optimize quality scores for production, rate production quality;

[0068] The specific calculation formula of the second optimization rate is:

[0069] ;

[0070] in, is the second optimization rate, Optimize quality ratings for your work, Rate the quality of the work.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] 1. This invention constructs a production quality scoring model. This model identifies key parameters in the crawler track production process: material property data, processing data, and quality inspection data. It then extracts high-dimensional features corresponding to these key parameters. By fusing these high-dimensional features, a production quality score is obtained and a first optimization parameter is generated. Combining data preprocessing with feature extraction techniques, this model effectively optimizes the production process and improves production efficiency. By analyzing crawler track production data, quality control of the crawler track manufacturing process is achieved.

[0073] 2. This invention constructs a performance quality scoring model. By using regression analysis on actual track operating conditions and wear data, such as wear rate, stress distribution, and fatigue life, a performance quality score is obtained, generating a second optimization parameter. The optimized parameters based on the performance data are used to reversely guide adjustments to the production process, ensuring that the track demonstrates durability and reliability in extreme operating environments. This data-driven optimization strategy significantly improves track quality optimization during production.

[0074] 3. This invention constructs a crawler correlation analysis model, performs multidimensional correlation analysis on a first optimization parameter derived from production quality scores and a second optimization parameter derived from work quality scores. Using a deep learning model, it integrates key features of production and work data to generate crawler production optimization parameters and optimizes parameter configuration across the production process. By calculating optimization rates and weighted fusion, the comprehensive performance of the crawler production process is accurately assessed, providing scientific and data-driven decision support. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 A schematic diagram of a process flow of a construction machinery production optimization method based on deep learning provided by the present invention;

[0076] Figure 2 A schematic diagram of the structure of a construction machinery production optimization system based on deep learning provided by the present invention;

[0077] Figure 3 A schematic diagram of a process flow for track production optimization provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0078] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0079] Example 1

[0080] The present invention proposes a construction machinery production optimization method based on deep learning, the process of the method is as follows Figure 1 Shown, including:

[0081] Acquire production data of excavator tracks during the production process through sensors and monitoring equipment; obtain working data by conducting experiments on sample tracks in a simulated real environment;

[0082] Furthermore, the production data includes track material characteristic data, processing parameter data and quality inspection data;

[0083] The track material characteristic data includes material parameter data, material physical properties and material coating parameters; the processing parameter data includes processing temperature, processing time, molding pressure and welding parameters; the quality inspection data includes geometric dimension parameters, surface defect parameters, hardness and stress distribution; the working data includes mechanical performance data, actual working condition data and extreme condition data; the mechanical performance data includes wear rate, stress distribution and fatigue life; the actual working condition data includes force, vibration frequency and displacement response in wetlands, gravel roads and steep slopes; the extreme condition data is the performance changes of the track under extreme environments of high temperature and high humidity.

[0084] In this example, by collecting production data from excavator tracks and simulating real-world operating conditions, comprehensive monitoring and evaluation of the production process and actual performance are achieved. Production processes can be optimized through in-depth analysis of material properties, processing technology, and operating data.

[0085] Constructing a production quality scoring model to identify the production data, obtain a production quality score, and generate a first optimization parameter;

[0086] Furthermore, the production quality scoring model includes a first data preprocessing layer, a first feature extraction layer, a production quality scoring output layer, and an optimization parameter acquisition layer;

[0087] The first data processing layer obtains first standard production data by normalizing the production data;

[0088] The first feature extraction layer extracts corresponding high-dimensional features by establishing fully connected layers for the track material characteristic data, processing parameter data, and quality inspection data in the first standard production data;

[0089] The production quality score output layer obtains the production quality score by fusing high-dimensional features. The specific calculation formula of the production quality score is:

[0090] ;

[0091] in, Rate production quality, Score weights for material properties, Score material properties, Processing technology scoring weight, Score the processing technology. Quality inspection score weight, Score the quality inspection. is the interaction function weight, is the interaction function;

[0092] The specific calculation formula for material property score is:

[0093] ;

[0094] in, is a logarithmic function with the natural number e as the base, is the material composition uniformity weight, For the uniformity of material composition, is the maximum value of material composition uniformity, is the material hardness weight, is the material hardness, is the maximum hardness of the material, is the material coating thickness weight, is the material coating thickness, is the maximum thickness of the material coating, is the adjustment factor of the material interaction term, It is an infinitesimal value to avoid the denominator being zero;

[0095] The specific calculation formula for the processing technology score is:

[0096] ;

[0097] in, is the actual processing temperature, Preset processing temperature for crawler tracks, is the actual processing time, Preset processing time for crawler, is the molding pressure, Preset forming pressure for the track, is the welding current intensity, Preset welding current intensity for tracks;

[0098] The specific calculation formula for quality inspection score is:

[0099] ;

[0100] in, is the geometric size deviation, is the surface defect ratio, Preset value for material hardness, is the track stress uniformity, Preset stress uniformity for the track, A factor to adjust data deviation and stress uniformity;

[0101] The specific calculation formula of the interaction function is: .

[0102] The optimization parameter acquisition layer obtains the first optimization parameter by identifying the production quality score.

[0103] In this embodiment, a production quality scoring model is constructed to analyze the material properties, processing techniques, and quality inspection indicators in track production data to achieve a comprehensive assessment of track production quality. Data preprocessing and high-dimensional feature extraction, combined with scientific scoring formulas and interactive functions, ensure accurate and reliable scoring results. The optimization parameter acquisition layer generates the first optimization parameters based on the scoring results, guiding adjustments and improvements to the production process. This significantly improves track production quality and consistency, meeting the requirements of complex operating conditions.

[0104] Constructing a work quality scoring model to analyze the work data, obtain a work quality score, and generate a second optimization parameter;

[0105] Furthermore, the work quality scoring model includes a data processing and analysis layer, a second feature extraction layer, a work quality scoring layer, and a second parameter optimization layer;

[0106] The data processing and analysis layer obtains standard working data by performing denoising and outlier detection on the working data;

[0107] The second feature extraction layer obtains working data features by extracting features from the standard working data; the working data features include wear rate, fatigue life, force distribution and vibration frequency;

[0108] The work quality scoring layer obtains a work quality score by performing regression analysis on work data features. The specific calculation formula for the work quality score is:

[0109] ;

[0110] in, Rate the quality of the work, is the mechanical performance scoring weight, Score the mechanical properties, is the actual working condition scoring weight, Score the actual working conditions. Score weights for extreme conditions, score for extreme conditions;

[0111] The specific calculation formula for mechanical performance score is:

[0112] ;

[0113] in, is an exponential function, To adjust the nonlinear factors of mechanical properties, is the track wear rate, is the maximum value of the track wear rate, is the maximum value of track stress uniformity, is the fatigue life, is the maximum fatigue life;

[0114] The specific calculation formula for the actual working condition score is:

[0115] ;

[0116] in, is the wetland anti-slip force weight, Anti-slip force in wetlands, is the vibration frequency weight of the gravel road surface, is the vibration frequency of the gravel road surface, is the steep slope stability weight, For steep slope stability;

[0117] The specific calculation formula for extreme condition score is:

[0118] ;

[0119] in, is the high temperature performance retention rate, To preset the high temperature performance retention rate, For high humidity environment anti-corrosion ability, To provide corrosion resistance in high humidity environment.

[0120] The second parameter optimization layer obtains the second optimization parameter by identifying the work quality score.

[0121] In this embodiment, a performance scoring model is constructed to comprehensively analyze and score the track's operating data under actual working conditions, effectively quantifying its mechanical properties, actual operating performance, and adaptability to extreme conditions. Data denoising and outlier detection techniques are used to ensure data accuracy, and key indicators such as wear rate, fatigue life, and force distribution are analyzed through linear regression and feature extraction. The scoring formula accurately weighs the impact of different factors to generate a secondary optimization parameter, providing a scientific basis for optimizing track performance and extending its lifespan, thereby enhancing the track's adaptability and reliability in various complex working environments.

[0122] Analyzing the first optimization parameter and the second optimization parameter by using a crawler correlation analysis model, obtaining a correlation between the first optimization parameter and the second optimization parameter, and generating crawler production optimization parameters;

[0123] Furthermore, the crawler correlation analysis model includes a feature correlation layer, a correlation analysis layer and a production optimization parameter generation layer;

[0124] The feature association layer obtains a feature association data set by performing feature extraction on the first optimization parameter and the second optimization parameter;

[0125] The correlation analysis layer performs correlation analysis on the feature correlation data set using the Pearson correlation coefficient to obtain the correlation between the first optimization parameter and the second optimization parameter;

[0126] The production optimization parameter generation layer obtains the crawler production optimization parameter by analyzing the relationship between the first optimization parameter and the second optimization parameter using an optimization algorithm.

[0127] In this embodiment, the production optimization parameter generation layer uses a multi-objective optimization algorithm to address the nonlinear relationship between production and performance parameters. Through continuous iteration, the optimal solution for simultaneously optimizing multiple objective functions is sought, thereby generating production optimization parameters. By comparing the mean square error, correlation coefficient, convergence rate, and computational complexity of each optimization algorithm, the best model is selected to obtain crawler production optimization parameters (see Table 1).

[0128] Table 1 Comparison of the effects of different optimization algorithms on obtaining crawler production optimization parameters

[0129]

[0130] The convergence speed is obtained through the number of iterations.

[0131] In this embodiment, the correlation between the first and second optimization parameters is analyzed to obtain optimized track production parameters, thereby optimizing the production parameters. For example, in terms of material optimization, by correlating the material property score with the mechanical performance score, the material composition and coating thickness of the track are optimized to improve wear resistance and fatigue resistance. In terms of process optimization, the processing temperature, molding pressure, and welding parameters are adjusted based on the processing technology score and the actual operating condition score to ensure a precise and reliable production process that meets the requirements of complex operating conditions. In terms of quality control, combined with the analysis of geometric dimensional parameters and surface defects, inspection indicators and tolerance control are optimized to reduce performance degradation caused by processing errors. In terms of environmental adaptability, based on the correlation between the extreme condition score and material properties, recommendations are made to enhance anti-corrosion treatment or improve coating performance to adapt to extreme environments such as high temperature and high humidity. In addition, by analyzing the correlation between force distribution and vibration frequency, the track structure design is optimized to provide greater stability and vibration resistance in typical operating conditions such as wetlands, gravel roads, and steep slopes.

[0132] Track production optimization parameters guide material selection, process adjustment, and product improvement, enabling tracks to perform better under a variety of complex working conditions while reducing production costs and increasing service life.

[0133] Optimizing crawler production data and working data using the crawler production optimization parameters to obtain a production optimization quality score and a working optimization quality score;

[0134] The specific calculation formula for production optimization quality score is:

[0135] ;

[0136] in, Optimize quality scores for production, Optimize scoring weights for material properties, Optimize scoring for material properties, Processing technology optimization scoring weight, Optimize the scoring for the processing technology, Quality inspection optimizes scoring weights, Optimize scoring for quality inspection, To optimize the interaction function weight, To optimize the interaction function;

[0137] The specific calculation formula for the material property optimization score is:

[0138] ;

[0139] in, is the material property optimization function, For the uniformity of material composition, is the material hardness, is the material coating thickness;

[0140] The specific calculation formula for the processing technology optimization score is:

[0141] ;

[0142] in, is the processing technology optimization function, is the actual processing temperature, is the actual processing time, is the molding pressure, is the welding current intensity;

[0143] The specific calculation formula for quality inspection optimization score is:

[0144] ;

[0145] in, Optimize the scoring function for quality detection, is the geometric size deviation, is the surface defect ratio, is the track stress uniformity.

[0146] The specific calculation formula for the work optimization quality score is:

[0147] ;

[0148] in, Optimize quality ratings for your work, Optimize scoring weights for mechanical properties, Optimize the scoring for mechanical properties, Optimize the scoring weights for actual working conditions, Optimize the score for actual working conditions, Optimize scoring weights for extreme conditions, Optimize scoring for extreme conditions;

[0149] The specific calculation formula for the mechanical performance optimization score is:

[0150] ;

[0151] in, is the mechanical performance optimization function, is the track wear rate, is fatigue life;

[0152] The specific calculation formula for the actual working condition optimization score is:

[0153] ;

[0154] in, Optimize the function for actual working conditions, Anti-slip force in wetlands, is the vibration frequency of the gravel road surface, For steep slope stability;

[0155] The specific calculation formula for extreme condition optimization score is:

[0156] ;

[0157] in, Optimize functions for extreme conditions, is the high temperature performance retention rate, Provides corrosion resistance in high humidity environments.

[0158] By comparing the production quality score with the production optimization quality score, the first optimization rate is obtained; by comparing the work quality score with the work optimization quality score, the second optimization rate is obtained;

[0159] Furthermore, the specific calculation formula of the first optimization rate is:

[0160] ;

[0161] in, is the first optimization rate, Optimize quality scores for production, rate production quality;

[0162] The specific calculation formula of the second optimization rate is:

[0163] ;

[0164] in, is the second optimization rate, Optimize quality ratings for your work, Rate the quality of the work.

[0165] In this example, by optimizing crawler production and operating data, we generate production and operating quality scores. We then calculate a first optimization rate and a second optimization rate to quantify the degree of improvement in production and operating quality. These rates are then compared with the original scores to clarify the optimization effect, providing a scientific basis for improving crawler production and performance.

[0166] The first optimization rate and the second optimization rate are weightedly integrated to obtain the crawler production optimization rate; the specific calculation formula of the crawler production optimization rate is:

[0167] ;

[0168] in, To optimize the production rate of crawler tracks, is the first optimization rate weight, is the first optimization rate, is the second optimization rate weight, is the second optimization rate.

[0169] Get 5 different production lines crawler to conduct experiments, according to Figure 1 The method flow shown is used to obtain the crawler production optimization rate, production optimization quality score, production quality score, work optimization quality score and work quality score, and further obtain the first optimization rate, the second optimization rate and the crawler production optimization rate, see Table 2 for details;

[0170] In this embodiment, the present invention realizes a comprehensive evaluation and optimization of the production quality and actual performance of the crawler by collecting the production data of the excavator crawler and simulating the working data under real working conditions. By constructing a production quality scoring model and a working quality scoring model, the production data and working data of the crawler are feature extracted and quantitatively scored respectively, and the material properties, processing technology, quality inspection indicators, mechanical properties, actual working condition performance and adaptability to extreme conditions are comprehensively evaluated. Furthermore, through the crawler correlation analysis model, the correlation between the production data and the working data is analyzed, the production material composition, processing parameters and inspection standards are optimized, and the production quality and consistency of the crawler are significantly improved. In addition, the present invention uses the production optimization parameters and working optimization parameters generated by the optimization algorithm to guide the material improvement, process optimization and environmental adaptability improvement of the crawler, respectively, so that the crawler has better performance and reliability under complex working conditions. Through the calculation and fusion of the first optimization rate and the second optimization rate, the degree of improvement in crawler production and performance is effectively quantified.

[0171] Table 2 Track sample production optimization rate and related scoring data

[0172]

[0173] Example 2

[0174] The present invention provides a construction machinery production optimization method based on deep learning. This method is applied to a construction machinery production optimization system based on deep learning. For the specific method flow chart and system structure diagram, please refer to Figure 1 and Figure 2 .

[0175] As an embodiment of the present invention, refer to Figure 1S10 in the figure is applied to a data acquisition unit of a construction machinery production optimization system based on deep learning. The data acquisition unit is used to obtain production data of excavator tracks during the production process through sensors and monitoring equipment; and obtain working data by conducting experiments on sample tracks in a simulated real environment.

[0176] Furthermore, the production data includes track material characteristic data, processing parameter data and quality inspection data;

[0177] The track material characteristic data includes material parameter data, material physical properties and material coating parameters;

[0178] The processing parameter data include processing temperature, processing time, forming pressure and welding parameters;

[0179] The quality inspection data includes geometric size parameters, surface defect parameters, hardness and stress distribution;

[0180] The working data includes mechanical performance data, actual working condition data and extreme condition data;

[0181] The mechanical performance data include wear rate, stress distribution and fatigue life;

[0182] The actual working condition data include force, vibration frequency and displacement response under wet conditions, gravel roads and steep slopes;

[0183] The extreme condition data refers to the performance changes of the track under extreme environments of high temperature and high humidity.

[0184] As an embodiment of the present invention, refer to Figure 1 S20 in the embodiment, S20 is applied to a first optimization parameter unit of a construction machinery production optimization system based on deep learning, the first optimization parameter unit being used to construct a production quality scoring model to identify the production data, obtain a production quality score, and generate a first optimization parameter;

[0185] Furthermore, the production quality scoring model includes a first data preprocessing layer, a first feature extraction layer, a production quality scoring output layer, and an optimization parameter acquisition layer;

[0186] The first data processing layer obtains first standard production data by normalizing the production data;

[0187] The first feature extraction layer extracts corresponding high-dimensional features by establishing fully connected layers for the track material characteristic data, processing parameter data, and quality inspection data in the first standard production data;

[0188] The production quality score output layer obtains the production quality score by fusing high-dimensional features. The specific calculation formula of the production quality score is:

[0189] ;

[0190] in, Rate production quality, Score weights for material properties, Score material properties, Processing technology scoring weight, Score the processing technology. Quality inspection score weight, Score the quality inspection. is the interaction function weight, is the interaction function;

[0191] The optimization parameter acquisition layer obtains the first optimization parameter by identifying the production quality score.

[0192] As an embodiment of the present invention, refer to Figure 1 S30 in which S30 is applied to a second optimization parameter unit of a construction machinery production optimization system based on deep learning, the second optimization parameter unit being used to construct work quality score data, analyze the work data, obtain a work quality score, and generate a second optimization parameter;

[0193] Furthermore, the work quality scoring model includes a data processing and analysis layer, a second feature extraction layer, a work quality scoring layer, and a second parameter optimization layer;

[0194] The data processing and analysis layer obtains standard working data by performing denoising and outlier detection on the working data;

[0195] The second feature extraction layer obtains working data features by extracting features from the standard working data; the working data features include wear rate, fatigue life, force distribution and vibration frequency;

[0196] The work quality scoring layer obtains a work quality score by performing regression analysis on work data features. The specific calculation formula for the work quality score is:

[0197] ;

[0198] in, Rate the quality of the work, is the mechanical performance scoring weight, Score the mechanical properties, is the actual working condition scoring weight, Score the actual working conditions. Score weights for extreme conditions, score for extreme conditions;

[0199] The second parameter optimization layer obtains the second optimization parameter by identifying the work quality score.

[0200] As an embodiment of the present invention, refer to Figure 1 S40 in the embodiment of the present invention is applied to an optimization parameter unit of a construction machinery production optimization system based on deep learning, the optimization parameter unit being configured to analyze the first optimization parameter and the second optimization parameter using a crawler association analysis model, obtain the association between the first optimization parameter and the second optimization parameter, and generate crawler production optimization parameters; optimize crawler production data and working data using the crawler production optimization parameters to obtain a production optimization quality score and a working optimization quality score;

[0201] Furthermore, the crawler correlation analysis model includes a feature correlation layer, a correlation analysis layer and a production optimization parameter generation layer;

[0202] The feature association layer obtains a feature association data set by performing feature extraction on the first optimization parameter and the second optimization parameter;

[0203] The correlation analysis layer performs correlation analysis on the feature correlation data set using the Pearson correlation coefficient to obtain the correlation between the first optimization parameter and the second optimization parameter;

[0204] The production optimization parameter generation layer analyzes the relationship between the first optimization parameter and the second optimization parameter by using an optimization algorithm to obtain the crawler production optimization parameter;

[0205] The crawler production data and the working data are optimized by using the crawler production optimization parameters to obtain a production optimization quality score and a working optimization quality score.

[0206] As an embodiment of the present invention, refer to Figure 1 S50 in the example is applied to a production optimization rate acquisition unit of a construction engineering machinery production optimization system based on deep learning. The production optimization rate acquisition unit is used to obtain a first optimization rate by comparing the production quality score with the production optimization quality score; to obtain a second optimization rate by comparing the work quality score with the work optimization quality score; and to obtain a crawler production optimization rate by weighted fusion of the first optimization rate and the second optimization rate. Figure 3 Schematic diagram of the process of track production optimization;

[0207] Furthermore, the specific calculation formula of the first optimization rate is:

[0208] ;

[0209] in, is the first optimization rate, Optimize quality scores for production, rate production quality;

[0210] Get 5 crawlers randomly selected from 5 different production lines for experiment. Figure 1 The method flow shown is to obtain the first optimization rate, see Table 3 for details;

[0211] Table 3 First optimization rate experimental data table

[0212]

[0213] The specific calculation formula of the second optimization rate is:

[0214] ;

[0215] in, is the second optimization rate, Optimize quality ratings for your work, Rate the quality of the work.

[0216] Get 5 different production lines crawler to conduct experiments, according to Figure 1 The method flow shown is to obtain the second optimization rate, see Table 4 for details;

[0217] Table 4 Second optimization rate experimental data table

[0218]

[0219] The first optimization rate and the second optimization rate are weightedly integrated to obtain the crawler production optimization rate; the specific calculation formula of the crawler production optimization rate is:

[0220] ;

[0221] in, To optimize the production rate of crawler tracks, is the first optimization rate weight, is the first optimization rate, is the second optimization rate weight, is the second optimization rate.

[0222] The present invention realizes the all-round evaluation and optimization of track performance and production process by comprehensively collecting and analyzing the production data and working data of the excavator track, and combining the deep learning model to build the production quality scoring and working quality scoring models. Through the in-depth analysis of material properties, processing technology, quality inspection and actual working performance data, key features are accurately extracted, and scientific scoring formulas and optimization algorithms are used to generate optimization parameters to guide the adjustment of production process. Furthermore, by constructing a track correlation analysis model, production quality is associated with working performance, and material selection, process parameters and quality inspection indicators are optimized, thereby significantly improving the durability, adaptability and reliability of the track under complex working conditions. In addition, by optimizing the quality score calculation and the quantitative evaluation of the optimization rate, the improvement of track production and working quality is effectively measured, providing a scientific basis for the improvement of track production process and performance improvement.

[0223] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A construction machinery production optimization method based on deep learning, characterized in that: include: Acquire production data of excavator tracks during the production process through sensors and monitoring equipment; Obtain working data by conducting experiments on sample crawlers in a simulated real environment; Constructing a production quality scoring model to identify production data, obtain a production quality score, and generate a first optimization parameter; Constructing a work quality scoring model to analyze work data, obtain a work quality score, and generate a second optimization parameter; Analyze the first optimization parameter and the second optimization parameter through the crawler correlation analysis model, obtain the correlation between the first optimization parameter and the second optimization parameter, and generate the crawler production optimization parameter; The crawler association analysis model includes a feature association layer, an association analysis layer and a production optimization parameter generation layer; The feature association layer obtains a feature association data set by performing feature extraction on the first optimization parameter and the second optimization parameter; The correlation analysis layer performs correlation analysis on the feature correlation data set using the Pearson correlation coefficient to obtain the correlation between the first optimization parameter and the second optimization parameter; The production optimization parameter generation layer analyzes the relationship between the first optimization parameter and the second optimization parameter by using an optimization algorithm to obtain the crawler production optimization parameter; Optimize crawler production data and work data through crawler production optimization parameters to obtain production optimization quality scores and work optimization quality scores; By comparing the production quality score with the production optimization quality score, the first optimization rate is obtained; by comparing the work quality score with the work optimization quality score, the second optimization rate is obtained; The first optimization rate and the second optimization rate are weightedly integrated to obtain the crawler production optimization rate. The specific calculation formula of the crawler production optimization rate is: in, To optimize the production rate of crawler tracks, is the first optimization rate weight, is the first optimization rate, is the second optimization rate weight, is the second optimization rate.

2. The method for optimizing construction machinery production based on deep learning according to claim 1, characterized in that: The production data includes track material characteristic data, processing parameter data and quality inspection data; The track material characteristic data includes material parameter data, material physical properties and material coating parameters; The processing parameter data include processing temperature, processing time, forming pressure and welding parameters; The quality inspection data includes geometric size parameters, surface defect parameters, hardness and stress distribution; The working data includes mechanical performance data, actual working condition data and extreme condition data; The mechanical performance data include wear rate, stress distribution and fatigue life; The actual working condition data include force, vibration frequency and displacement response under wet conditions, gravel roads and steep slopes; The extreme condition data refers to the performance changes of the track under extreme environments of high temperature and high humidity.

3. The method for optimizing construction machinery production based on deep learning according to claim 1, characterized in that: The production quality scoring model includes a first data preprocessing layer, a first feature extraction layer, a production quality scoring output layer and an optimization parameter acquisition layer; The first data processing layer obtains first standard production data by normalizing the production data; The first feature extraction layer extracts corresponding high-dimensional features by establishing fully connected layers for the track material characteristic data, processing parameter data, and quality inspection data in the first standard production data; The production quality score output layer obtains the production quality score by fusing high-dimensional features. The specific calculation formula of the production quality score is: in, Rate production quality, Score weights for material properties, Score material properties, Processing technology scoring weight, Score the processing technology. Quality inspection score weight, Score the quality inspection. is the interaction function weight, is the interaction function; The optimization parameter acquisition layer obtains the first optimization parameter by identifying the production quality score.

4. The method for optimizing construction machinery production based on deep learning according to claim 1, characterized in that: The work quality scoring model includes a data processing and analysis layer, a second feature extraction layer, a work quality scoring layer, and a second parameter optimization layer; The data processing and analysis layer obtains standard working data by performing denoising and outlier detection on the working data; The second feature extraction layer obtains working data features by extracting features from the standard working data; the working data features include wear rate, fatigue life, force distribution and vibration frequency; The work quality scoring layer obtains a work quality score by performing regression analysis on work data features. The specific calculation formula for the work quality score is: in, Rate the quality of the work, is the mechanical performance scoring weight, Score the mechanical properties, is the actual working condition scoring weight, Score the actual working conditions. Score weights for extreme conditions, score for extreme conditions; The second parameter optimization layer obtains the second optimization parameter by identifying the work quality score.

5. The method for optimizing construction machinery production based on deep learning according to claim 1, characterized in that: The specific calculation formula of the first optimization rate is: in, is the first optimization rate, Optimize quality scores for production, rate production quality; The specific calculation formula of the second optimization rate is: in, is the second optimization rate, Optimize quality ratings for your work, Rate the quality of the work.

6. A construction machinery production optimization system based on deep learning, characterized in that: Executing the deep learning-based construction machinery production optimization method according to claim 1 comprises: The data acquisition unit uses sensors and monitoring equipment to obtain production data of excavator crawlers during the production process; and obtains working data by conducting experiments on sample crawlers in a simulated real environment; A first optimization parameter unit is configured to construct a production quality scoring model to identify the production data, obtain a production quality score, and generate a first optimization parameter; A second optimization parameter unit constructs work quality score data, analyzes the work data, obtains a work quality score, and generates a second optimization parameter; an optimization parameter unit, which analyzes the first optimization parameter and the second optimization parameter using a crawler correlation analysis model, obtains a correlation between the first optimization parameter and the second optimization parameter, and generates crawler production optimization parameters; optimizes crawler production data and work data using the crawler production optimization parameters to obtain a production optimization quality score and a work optimization quality score; The production optimization rate acquisition unit obtains a first optimization rate by comparing the production quality score with the production optimization quality score; obtains a second optimization rate by comparing the work quality score with the work optimization quality score; and obtains a crawler production optimization rate by weighted fusion of the first optimization rate and the second optimization rate. The specific calculation formula of the crawler production optimization rate is: in, To optimize the production rate of crawler tracks, is the first optimization rate weight, is the first optimization rate, is the second optimization rate weight, is the second optimization rate.

Images