CN114491839A - Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool - Google Patents

Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool Download PDF

Info

Publication number
CN114491839A
CN114491839A CN202210027023.6A CN202210027023A CN114491839A CN 114491839 A CN114491839 A CN 114491839A CN 202210027023 A CN202210027023 A CN 202210027023A CN 114491839 A CN114491839 A CN 114491839A
Authority
CN
China
Prior art keywords
data
neural network
prediction model
network prediction
numerical control
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.)
Pending
Application number
CN202210027023.6A
Other languages
Chinese (zh)
Inventor
张炜
蔡杰
李晓俊
邓劲莲
赵伟胜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Institute of Mechanical and Electrical Engineering Co Ltd
Original Assignee
Zhejiang Institute of Mechanical and Electrical Engineering Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang Institute of Mechanical and Electrical Engineering Co Ltd filed Critical Zhejiang Institute of Mechanical and Electrical Engineering Co Ltd
Priority to CN202210027023.6A priority Critical patent/CN114491839A/en
Publication of CN114491839A publication Critical patent/CN114491839A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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; CALCULATING OR 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; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Computer Hardware Design (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Physiology (AREA)
  • Genetics & Genomics (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Electrical Discharge Machining, Electrochemical Machining, And Combined Machining (AREA)

Abstract

The invention relates to the technical field of artificial intelligence, provides a method and a system for intelligently optimizing processing parameters of a numerical control electric spark machine tool, the method optimizes the structural parameters of a BP neural network prediction model P through a genetic algorithm, then adopts the standard technological processing parameter data of a numerical control electric spark machine tool to carry out training and performance testing, so that the numerical control electric spark machine carries out self-adaptive adjustment and correction on the processing parameters in the processing process, and also provides a system for implementing the method, the pulse interval, the pulse width and the peak current in the electric spark machining electrode can be continuously adjusted in the machining process, the numerical control electric spark machine tool can be adaptively adjusted according to the machining parameters of the next working condition, the change of the machining parameters caused by electrode abrasion, the distance between the machined surfaces and the like in the machining process is effectively avoided, and the machining precision and the adaptability of the whole numerical control electric spark machine tool are improved.

Description

Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a system for intelligently optimizing machining parameters of a numerical control electric spark machine tool.
Background
A numerical control Electric Discharge Machine (EDM) is a special machine tool which processes conductive materials by using the principle of Electric spark processing, and is also called an electroerosion machine tool; the tool is mainly used for processing various high-hardness materials, such as hard alloy, quenched steel and the like, dies and parts with complex shapes, and tools for cutting, slotting and removing broken parts in workpiece holes, such as drill bits, screw taps and the like; with the development of digital control technology, the electric spark machine tool is numerically controlled and is controlled by a microcomputer, so that the machine tool has more complete functions and greatly improved automation degree, the automatic positioning of an electrode and a workpiece, the automatic conversion of processing conditions, the automatic exchange of the electrode, the automatic feeding of a workbench, the multidirectional servo control of a translation head and the like are realized, the processing speed, the processing precision and the processing stability are obviously improved by technologies such as a low-loss power supply, a micro-finishing power supply, an adaptive control technology, a perfect clamp system and the like, and the application range of the electric spark machine tool is developed towards small size, precision and special directions.
CN111014851B discloses a method for processing multiple compound angle rubber openings based on EDM, which is characterized by comprising: s1, designing the shape of the electrode head according to the shape of the rubber opening of the workpiece, wherein the tail of the electrode tip is provided with a sub-center frame which exceeds the highest surface of the workpiece and faces to the Z direction and is used for processing and positioning the spindle of the EDM machine tool; s2, placing the workpiece on a machine tool workbench according to the 2D drawing direction of the workpiece, and righting the workpiece in the X and Y directions; s3, mounting the electrode at the original point A of the machine tool spindle in an inverted manner and locking; s4, confirming that the central line of a first glue opening of the machined workpiece forms 45 degrees with the X axis and 40 degrees with the Z axis, and controlling the main shaft and the electrode to move to an initial position B; s5, controlling the electrode to start from an initial position B, simultaneously feeding along three directions of XYZ, performing electric discharge machining along a direction forming an included angle of 40 degrees with the Z axis to a final position C, and finishing electric discharge machining, namely, the initial position of the electrode is placed at a position which is concentric with the center line of the glue opening and completely avoids the workpiece, wherein X is 22.071, Y is 6.091, and Z is 37.000, controlling the three directions of the main shaft XYZ to simultaneously feed, finally reaching a position required by the glue opening, X is 20.386, Y is 7.776, and Z is 35.000, and after the workpiece is subjected to electric discharge machining, the electrode safely exits the workpiece according to a feeding route; s6, controlling the electrode to return to an initial position B along the original path of the feeding direction and to displace to a main shaft original point A; s7, confirming that the center line of the second glue opening of the machined workpiece forms 27 degrees with the X axis, and repeating the step S3-the step S6 to perform the electric discharge machining of the next glue opening until all the glue openings are machined, wherein the workpiece is always unchanged in position.
At present, the traditional electric spark machining efficiency is often lower than that of other mechanical machining methods, the machining precision is obviously limited, particularly for the minimum corner radius, the radius obtained by electric spark machining is slightly larger than a machining discharge gap, and is usually between 0.02 mm and 0.03 mm; if the electrode is worn or machined with a translation head, the corner radius will have a large deviation and it is not possible to achieve a true full right angle; the electro-discharge machined surface often has an altered layer or even micro-cracks, the machined surface often has many pulse discharge pits, and the surface gloss is often poor.
Disclosure of Invention
In the processing practice of the numerical control electric spark machine tool, when an electrode is worn or a translation head is used for processing in the traditional numerical control electric spark processing process, the radius of a corner has larger deviation and can not reach a real full right angle; the problem of low precision is caused by that the machine tool is difficult to adaptively adjust and correct the machining parameters of the next working condition in the numerical control electric spark machining process, namely the machining parameters in the process, such as pulse interval, pulse width, peak current and the like, cannot be predicted; the accumulated superposition of deviation is caused by the abrasion of the numerical control spark electrode and the change of the distance between the processing surfaces, and further a series of problems of low processing precision are caused.
In view of the above, the present invention is directed to a method for intelligently optimizing processing parameters of a numerical control electric spark machine tool, the method comprising,
s1, selecting M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of a neural network, and using the rest data in the M groups of data as performance test sample data, wherein M, N are positive integers, and M is larger than N;
step S2, establishing an initial BP neural network prediction model P, optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, wherein the structural parameters at least comprise hidden layer node number, threshold value and connection weight, and constructing the optimized BP neural network prediction model P1；
Step S3, after the N groups of data in the step S1 are normalized, training a BP neural network prediction model P1Waiting for BP neural network prediction model P1After the training error precision is less than or equal to r and the residual data in the M groups of data are input for normalization processing, a BP neural network prediction model P is subjected to1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1；
Step S4, selecting a numerical control electric spark machine tool M1Group data at least including pulse interval, pulse width and peak current, normalizing, and inputting into BP neural network prediction model P1And obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool.
Preferably, in step S2, the genetic algorithm step includes,
step S21, dividing chromosome structuralization into control genes and parameter genes, optimizing the number of hidden layer nodes of the BP neural network prediction model P by using the control genes, and optimizing the connection weight and the threshold of the BP neural network prediction model P by using the parameter genes;
step S22, a fitness function is set,
f＝a·frmse+b·fcom，a＞0；b＜1
Figure BDA0003464437180000031
Figure BDA0003464437180000032
where n is the number of training data samples, frmseThe root mean square error of the training data, i.e. the square root of the ratio of the square of the deviation of the predicted value from the true value to the number n of samples of training data, yiIn the form of an actual value of the value,
Figure BDA0003464437180000033
is a predicted value; a. b is weight, a, b belongs to (0, 1), Q (0) is original population individual quantity, and Q (1) is optimized population individual quantity;
s23, selecting an optimal individual to store a genetic algorithm, and performing crossover and mutation, wherein control genes and parameter genes adopt different codes, the control gene codes adopt single-point crossover and basic mutation operators, and the parameter genes adopt global arithmetic crossover and non-uniform mutation operators;
step S24, setting an adaptive crossover probability,
Figure RE-GDA0003558608510000041
wherein, favrAs the average fitness of the population, fminIs the minimum fitness of the population, fcFor small cross individual adaptation value, k1,k2Setting the value to 1;
step S25, setting mutation probability,
Figure RE-GDA0003558608510000042
wherein f ismTo wait for the individual adaptation value of the mutation, k3,k4Setting the value to 0.5;
in step S26, the data is normalized,
Figure BDA0003464437180000043
wherein x isiIs raw data, u is normalized data, xmax、xminIs the maximum value, the minimum value, u of the original datamax、uminAnd respectively normalizing the upper limit and the lower limit of the processed data.
Preferably, the BP neural network predictive model P comprises an input layer, an output layer and at least one hidden layer.
Preferably, in step S3, when the training error precision is greater than r, the retraining BP neural network prediction model P is returned.
Preferably, when the number of times of retraining the BP neural network prediction model P is greater than T, returning to step S2 to re-optimize the structural parameters of the BP neural network prediction model P by using a genetic algorithm.
Preferably, M is greater than or equal to 50, and N is at least greater than or equal to 20.
The invention also provides a system for implementing the intelligent optimization method of the machining parameters of the numerical control electric discharge machine tool, which comprises,
the data operation unit is used for storing and processing M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of a neural network, and using the rest data in the M groups of data as performance test sample data, wherein M, N is positive integer and M is greater than N;
a model construction unit for establishing an initial BP neural network prediction model P, optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, wherein the structural parameters at least comprise hidden layer node number, threshold value and connection weight, and constructing the optimized BP neural network prediction model P1；
A model training unit for normalizing the N groups of data in the data operation unit and training the BP neural network prediction model P1Prediction model P of neural network to be BP1After training error precision is less than or equal to r, inputting residual data in M groups of data to be normalizedFor BP neural network prediction model P1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1；
An optimization processing unit for selecting the numerical control electric spark machine M1Group data at least including pulse interval, pulse width and peak current is normalized and input into BP neural network prediction model P1Obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool.
Preferably, the model training unit and the optimization processing unit both include a normalization processing module for normalization processing of data.
Preferably, the system further comprises a genetic algorithm module for optimizing the structural parameters of the BP neural network prediction model P, wherein the structural parameters at least comprise the number of hidden layer nodes, a threshold value and a connection weight.
According to another aspect of the embodiments of the present invention, there is provided a storage medium including a stored program, wherein when the program runs, a device in which the storage medium is located is controlled to execute the above method.
Compared with the prior art, the numerical control electric spark machine tool machining parameter intelligent optimization method provided by the invention comprises the steps of selecting M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of a neural network, using the residual data in the M groups of data as performance test sample data, establishing an initial BP neural network prediction model P, optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, and constructing the optimized BP neural network prediction model P1(ii) a After N groups of data are normalized, training a BP neural network prediction model P1Waiting for BP neural network prediction model P1After training error precision is less than or equal to r and after normalization processing of residual data in M groups of data is input, a BP neural network prediction model P is subjected to1Performing performance test, and outputting BP neural network prediction model P when training error precision is less than or equal to r1(ii) a Selective numerical control electric spark machine tool M1After group data normalization processing, inputting the data into a BP neural network prediction model P1Obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool; the invention also discloses a system for implementing the method, which optimizes the structural parameters of the BP neural network prediction model P through a genetic algorithm, and then adopts the data of the standard process machining parameters of the numerical control electric spark machine tool to train and test the performance, so that the numerical control electric spark machine tool can adaptively adjust and correct the machining parameters in the machining process, thereby continuously adjusting the pulse interval, the pulse width and the peak value current in the electric spark machining electrode, realizing that the numerical control electric spark machine tool can adaptively adjust the machining parameters according to the next working condition, effectively avoiding the change of the machining parameters caused by electrode abrasion, the machining surface distance and the like in the machining process, and further improving the machining precision and the adaptability of the whole numerical control electric spark machine tool.
Additional features and advantages of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate an embodiment of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of an embodiment of the intelligent optimization method for machining parameters of the numerical control electric discharge machine tool.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the invention, are intended for purposes of illustration and explanation only and are not intended to limit the scope of the invention.
In order to make the technical solutions of the present invention better understood by those skilled in the art, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," "third," "fourth," and the like in the description and in the claims of the invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged under appropriate circumstances such that embodiments of the invention are described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The numerical control electric spark machining method aims to solve a series of problems that in the prior art, in the background technology part, a machine tool is difficult to adaptively adjust and correct machining parameters according to the next working condition, namely the machining parameters in the process cannot be predicted, and the machining deviation is accumulated and superposed due to electrode abrasion, machining surface distance and other changes caused in the numerical control electric spark machining process, so that the machining precision is not high and the like. The invention provides an intelligent optimization method for processing parameters of a numerical control electric spark machine tool, as shown in figure 1, the intelligent optimization method for the processing parameters of the numerical control electric spark machine tool comprises the following steps,
s1, selecting M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of a neural network, and using the rest data in the M groups of data as performance test sample data, wherein M, N are positive integers, and M is larger than N;
step S2, establishing an initial BP neural network prediction model P, and optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, wherein the structural parametersAt least including hidden layer node number, threshold value and connection weight, constructing optimized BP neural network prediction model P1；
Step S3, after the N groups of data in the step S1 are normalized, training a BP neural network prediction model P1Waiting for BP neural network prediction model P1After the training error precision is less than or equal to r and the residual data in the M groups of data are input for normalization processing, a BP neural network prediction model P is subjected to1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1；
Step S4, selecting a numerical control electric spark machine tool M1Group data at least including pulse interval, pulse width and peak current, normalizing, and inputting into BP neural network prediction model P1And obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool.
The invention provides an intelligent optimization method for machining parameters of a numerical control electric spark machine tool, which comprises the steps of selecting M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups from the M groups of data as training sample data of a neural network, using the residual data in the M groups of data as performance test sample data, establishing an initial BP neural network prediction model P, optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, and establishing an optimized BP neural network prediction model P1(ii) a After N groups of data are normalized, training a BP neural network prediction model P1Waiting for BP neural network prediction model P1After training error precision is less than or equal to r, after normalization processing is carried out on residual data in input M groups of data, a BP neural network prediction model P is subjected to1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1(ii) a Selecting numerical control electric spark machine tool M1After group data normalization processing, inputting the data into a BP neural network prediction model P1Obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool; optimizing P structural parameters of the BP neural network prediction model by a genetic algorithm, and thenThe numerical control electric spark machine tool carries out training and performance testing by adopting the standard process machining parameter data of the numerical control electric spark machine tool, so that the numerical control electric spark machine tool carries out self-adaptive adjustment and correction on the machining parameters in the machining process, the pulse interval, the pulse width and the peak current in an electric spark machining electrode are continuously adjusted, the numerical control electric spark machine tool can carry out self-adaptive adjustment on the machining parameters according to the next working condition, the change of the machining parameters caused by electrode abrasion, the machining surface distance and the like in the machining process is effectively avoided, and the machining precision and the self-adaptability of the whole numerical control electric spark machine tool are improved.
For example, in the numerical control electric spark machine tool machining parameter intelligent optimization method provided by the invention, M groups of data of standard process machining parameters are selected for the numerical control electric spark machine tool, wherein M is 50, N groups of data are randomly extracted to serve as training sample data of a neural network, N is 30, and the rest M-N is 20 groups of data to serve as performance test sample data.
In order to make the numerical control electric spark machine tool processing parameter data easier to input into the BP neural network prediction model, under the more preferable condition of the invention, for example, the numerical control electric spark machine tool processing speed vmThe range is 0.002-2.0 g/min, and the numerical span is 1000 times; surface roughness Rmax5-90 μm, and a machining gap Sβ20-280 μm, and converting the processing parameters into a range value from 0 to 1 by normalization treatment.
In order to better adjust and modify the optimization of the processing parameter data input into the neural network, it is necessary to optimize the number of hidden layer nodes, threshold values, connection weights, and other structural parameters in the BP neural network prediction model P, in step S2, the genetic algorithm step includes,
step S21, dividing chromosome structuralization into control genes and parameter genes, optimizing the number of hidden layer nodes of the BP neural network prediction model P by using the control genes, and optimizing the connection weight and the threshold of the BP neural network prediction model P by using the parameter genes;
step S22, a fitness function is set,
f＝a·frmse+b·fcom，a＞0；b＜1
Figure BDA0003464437180000101
Figure BDA0003464437180000102
where n is the number of training data samples, frmseThe root mean square error of the training data, i.e. the square root of the ratio of the square of the deviation of the predicted value from the true value to the number n of samples of training data, yiIn the form of an actual value of the value,
Figure BDA0003464437180000103
is a predicted value; a. b is weight, a, b belongs to (0, 1), Q (0) is original population individual quantity, and Q (1) is optimized population individual quantity;
s23, selecting the optimal individual to store the genetic algorithm, crossing and mutating, wherein, the control gene and the parameter gene adopt different codes, the control gene code adopts single-point crossing and basic mutation operator, the parameter gene adopts global arithmetic crossing and non-uniform mutation operator;
step S24, setting an adaptive crossover probability,
Figure BDA0003464437180000104
wherein f isavrAs the average fitness of the population, fminIs the minimum fitness of the population, fcFor small cross individual adaptation values, k1,k2Setting the value to 1;
step S25, setting the mutation probability,
Figure BDA0003464437180000105
wherein f ismIn order to wait for the individual fitness value of the mutation,k3,k4setting the value to 0.5;
in step S26, the data is normalized,
Figure BDA0003464437180000106
wherein x isiIs raw data, u is normalized data, xmax、xminIs the maximum value, the minimum value, u, of the original datamax、uminAnd respectively normalizing the upper limit and the lower limit of the processed data.
For example, in the genetic algorithm step, the population size is 200, the maximum evolution algebra G is 400, the network complex number adjustment coefficient a is 1-b, b is 0.1, and k1＝k2＝1.0、k3＝k4The initial setting of the BP neural network prediction model P comprises 4 input nodes, 3 output nodes and 15 hidden layer nodes, wherein the BP neural network prediction model P comprises 0.5; setting the threshold values of the initialization state between-2 and 2, setting the weight average of the connection weight of the initialization state between-3 and 3, setting the maximum training times of the BP neural network prediction model P to 2000, and setting the learning rate to 10-2The training error precision r is 10-3. Normalizing the input value of the BP neural network prediction model P, wherein the input value is the data of the numerical control electric spark machine standard process machining parameter M group, and at least comprises the speed vmSurface roughness R of the workpiecemaxElectrode wear ratio θ, machining gap SβInterval of pulse t0Pulse width t1And peak current ieSeven types of parameters. For example, under the parameter condition, after 185 times of iterative computations by the genetic algorithm, the number m of the optimal hidden layer nodes of the BP neural network prediction model P is 11, that is, the network connection structure of the BP neural network prediction model P can be further optimized.
In order to more accurately predict the next machining parameters of the numerical control electric spark machine tool, the pulse interval t is adjusted0Pulse width t1And peak current ieIn a more preferred aspect of the present invention, the BP neural network prediction model P includes an input layer,An output layer and at least one hidden layer.
In order to obtain the BP neural network prediction model P with higher accuracy and require that the BP neural network prediction model P reaches a certain maturity, in a more preferable case of the present invention, in step S3, when the training error accuracy is greater than r, the BP neural network prediction model P is retrained again. In order to improve the prediction accuracy, r is more preferably 10 or less-3。
In order to prevent the retrained BP neural network prediction model P from entering a premature mature state, the method returns to step S2 to perform optimization adjustment on the structural parameters of the neural network prediction model P again by using the genetic algorithm, and in a more preferable case of the present invention, when the number of retraining the BP neural network prediction model P is greater than T, the method returns to step S2 to perform re-optimization on the structural parameters of the BP neural network prediction model P by using the genetic algorithm. In order to take the accuracy of calculation and time cost into consideration, T is more preferably equal to or greater than 10.
In order to enable the neural network prediction model P to obtain a better prediction effect, the data of the standard process machining parameters M of the numerical control electric spark machine tool are not too small, and the training sample data of the neural network which is randomly extracted are not too small, so that the training effect is influenced.
The invention also discloses a system for implementing the intelligent optimization method of the machining parameters of the numerical control electric discharge machine tool, which comprises,
the data operation unit is used for storing and processing M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of a neural network, and using the rest data in the M groups of data as performance test sample data, wherein M, N is positive integer and M is greater than N;
a model construction unit for establishing an initial BP neural network prediction model P, and optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, wherein the structural parameters at least comprise hidden parametersConstructing an optimized BP neural network prediction model P containing the number of layer nodes, a threshold value and a connection weight1；
A model training unit for normalizing the N groups of data in the data operation unit and training the BP neural network prediction model P1After training error precision of the BP neural network prediction model P1 is less than or equal to r, inputting residual data in M groups of data to be normalized, and then carrying out normalization processing on the BP neural network prediction model P1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1；
An optimization processing unit for selecting the numerical control electric spark machine M1Group data at least including pulse interval, pulse width and peak current is normalized and input into BP neural network prediction model P1Obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool.
The invention provides a system for implementing the intelligent optimization method of the machining parameters of the numerical control electric spark machine tool, which selects M groups of data of the standard process machining parameters of the numerical control electric spark machine tool from a data operation unit, randomly selects N groups from the M groups of data as training sample data of a neural network, uses the residual data in the M groups of data as performance test sample data, establishes an initial BP neural network prediction model P in a model construction unit, optimizes the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, and constructs the optimized BP neural network prediction model P1(ii) a After N groups of data are normalized in a model training unit, training a BP neural network prediction model P1Model for predicting neural network to be BP1After training error precision is less than or equal to r, after inputting residual data normalization processing in M groups of data, predicting model P of BP neural network1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1(ii) a Selecting a numerical control electric spark machine tool M in an optimization processing unit1After group data normalization processing, inputting the data into a BP neural network prediction model P1Obtaining optimized processed ginsengCorrecting and adjusting the machining process of the numerical control electric spark machine tool; the system can continuously adjust the pulse interval, the pulse width and the peak current in the electric spark machining electrode, realize that the numerical control electric spark machine tool can adaptively adjust the machining parameters according to the next working condition, effectively avoid the change of the machining parameters caused by electrode abrasion, the distance between the machined surfaces and the like in the machining process, and further improve the machining precision and the adaptability of the whole numerical control electric spark machine tool.
In order to make the dimension and the scale of the input data within a certain range, thereby reducing the dispersity of data distribution and better performing prediction processing on the data input model, in a preferred case of the invention, the model training unit and the optimization processing unit both comprise normalization processing modules for normalization processing of data.
In order to obtain the optimized structural parameters of the BP neural network prediction model through a genetic algorithm and enable the model to have higher self-adaptation degree, under the preferable condition of the invention, the system further comprises a genetic algorithm module which is used for optimizing the structural parameters of the BP neural network prediction model P, wherein the structural parameters at least comprise hidden layer node number, threshold value and connection weight.
The embodiment of the invention also provides a storage medium, which comprises a stored program, wherein when the program runs, the device where the storage medium is located is controlled to execute the method.
It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art will also appreciate that the embodiments described in the specification are presently preferred and that no acts or modules are necessarily required of the invention.
In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus can be implemented in other manners. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is merely a logical division, and the actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of some interfaces, devices or units, and may be an electric or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in the form of hardware, or may also be implemented in the form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a separate product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a mobile terminal, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk, which can store program codes.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. An intelligent optimization method for processing parameters of a numerical control electric spark machine tool is characterized by comprising the following steps,
s1, selecting M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of a neural network, and using the rest data in the M groups of data as performance test sample data, wherein M, N are positive integers, and M is greater than N;
step S2, establishing an initial BP neural network prediction model P, optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, wherein the structural parameters at least comprise hidden layer node number, threshold value and connection weight, and constructing the optimized BP neural network prediction model P1；
Step S3, after the N groups of data in the step S1 are normalized, training a BP neural network prediction model P1Waiting for BP neural network prediction model P1After training error precision is less than or equal to r and after normalization processing of residual data in M groups of data is input, a BP neural network prediction model P is subjected to1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1；
Step S4, selecting a numerical control electric spark machine tool M1Group data, including at least pulse interval, pulse width and peak current, normalizedInputting BP neural network prediction model P1And obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool.
2. The intelligent optimization method for machining parameters of the numerical control electric discharge machine according to claim 1, wherein in step S2, the genetic algorithm step comprises,
step S21, dividing chromosome structuralization into control genes and parameter genes, optimizing the number of hidden layer nodes of the BP neural network prediction model P by using the control genes, and optimizing the connection weight and the threshold of the BP neural network prediction model P by using the parameter genes;
step S22, a fitness function is set,
f＝a·frmse+b·fcom,a，b∈(0，1)
Figure FDA0003464437170000021
Figure FDA0003464437170000022
where n is the number of training data samples, frmseThe root mean square error of the training data, i.e. the square root of the ratio of the square of the deviation of the predicted value from the true value to the number n of samples of training data, yiIn the form of an actual value of the value,
Figure FDA0003464437170000023
is a predicted value; a. b is weight, Q (0) is the number of original population individuals, and Q (1) is the number of optimized population individuals;
s23, selecting an optimal individual to store a genetic algorithm, and performing crossover and mutation, wherein control genes and parameter genes adopt different codes, the control gene codes adopt single-point crossover and basic mutation operators, and the parameter genes adopt global arithmetic crossover and non-uniform mutation operators;
step S24, setting an adaptive crossover probability,
Figure DEST_PATH_GDA0003558608510000041
wherein f isavrAs the average fitness of the population, fminIs the minimum fitness of the population, fcFor small cross individual adaptation value, k1,k2Setting the value to 1; step S25, setting the mutation probability,
Figure DEST_PATH_GDA0003558608510000042
wherein f ismTo wait for the individual adaptation value of the mutation, k3,k4Setting the value to 0.5;
in step S26, the data is normalized,
Figure FDA0003464437170000026
wherein x isiIs raw data, u is normalized data, xmax、xminIs the maximum value, the minimum value, u of the original datamax、uminAnd respectively normalizing the upper limit and the lower limit of the processed data.
3. The intelligent optimization method for machining parameters of the numerical control electric discharge machine according to claim 1, wherein the BP neural network prediction model P comprises an input layer, an output layer and at least one hidden layer.
4. The intelligent optimization method for machining parameters of the numerical control electric discharge machine according to claim 1, wherein in step S3, when the precision of the training error is greater than r, the retraining of the BP neural network prediction model P is returned.
5. The intelligent optimization method for machining parameters of the numerical control electric discharge machine according to claim 4, characterized in that when the number of retraining the BP neural network prediction model P is more than T, the method returns to the step S2 to re-optimize the structural parameters of the BP neural network prediction model P by using the genetic algorithm.
6. The intelligent optimization method for the machining parameters of the numerical control electric discharge machine according to any one of claims 1 to 5, characterized in that the value of M is greater than or equal to 50, and the value of N is at least greater than or equal to 20.
7. A system for implementing the intelligent optimization method for the machining parameters of the numerical control electric discharge machine tool according to any one of the claims 1 to 6, which is characterized by comprising,
the data operation unit is used for storing and processing M groups of data of standard process machining parameters of the numerical control electric spark machine tool, randomly selecting N groups of data from the M groups of data as training sample data of the neural network, and using the rest data in the M groups of data as performance test sample data, wherein M, N is a positive integer, and M is greater than N;
a model construction unit for establishing an initial BP neural network prediction model P, optimizing the structural parameters of the BP neural network prediction model P by adopting a genetic algorithm, wherein the structural parameters at least comprise hidden layer node number, threshold value and connection weight, and constructing the optimized BP neural network prediction model P1；
A model training unit for training the BP neural network prediction model P after normalizing the N groups of data in the data operation unit1Waiting for BP neural network prediction model P1After training error precision is less than or equal to r and after normalization processing of residual data in M groups of data is input, a BP neural network prediction model P is subjected to1Performing performance test, and outputting a BP neural network prediction model P when the training error precision is less than or equal to r1；
An optimization processing unit for selecting the numerical control electric spark machine M1Group data at least including pulse interval, pulse width and peak current is normalized and input into BP neural network prediction modelP1And obtaining optimized processing parameters, and correcting and adjusting the processing process of the numerical control electric spark machine tool.
8. The system of claim 7, wherein the model training unit and the optimization processing unit each comprise a normalization processing module for normalization processing of data.
9. The system according to any one of claims 7 to 8, wherein the system further comprises a genetic algorithm module for optimizing structural parameters of the BP neural network prediction model P, wherein the structural parameters at least comprise hidden layer node number, threshold value and connection weight.
10. A storage medium, comprising a stored program, wherein the program, when executed, controls an apparatus in which the storage medium is located to perform the method of any one of claims 1-6.
CN202210027023.6A 2022-01-11 2022-01-11 Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool Pending CN114491839A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210027023.6A CN114491839A (en) 2022-01-11 2022-01-11 Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210027023.6A CN114491839A (en) 2022-01-11 2022-01-11 Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool

Publications (1)

Publication Number Publication Date
CN114491839A true CN114491839A (en) 2022-05-13

Family

ID=81510315

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210027023.6A Pending CN114491839A (en) 2022-01-11 2022-01-11 Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool

Country Status (1)

Country Link
CN (1) CN114491839A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116224930A (en) * 2023-01-17 2023-06-06 扬州市职业大学(扬州开放大学) Processing control method and system for numerically controlled grinder product
CN116415434A (en) * 2023-04-07 2023-07-11 平湖市山特螺纹工具有限公司 Screw tap processing technique and system for high-strength steel

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116224930A (en) * 2023-01-17 2023-06-06 扬州市职业大学(扬州开放大学) Processing control method and system for numerically controlled grinder product
CN116224930B (en) * 2023-01-17 2023-08-22 扬州市职业大学(扬州开放大学) Processing control method and system for numerically controlled grinder product
CN116415434A (en) * 2023-04-07 2023-07-11 平湖市山特螺纹工具有限公司 Screw tap processing technique and system for high-strength steel
CN116415434B (en) * 2023-04-07 2024-05-24 平湖市山特螺纹工具有限公司 Screw tap processing technique and system for high-strength steel

Similar Documents

Publication Publication Date Title
CN114491839A (en) Intelligent optimization method and system for machining parameters of numerical control electric spark machine tool
CN108389001B (en) Intelligent nonlinear process planning method based on STEP-NC
CN110488754B (en) Machine tool self-adaptive control method based on GA-BP neural network algorithm
CN111563301A (en) Thin-wall part milling parameter optimization method
CN110221580B (en) Feed speed optimization method based on main shaft data simulation
CN114384861A (en) Multi-working-condition cutting parameter optimization method for numerical control machine tool based on multilayer perceptron
CN106875094A (en) A kind of multiple target Job-Shop method based on polychromatic sets genetic algorithm
CN105608295B (en) The multi-objective genetic algorithm of coking furnace pressure and RBF neural Optimization Modeling method
CN114925596B (en) Optimization method for finish milling machining parameters of casting platform based on wafer test equipment
CN116522806A (en) Polishing process parameter optimization method, polishing system, electronic device and storage medium
Cao et al. A continuous optimization decision making of process parameters in high-speed gear hobbing using IBPNN/DE algorithm
CN114611379A (en) Machining process energy-saving planning method based on data driving
CN115130749A (en) NSGA-III and TOPSIS fused data-driven multi-objective optimization method
CN115796510A (en) Multi-target flexible job shop scheduling method based on improved variable neighborhood genetic algorithm
Bhuiyan et al. Optimization of cutting parameters in turning process
CN112541694A (en) Flexible job shop scheduling method considering preparation time and workpiece batching
CN113084456A (en) Processing technology of aluminum alloy thin-wall shell
Nayak et al. Parametric appraisal of WEDM using harmony search algorithm
CN116360355A (en) Method for solving workshop scheduling problem based on NSGA-III algorithm
CN113065630B (en) Technological parameter optimization method for abrasive belt grinding
Karnik et al. Integrating Taguchi principle with genetic algorithm to minimize burr size in drilling of AISI 316L stainless steel using an artificial neural network model
CN108829030B (en) Characteristic tool path optimization method for blade root of integral impeller
CN112861433B (en) Product low-carbon design method based on multi-level integrated framework
CN114237166A (en) Method for solving multi-rotating-speed energy-saving scheduling problem based on improved SPEA2 algorithm
CN110096741B (en) Pre-forging forming die design method based on prediction model and improved genetic algorithm

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