CN113051683A - Method, system, equipment and storage medium for predicting service life of numerical control machine tool cutter - Google Patents
Method, system, equipment and storage medium for predicting service life of numerical control machine tool cutterDownload PDFInfo
- Publication number
- CN113051683A CN113051683ACN202110326604.5ACN202110326604ACN113051683ACN 113051683 ACN113051683 ACN 113051683ACN 202110326604 ACN202110326604 ACN 202110326604ACN 113051683 ACN113051683 ACN 113051683A
- Authority
- CN
- China
- Prior art keywords
- tool
- tool life
- life
- sample data
- training sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/25—Design optimisation, verification or simulation using particle-based methods
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Geometry (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Numerical Control (AREA)
Abstract
Translated fromChinese
Description
Translated fromChinese技术领域technical field
本发明属于车间智能化加工领域，涉及一种数控机床刀具寿命预测方法、系统、设备及存储介质。The invention belongs to the field of workshop intelligent processing, and relates to a tool life prediction method, system, equipment and storage medium of a numerically controlled machine tool.
背景技术Background technique
车间智能化加工过程中，配置合理的数控刀具系统是保证工件加工精度和质量的关键。刀具寿命作为衡量刀具系统性能和评估刀具可靠性的重要指标，若在加工过程中没有对刀具寿命进行预测或者预测结果不准确，都会影响工件加工质量，严重时将导致工件报废，引起生产线停滞。因此，数控刀具寿命预测可作为切削参数配置和提前换刀的依据，准确的寿命预测可提高数控刀具的实际使用率，降低刀具使用成本。In the process of intelligent machining in the workshop, the configuration of a reasonable CNC tool system is the key to ensuring the machining accuracy and quality of the workpiece. Tool life is an important indicator to measure the performance of the tool system and evaluate the reliability of the tool. If the tool life is not predicted or the prediction result is inaccurate during the machining process, the machining quality of the workpiece will be affected. In severe cases, the workpiece will be scrapped and the production line will be stagnant. Therefore, CNC tool life prediction can be used as the basis for cutting parameter configuration and tool change in advance. Accurate life prediction can improve the actual utilization rate of CNC tools and reduce the cost of tool use.
目前，刀具寿命预测技术已经有了一些方法，一方面主要是将遗传算法GA、粒子群算法PSO、BP神经网络等智能算法引入刀具寿命预测问题中。另一方面，基于统计学习发展起来的支持向量回归机(Support Vector Regression，SVR)在小样本、非线性问题中具有优越的预测性能，也有一些方法将SVR引入刀具寿命预测问题中。然而，上述算法在解决刀具寿命预测时仍然存在一定的问题。例如：虽然神经网络具有良好的非线性逼近能力，但在刀具寿命预测过程中，反向传播算法存在收敛速度慢、易陷入局部极小点、全局搜索能力弱等缺点；虽然粒子群算法搜索速度快、效率高、易于实现的群体智能优化算法，但同时存在着容易早熟收敛、搜索精度较低等劣势；SVR中的训练参数缺乏固定的选取准则，随机选取无法保证良好的预测效果。At present, there are some methods for tool life prediction technology. On the one hand, intelligent algorithms such as genetic algorithm GA, particle swarm algorithm PSO, and BP neural network are introduced into the problem of tool life prediction. On the other hand, Support Vector Regression (SVR) developed based on statistical learning has superior prediction performance in small sample and nonlinear problems, and there are also some methods that introduce SVR into tool life prediction problems. However, the above algorithms still have certain problems in solving tool life prediction. For example, although the neural network has good nonlinear approximation ability, in the process of tool life prediction, the backpropagation algorithm has shortcomings such as slow convergence speed, easy to fall into local minimum points, and weak global search ability; although the particle swarm algorithm search speed It is a fast, efficient and easy-to-implement swarm intelligence optimization algorithm, but at the same time, it has the disadvantages of easy premature convergence and low search accuracy; the training parameters in SVR lack fixed selection criteria, and random selection cannot guarantee a good prediction effect.
发明内容SUMMARY OF THE INVENTION
本发明的目的在于克服上述现有技术的缺点，提供一种数控机床刀具寿命预测方法、系统、设备及存储介质，提高算法搜索SVR参数最优值的可能性，以提高刀具寿命预测的精度。The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art, and to provide a method, system, equipment and storage medium for predicting the tool life of a numerically controlled machine tool, so as to improve the possibility of algorithm searching for the optimal value of SVR parameters, so as to improve the accuracy of tool life prediction.
为达到上述目的，本发明采用以下技术方案予以实现：To achieve the above object, the present invention adopts the following technical solutions to realize:
一种数控机床刀具寿命预测方法，包括以下步骤；A method for predicting tool life of a CNC machine tool, comprising the following steps;
步骤一，确定n种影响刀具寿命的因素，并选择m组刀具数据组成训练样本数据集，每组刀具数据均包含对应于n个影响因素的参数值和刀具实际寿命值；Step 1: Determine n factors that affect tool life, and select m groups of tool data to form a training sample data set, each group of tool data includes parameter values and actual tool life values corresponding to the n influencing factors;
步骤二，对训练样本数据进行归一化预处理至[-1,1]区间内；Step 2: Normalize and preprocess the training sample data to be within the [-1,1] interval;
步骤三，将归一化后的训练样本数据作为AMPSO算法的输入，根据AMPSO算法的适应度函数值不断训练得到SVR模型最优的惩罚参数和核函数参数组合；
步骤四，基于所得的最佳参数和核函数参数组合，以归一化的训练样本数据集中的各参数值作为输入，刀具实际寿命值作为输出，确定预测模型的最佳核函数，采用SVR模型对样本数据进行训练，得到刀具寿命预测模型；Step 4: Based on the obtained combination of optimal parameters and kernel function parameters, each parameter value in the normalized training sample data set is used as input, and the actual tool life value is used as output to determine the optimal kernel function of the prediction model, and the SVR model is used. Train the sample data to obtain a tool life prediction model;
步骤五，将影响刀具寿命的因素输入到刀具寿命预测模型中，输出得到预测的刀具寿命。Step 5: Input the factors affecting the tool life into the tool life prediction model, and output the predicted tool life.
优选的，步骤一中，影响刀具寿命的因素为切削速度、每齿进给量、切削深度、切削宽度、刀具直径和刀具齿数。Preferably, in
优选的，步骤一中，训练样本数据集的输入数据集X＝[xi]m×n，其中，xi为各组刀具数据不同影响因素的具体参数值，输出数据集Y＝[y1，y2，y3，...，ym]T，其中y为刀具实际寿命值。Preferably, in
优选的，步骤二中，归一化预处理函数为
优选的，步骤三中，将刀具实际寿命与预测寿命之间的偏差转换为AMPSO算法的适应度函数值，适应度函数为误差平方和的倒数，即
优选的，步骤四中，选择误差最小的径向基核函数作为最佳核函数，径向基函数选为高斯核函数
优选的，得到刀具寿命预测模型后，将待测刀具的数据输入所建立的刀具寿命预测模型中，反归一化输出最终预测结果，通过计算实际寿命与预测寿命之间的偏差，来检验所建立的铣刀寿命预测模型的预测准确率，如果准确率小于要求值，则返回步骤一，如果准确率大于等于要求值，则进行步骤五。Preferably, after the tool life prediction model is obtained, the data of the tool to be tested is input into the established tool life prediction model, the final prediction result is output by inverse normalization, and the deviation between the actual life and the predicted life is calculated to verify the The prediction accuracy rate of the established milling cutter life prediction model, if the accuracy rate is less than the required value, go back to step one, if the accuracy rate is greater than or equal to the required value, go to step five.
一种数控机床刀具寿命预测系统，包括：A tool life prediction system for CNC machine tools, comprising:
训练样本数据集构建模块，用于分析确定n种影响刀具寿命的因素，并选择m组刀具数据组成训练样本数据集，每组刀具数据均包含对应于n个影响因素的参数值和刀具实际寿命值；The training sample data set building module is used to analyze and determine n factors that affect tool life, and select m sets of tool data to form a training sample data set. Each set of tool data contains parameter values corresponding to the n influencing factors and the actual tool life. value;
归一化模块，用于对训练样本数据进行归一化预处理至[-1,1]区间内；The normalization module is used to normalize and preprocess the training sample data to the [-1,1] interval;
SVR模型参数优化模块，用于将归一化后的训练样本数据作为AMPSO算法的输入，根据AMPSO算法的适应度函数值不断训练得到SVR模型最优的惩罚参数和核函数参数组合；The SVR model parameter optimization module is used to take the normalized training sample data as the input of the AMPSO algorithm, and continuously train to obtain the optimal combination of penalty parameters and kernel function parameters of the SVR model according to the fitness function value of the AMPSO algorithm;
刀具寿命预测模型构建模块，用于基于所得的最佳参数和核函数参数组合，以归一化的训练样本数据集中的各参数值作为输入，刀具实际寿命值作为输出，确定预测模型的最佳核函数，采用SVR模型对样本数据进行训练，得到刀具寿命预测模型；The tool life prediction model building module is used to determine the best prediction model based on the obtained combination of the best parameters and kernel function parameters, using the parameter values in the normalized training sample data set as input and the actual tool life value as output. Kernel function, the SVR model is used to train the sample data, and the tool life prediction model is obtained;
应用模块，用于将影响刀具寿命的因素输入到刀具寿命预测模型中，输出得到预测的刀具寿命。The application module is used to input the factors affecting the tool life into the tool life prediction model, and output the predicted tool life.
一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如上述任意一项所述数控机床刀具寿命预测方法的步骤。A computer device, comprising a memory, a processor and a computer program stored in the memory and running on the processor, when the processor executes the computer program, the numerical control machine tool described in any one of the above is implemented Steps of a tool life prediction method.
一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现如上述任意一项数控机床刀具寿命预测方法的步骤。A computer-readable storage medium storing a computer program, when the computer program is executed by a processor, implements the steps of any one of the above-mentioned methods for predicting tool life of a CNC machine tool.
与现有技术相比，本发明具有以下有益效果：Compared with the prior art, the present invention has the following beneficial effects:
本发明针对SVR中的训练参数缺乏固定的选取准则，在普通粒子群算法的基础上引入简单变异算子，以一定概率重新初始化粒子，算法中每个粒子都代表问题的一个潜在解，每个粒子对应一个由适应度函数决定的适应度值，更新粒子从而实现个体在可解空间的寻优。与现有方法相比，本方法的预测误差小、精度高、泛化能力强，可准确预测数控铣刀寿命，同时训练时间短、效率高，可以为车间刀具准备和数控加工过程中的换刀决策提供依据。Aiming at the lack of fixed selection criteria for the training parameters in the SVR, the invention introduces a simple mutation operator on the basis of the ordinary particle swarm algorithm, and re-initializes the particles with a certain probability. Each particle in the algorithm represents a potential solution of the problem. The particle corresponds to a fitness value determined by the fitness function, and the particle is updated to realize the optimization of the individual in the solvable space. Compared with the existing method, the method has small prediction error, high precision and strong generalization ability, and can accurately predict the life of CNC milling cutters. At the same time, the training time is short and the efficiency is high. The knife decision provides the basis.
附图说明Description of drawings
图1为本发明的方法流程图；Fig. 1 is the method flow chart of the present invention;
图2为本发明的案例刀具在训练样本集的实际寿命和预测寿命图。FIG. 2 is a graph of the actual life and predicted life of the case tool of the present invention in the training sample set.
具体实施方式Detailed ways
下面结合附图对本发明做进一步详细描述：Below in conjunction with accompanying drawing, the present invention is described in further detail:
本发明所述的数控机床刀具寿命预测方法，即基于AMPSO-SVR的数控机床刀具寿命预测方法的流程图如图1所示，包括以下步骤：The method for predicting tool life of CNC machine tools according to the present invention, that is, the flow chart of the method for predicting tool life of CNC machine tools based on AMPSO-SVR, is shown in Figure 1, and includes the following steps:
(1)分析确定n种影响刀具寿命的因素，并选择m组刀具数据组成训练样本数据集，每组刀具数据均包含对应于n个影响因素的参数值和刀具实际寿命值，则训练数据集中样本的个数为m，训练样本数据集的输入数据集X＝[xi]m×n，其中，xi为各组刀具数据不同影响因素的具体参数值，输出数据集Y＝[y1，y2，y3，...，ym]T，其中y为刀具实际寿命值。(1) Analyze and determine n factors that affect tool life, and select m groups of tool data to form a training sample data set. Each group of tool data contains parameter values and actual tool life values corresponding to the n influencing factors. The number of samples is m, the input data set X=[xi ]m×n of the training sample data set, wherein,xi is the specific parameter value of the different influencing factors of each group of tool data, and the output data set Y=[y1 , y2 , y3 , ..., ym ]T , where y is the actual tool life value.
刀具寿命是指一把新刀具从投入使用起，直到刀具报废为止的切削时间总和。刀具从开始使用至达到磨钝标准时应保证的切削时间为刀具耐用度。在数值上，刀具寿命等于刀具耐用度乘以刃磨次数。在实际加工中，刀具寿命通常指刀具耐用度。Tool life refers to the total cutting time of a new tool from the time it is put into use until the tool is scrapped. The cutting time that should be guaranteed from the beginning of use to the time when the tool reaches the blunt standard is the tool durability. Numerically, tool life is equal to tool durability multiplied by the number of sharpenings. In actual machining, tool life usually refers to tool durability.
由实践和经验总结出刀具寿命计算公式如下：Summarized from practice and experience, the tool life calculation formula is as follows:
式中，T为刀具使用寿命，Cr为刀具寿命系数，D0为刀具直径，vc为切削速度，ap为切削深度，f为每齿进给量，aw为切削宽度，Z为刀具齿数。x、y、u、p、q为各个参数对刀具寿命的影响指数。In the formula, T is the tool life, Cr is the tool life coefficient, D0 is the tool diameter, vc is the cutting speed, ap is the cutting depth, f is the feed per tooth, aw is the cutting width, and Z is Number of tool teeth. x, y, u, p, q are the influence indexes of each parameter on tool life.
由公式可知，影响刀具寿命的主要因素有切削速度、每齿进给量、切削深度、切削宽度、刀具直径和刀具齿数。通过以上分析，设定影响刀具寿命的因素为6种，即n＝6，分别为切削速度、每齿进给量、切削深度、切削宽度、刀具直径和刀具齿数，本发明将这6个影响因素作为刀具寿命预测模型的输入。It can be seen from the formula that the main factors affecting tool life are cutting speed, feed per tooth, depth of cut, width of cut, tool diameter and the number of tool teeth. Through the above analysis, six factors are set to affect the tool life, namely n=6, which are cutting speed, feed per tooth, cutting depth, cutting width, tool diameter and number of teeth. factors as input to the tool life prediction model.
(2)采用归一化预处理函数
(3)将归一化后的训练样本数据作为AMPSO算法的输入，将铣刀实际寿命与预测寿命之间的偏差转换为AMPSO算法的适应度函数值，适应度函数为误差平方和的倒数，即
为了构造具有较强预测性能的SVR模型来预测刀具寿命，需要设置较优的SVR模型的惩罚参数c和核函数参数σ组合。本文在PSO算法的基础上引入自适应变异算子，构建AMPSO算法来优化SVR模型的参数组合c和σ。所述的AMPSO算法，基本思想是：粒子每次更新之后，以一定的概率重新初始化粒子。这一操作扩展了在迭代过程中不断缩小的搜索空间，使粒子能够跳出当前的最优解，提高模型寻找到更优解的可能性。AMPSO算法中的粒子更新具体实施步骤如下：In order to construct an SVR model with strong prediction performance to predict tool life, it is necessary to set a combination of the penalty parameter c and the kernel function parameter σ of the optimal SVR model. In this paper, an adaptive mutation operator is introduced on the basis of the PSO algorithm, and the AMPSO algorithm is constructed to optimize the parameter combination c and σ of the SVR model. The basic idea of the AMPSO algorithm is: after each update of the particle, the particle is re-initialized with a certain probability. This operation expands the search space that is continuously reduced in the iterative process, so that the particles can jump out of the current optimal solution and improve the possibility of the model finding a better solution. The specific implementation steps of particle update in AMPSO algorithm are as follows:
经过k次迭代，第i个粒子在N维搜索空间的位置为pop(j，k)，p为变异阈值，当粒子满足大于阈值时，将跳出当前位置，出现新的位置，否则保持不变。可表示为After k iterations, the position of the i-th particle in the N-dimensional search space is pop(j, k), and p is the mutation threshold. When the particle meets the threshold, it will jump out of the current position and a new position will appear, otherwise it will remain unchanged. . can be expressed as
其中，p为0到1之间的常数，r1和r2为0到1之间的随机数，ceil(x)为将x四舍五入为大于或等于最接近x的整数，sizepop为种群最大数量，popgmax、popgmin分别为SVR核函数参数变化的最大值和最小值。where p is a constant between 0 and 1, r1 and r2 are random numbers between 0 and 1, ceil(x) is the rounding of x to the nearest integer greater than or equal to x, and sizepop is the maximum population size , popgmax and popgmin are the maximum and minimum values of the parameter changes of the SVR kernel function, respectively.
(4)基于所得的最佳参数c和σ组合，以归一化的训练样本数据集中的各参数值作为输入，刀具实际寿命值作为输出，采用SVR对样本数据进行训练，得到基于AMPSO-SVR的铣刀寿命预测模型。其中：在多次试验的基础上选择误差最小的径向基函数作为SVR模型的核函数，径向基函数选为高斯核函数
(5)将待测刀具的数据输入所建立的刀具寿命预测模型中，反归一化输出最终预测结果，通过计算实际寿命与预测寿命之间的偏差，来检验所建立的铣刀寿命预测模型的预测准确率。(5) Input the data of the tool to be tested into the established tool life prediction model, denormalize and output the final prediction result, and check the established milling cutter life prediction model by calculating the deviation between the actual life and the predicted life. prediction accuracy.
以下为本发明的实施例，本实施例采用的刀具材料为硬质合金钢，采用立铣加工方式，在加工要求为粗铣的条件下，加工材料为45钢的工件，实验数据如表1所示。其包括以下步骤：The following are embodiments of the present invention, the tool material used in this embodiment is cemented carbide steel, and vertical milling is adopted. Under the condition that the processing requirement is rough milling, the processing material is a workpiece of 45 steel, and the experimental data are shown in Table 1. shown. It includes the following steps:
表1实验样本数据(训练+测试)Table 1 Experimental sample data (training + testing)
首先，将表1中的第2-7列作为模型的输入，最后一列作为模型的输出，训练并测试基于AMPSO-SVR的铣刀寿命预测模型。将训练样本集(编号1-7)的输入数据归一化后，采用AMPSO算法优化SVR模型参数，得到优化后的参数为：c＝12.94，σ＝0.01。First, take columns 2-7 in Table 1 as the input of the model, and the last column as the output of the model, to train and test the AMPSO-SVR-based milling cutter life prediction model. After normalizing the input data of the training sample set (numbered 1-7), the AMPSO algorithm is used to optimize the parameters of the SVR model, and the optimized parameters are obtained as: c=12.94, σ=0.01.
其次，采用SVR训练得出铣刀寿命预测模型，计算该模型在训练集上的预测结果如图2所示，图中标出了针对每个训练样本的铣刀寿命预测相对误差，最小为0.35046％，最大为1.3496％。因此，所训练的基于AMPSO-SVR的铣刀寿命预测模型在训练样本集上可较准确地预测铣刀寿命。Secondly, the milling cutter life prediction model is obtained by SVR training, and the prediction result of the model on the training set is shown in Figure 2. The figure shows the relative error of milling cutter life prediction for each training sample, the minimum is 0.35046% , with a maximum of 1.3496%. Therefore, the trained AMPSO-SVR-based milling cutter life prediction model can more accurately predict the milling cutter life on the training sample set.
最后，在测试样本集(编号8-10)上对所训练的AMPSO-SVR模型进行验证，得到的铣刀寿命预测结果如表2所示。由表可知：预测寿命与实际寿命之间的相对误差最小为0.5872％，最大为0.9088％。因此，所训练的基于AMPSO-SVR的铣刀寿命预测模型在测试样本集上也可较准确地预测铣刀寿命。Finally, the trained AMPSO-SVR model is verified on the test sample set (numbered 8-10), and the obtained milling cutter life prediction results are shown in Table 2. It can be seen from the table that the relative error between the predicted life and the actual life is at least 0.5872% and at most 0.9088%. Therefore, the trained milling cutter life prediction model based on AMPSO-SVR can also more accurately predict milling cutter life on the test sample set.
表2测试样本集上的铣刀寿命预测结果Table 2 Prediction results of milling cutter life on the test sample set
SVR中的惩罚参数c和核函数参数σ对寿命预测结果影响很大，常用网格搜索(GridSearch，GS)等方法进行参数的选择。针对测试数据集，将GS-SVR、BP神经网络和本发明提出的AMPSO-SVR算法的预测结果进行比较，如表3所示。The penalty parameter c and the kernel function parameter σ in SVR have a great influence on the life prediction results, and methods such as grid search (GS) are often used to select parameters. For the test data set, the prediction results of GS-SVR, BP neural network and the AMPSO-SVR algorithm proposed by the present invention are compared, as shown in Table 3.
表3算法预测结果对比Table 3 Comparison of algorithm prediction results
通过对比可知，采用AMPSO-SVR算法能够得到更小的误差和更高的预测精度。此外，在模型训练时间上，采用BP神经网络的训练时间为2.095025秒，而采用AMPSO-SVR的训练时间为0.877488秒，可见本发明提出的AMPSO-SVR模型的训练效率较高。By comparison, it can be seen that the AMPSO-SVR algorithm can obtain smaller errors and higher prediction accuracy. In addition, in terms of model training time, the training time using the BP neural network is 2.095025 seconds, while the training time using AMPSO-SVR is 0.877488 seconds. It can be seen that the training efficiency of the AMPSO-SVR model proposed by the present invention is high.
本发明所述的数控机床刀具寿命预测系统，包括：The CNC machine tool tool life prediction system according to the present invention includes:
训练样本数据集构建模块，用于分析确定n种影响刀具寿命的因素，并选择m组刀具数据组成训练样本数据集，每组刀具数据均包含对应于n个影响因素的参数值和刀具实际寿命值。The training sample data set building module is used to analyze and determine n factors that affect tool life, and select m sets of tool data to form a training sample data set. Each set of tool data contains parameter values corresponding to the n influencing factors and the actual tool life. value.
归一化模块，用于对训练样本数据进行归一化预处理至[-1,1]区间内。The normalization module is used to normalize and preprocess the training sample data into the [-1,1] interval.
SVR模型参数优化模块，用于将归一化后的训练样本数据作为AMPSO算法的输入，根据AMPSO算法的适应度函数值不断训练得到SVR模型最优的惩罚参数和核函数参数组合。The SVR model parameter optimization module is used to take the normalized training sample data as the input of the AMPSO algorithm, and continuously train according to the fitness function value of the AMPSO algorithm to obtain the optimal combination of penalty parameters and kernel function parameters of the SVR model.
刀具寿命预测模型构建模块，用于基于所得的最佳参数和核函数参数组合，以归一化的训练样本数据集中的各参数值作为输入，刀具实际寿命值作为输出，确定预测模型的最佳核函数，采用SVR模型对样本数据进行训练，得到刀具寿命预测模型。The tool life prediction model building module is used to determine the best prediction model based on the obtained combination of the best parameters and kernel function parameters, using the parameter values in the normalized training sample data set as input and the actual tool life value as output. The kernel function is used to train the sample data with the SVR model, and the tool life prediction model is obtained.
应用模块，用于将影响刀具寿命的因素输入到刀具寿命预测模型中，输出得到预测的刀具寿命。The application module is used to input the factors affecting the tool life into the tool life prediction model, and output the predicted tool life.
本发明所述的计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如上述任意一项所述数控机床刀具寿命预测方法的步骤。The computer device of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program, the processor implements any of the above Describe the steps of the tool life prediction method for CNC machine tools.
本发明所述的计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现如上述任意一项数控机床刀具寿命预测方法的步骤。In the computer-readable storage medium of the present invention, the computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of any one of the above-mentioned methods for predicting the tool life of a CNC machine tool are implemented.
以上内容仅为说明本发明的技术思想，不能以此限定本发明的保护范围，凡是按照本发明提出的技术思想，在技术方案基础上所做的任何改动，均落入本发明权利要求书的保护范围之内。The above content is only to illustrate the technical idea of the present invention, and cannot limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention all fall within the scope of the claims of the present invention. within the scope of protection.
Claims (10)
Translated fromChinese
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110326604.5ACN113051683B (en) | 2021-03-26 | 2021-03-26 | Numerical control machine tool life prediction method, system, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110326604.5ACN113051683B (en) | 2021-03-26 | 2021-03-26 | Numerical control machine tool life prediction method, system, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113051683Atrue CN113051683A (en) | 2021-06-29 |
| CN113051683B CN113051683B (en) | 2023-05-26 |
Family
ID=76515661
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110326604.5AActiveCN113051683B (en) | 2021-03-26 | 2021-03-26 | Numerical control machine tool life prediction method, system, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113051683B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113743011A (en)* | 2021-08-31 | 2021-12-03 | 华中科技大学 | Device life prediction method and system based on PSO-BP neural network |
| CN113821419A (en)* | 2021-08-20 | 2021-12-21 | 西安理工大学 | A cloud server aging prediction method based on SVR and Gaussian function |
| CN114417734A (en)* | 2022-03-09 | 2022-04-29 | 深圳市信润富联数字科技有限公司 | Method and device for predicting service life of tool |
| CN115563872A (en)* | 2022-10-10 | 2023-01-03 | 中铁十八局集团有限公司 | A Prediction Method of TBM Hob Wear Amount Based on DE-SVR Algorithm |
| CN115903662A (en)* | 2022-12-01 | 2023-04-04 | 扬州环锐科技有限公司 | Method and system for managing research and development data of numerical control machine tool |
| CN116484752A (en)* | 2023-06-21 | 2023-07-25 | 深圳精匠云创科技有限公司 | Design method, system, device and storage medium of cutting edge shape |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778010A (en)* | 2016-12-29 | 2017-05-31 | 中铁十八局集团隧道工程有限公司 | TBM cutter life Forecasting Methodologies based on data-driven support vector regression |
| WO2019001220A1 (en)* | 2017-06-26 | 2019-01-03 | 江南大学 | Decision tree based turning process energy consumption prediction system and method of numerically controlled lathe |
| WO2021004154A1 (en)* | 2019-07-05 | 2021-01-14 | 重庆邮电大学 | Method for predicting remaining life of numerical control machine tool |
| US20210049515A1 (en)* | 2019-08-16 | 2021-02-18 | China Institute Of Water Resources And Hydropower Research | Prediction method and system of high slope deformation |
- 2021
- 2021-03-26CNCN202110326604.5Apatent/CN113051683B/enactiveActive
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778010A (en)* | 2016-12-29 | 2017-05-31 | 中铁十八局集团隧道工程有限公司 | TBM cutter life Forecasting Methodologies based on data-driven support vector regression |
| WO2019001220A1 (en)* | 2017-06-26 | 2019-01-03 | 江南大学 | Decision tree based turning process energy consumption prediction system and method of numerically controlled lathe |
| WO2021004154A1 (en)* | 2019-07-05 | 2021-01-14 | 重庆邮电大学 | Method for predicting remaining life of numerical control machine tool |
| US20210049515A1 (en)* | 2019-08-16 | 2021-02-18 | China Institute Of Water Resources And Hydropower Research | Prediction method and system of high slope deformation |
Non-Patent Citations (3)
| Title |
|---|
| A. F. TAVAEVA: "Investigation of cutting speed influence on optimality of the tool path route for CNC laser cutting machines", 《2017 INTERNATIONAL CONFERENCE ON INDUSTRIAL ENGINEERING, APPLICATIONS AND MANUFACTURING (ICIEAM)》* |
| 周光辉; 王杰; 韩占磊; 江平宇: "基于RFID的车间刀具自动识别技术与系统实现", 《四川兵工学报》* |
| 夏颖怡: "基于GA-BP神经网络的刀具寿命预测研究", 《精密制造与自动化》* |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113821419A (en)* | 2021-08-20 | 2021-12-21 | 西安理工大学 | A cloud server aging prediction method based on SVR and Gaussian function |
| CN113821419B (en)* | 2021-08-20 | 2025-01-07 | 西安理工大学 | A cloud server aging prediction method based on SVR and Gaussian function |
| CN113743011A (en)* | 2021-08-31 | 2021-12-03 | 华中科技大学 | Device life prediction method and system based on PSO-BP neural network |
| CN114417734A (en)* | 2022-03-09 | 2022-04-29 | 深圳市信润富联数字科技有限公司 | Method and device for predicting service life of tool |
| CN114417734B (en)* | 2022-03-09 | 2022-07-12 | 深圳市信润富联数字科技有限公司 | Method and device for predicting service life of tool |
| CN115563872A (en)* | 2022-10-10 | 2023-01-03 | 中铁十八局集团有限公司 | A Prediction Method of TBM Hob Wear Amount Based on DE-SVR Algorithm |
| CN115563872B (en)* | 2022-10-10 | 2024-04-12 | 中铁十八局集团有限公司 | TBM hob abrasion loss prediction method based on DE-SVR algorithm |
| CN115903662A (en)* | 2022-12-01 | 2023-04-04 | 扬州环锐科技有限公司 | Method and system for managing research and development data of numerical control machine tool |
| CN115903662B (en)* | 2022-12-01 | 2024-01-12 | 扬州环锐科技有限公司 | Numerical control machine tool research and development data management method and system |
| CN116484752A (en)* | 2023-06-21 | 2023-07-25 | 深圳精匠云创科技有限公司 | Design method, system, device and storage medium of cutting edge shape |
| CN116484752B (en)* | 2023-06-21 | 2024-04-16 | 深圳精匠云创科技有限公司 | Cutter blade design method, system, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113051683B (en) | 2023-05-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113051683A (en) | Method, system, equipment and storage medium for predicting service life of numerical control machine tool cutter | |
| CN111444247A (en) | KPI (Key performance indicator) -based root cause positioning method and device and storage medium | |
| CN107220734A (en) | CNC Lathe Turning process Energy Consumption Prediction System based on decision tree | |
| CN113664612A (en) | Numerical control machine tool milling cutter abrasion real-time monitoring method based on deep convolutional neural network | |
| CN110782658B (en) | Traffic prediction method based on LightGBM algorithm | |
| CN111300146A (en) | Numerical control machine tool cutter abrasion loss online prediction method based on spindle current and vibration signal | |
| CN114297802A (en) | Multi-objective optimization method for laser cutting technological parameters of thin plate | |
| CN114918581B (en) | Welding parameter processing method and device, storage medium and processor | |
| CN114492198B (en) | A cutting force prediction method based on improved PSO algorithm assisted SVM algorithm | |
| CN111797469A (en) | Aeroengine casing technological parameter optimization method based on machining cutter back-off deformation constraint | |
| CN107037306A (en) | Transformer fault dynamic early-warning method based on HMM | |
| CN114678080A (en) | Prediction model and construction method of phosphorus content at the end point of converter, and phosphorus content prediction method | |
| CN107358294A (en) | A kind of water demand prediction method based on Elman neutral nets | |
| CN117520825B (en) | Industrial master machining workpiece quality prediction method based on multi-scale feature fusion | |
| CN112420132A (en) | A kind of gasoline catalytic cracking process product quality optimization control method | |
| US20220277330A1 (en) | Method and system for product processing price prediction based on multiple regression model | |
| CN118466397B (en) | Monitoring system of numerical control machine tool | |
| CN111177217A (en) | Data preprocessing method, device, computer equipment and storage medium | |
| CN107947156B (en) | Power grid fault critical clearing time discrimination method based on improved Softmax regression | |
| CN118839960A (en) | Mine mining area water inrush early warning method based on similarity analysis | |
| CN116910579A (en) | Variable working condition machining chatter monitoring method based on transfer learning | |
| CN102621953A (en) | Automatic online quality monitoring and prediction model updating method for rubber hardness | |
| CN114926075A (en) | Mechanical part production scheduling method based on man-hour prediction | |
| CN113743453A (en) | Population quantity prediction method based on random forest | |
| CN117149551B (en) | Test method of vehicle-mounted wireless communication chip |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |