CN111948977A - Multi-objective optimization method and system for stainless steel processing - Google Patents
Multi-objective optimization method and system for stainless steel processing Download PDFInfo
- Publication number
- CN111948977A CN111948977A CN202010847400.1A CN202010847400A CN111948977A CN 111948977 A CN111948977 A CN 111948977A CN 202010847400 A CN202010847400 A CN 202010847400A CN 111948977 A CN111948977 A CN 111948977A
- Authority
- CN
- China
- Prior art keywords
- cutting
- stainless steel
- surface roughness
- material removal
- removal rate
- 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
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
- G05B19/19—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by positioning or contouring control systems, e.g. to control position from one programmed point to another or to control movement along a programmed continuous path
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/35—Nc in input of data, input till input file format
- G05B2219/35349—Display part, programmed locus and tool path, traject, dynamic locus
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Manufacturing & Machinery (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Numerical Control (AREA)
Abstract
The invention relates to a multi-objective optimization method and a system for stainless steel processing, which are used for obtaining extreme values of feed rate, cutting speed and cutting depth of a stainless steel material, establishing an L9 orthogonal table and carrying out a cutting experiment, collecting cutting force and surface roughness in the cutting experiment, calculating material removal rate to obtain an experiment data set, normalizing the experiment data set, inputting the normalized experiment data set into a support vector machine, establishing a correlation model of the cutting force, the surface roughness and the material removal rate, and finally optimizing the correlation model by using an NSPSO algorithm to obtain a multi-objective optimized Pareto front and a corresponding cutting parameter combination. The purpose of designing cutting parameters according to energy consumption is achieved, and the method conforms to the current large trend of sustainable development of manufacturing industry.
Description
Technical Field
The invention relates to the field of efficient and high-precision metal processing, in particular to a multi-objective optimization method and system for stainless steel processing.
Background
The optimization of the metal processing process of stainless steel and the like is always a concern in the metal processing field, and the same workpiece material and cutter combination can effectively improve the production efficiency, the part quality, the system stability and the like by changing the cutting parameters. In recent years, due to the introduction of the sustainable manufacturing concept, environmental objectives such as energy consumption and carbon emission are considered in the optimization of metal processing, which brings more challenges to the optimization of metal processing. In addition, machine learning and computer technology development also provide new ideas and methods for metal working optimization.
In the art, conventional methods typically model the optimization objective through empirical models or simple mechanistic models, and then find the optimal cutting parameter combination. However, the traditional empirical model is often limited and has low accuracy, and a detailed mechanism model of the targets such as energy consumption is difficult to establish, so that the research difficulty is increased.
Therefore, at present, a multi-objective optimization method including an energy consumption mechanism model is needed to solve the above problems, so as to satisfy the current sustainable manufacturing trend, consider the processing optimization requirements of metal materials in various aspects, obtain the optimal processing and cutting parameter combination, reduce energy consumption, realize energy saving and environmental protection, and improve the quality and production efficiency of parts.
Disclosure of Invention
Aiming at the defects and improvement requirements in the prior art, the invention provides a multi-objective optimization method and a multi-objective optimization system for stainless steel processing, wherein a virtual model of cutting force and surface integrity is established through a support vector machine based on a data set required by orthogonal experimental design, the material removal rate is calculated through a formula, the cutting force is used for reflecting energy consumption, the surface integrity is used for reflecting part quality, and the material removal rate is used for reflecting production efficiency. The three models are led into a multi-objective algorithm NSPSO, and a final Pareto front edge of multi-objective optimization and a corresponding cutting parameter combination are obtained through iterative updating, so that the cutting parameter combination with low energy consumption and high precision is obtained, energy consumption is reduced, energy conservation and environmental protection are realized, and meanwhile, the product quality and the production efficiency are obviously improved.
In order to achieve the purpose, the invention provides the following scheme:
a multi-objective optimization method for stainless steel processing, comprising:
obtaining extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material;
creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth;
performing a cutting experiment according to the L9 orthogonal table;
collecting cutting force and surface roughness in the cutting experiment, and simultaneously calculating the material removal rate to obtain an experiment data set;
normalizing the experimental data set, inputting the result of the normalization into a support vector machine, and establishing a correlation model of cutting force, surface roughness and material removal rate;
and optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto frontier and a corresponding cutting parameter combination.
The obtaining of the extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material specifically comprises:
and determining the maximum value and the minimum value of the feed rate, the cutting speed and the cutting depth when the tool machines the stainless steel material according to the rated parameters of the machining system and a tool manual.
Creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth, specifically comprising:
taking the feeding rate, the cutting speed and the cutting depth as three factors, and numbering the three factors as (i) - (iii);
respectively selecting the minimum value, the middle value and the maximum value of the feed rate, the cutting speed and the cutting depth from the extreme value intervals of the feed rate, the cutting speed and the cutting depth, taking the minimum value, the middle value and the maximum value as three levels, and respectively numbering the three levels to be 1-3;
creating an L9 orthogonal table according to the three factors and the three levels; the L9 orthogonal table is a table with 3 columns and 9 rows, the number of columns represents factors, the number of rows represents experiment times, and the number of the experiment times is one to nine;
the number of 9 experiments was grouped by the number of levels of 3 so that each 3 experiments was 1 group, and the level number of each factor was filled in the L9 orthogonal table according to the filling method of the orthogonal table.
The cutting experiment is carried out according to the L9 orthogonal table, and the method specifically comprises the following steps:
carrying out nine groups of experiments according to the experimental information in the L9 orthogonal table, wherein each group of experiments is repeated three times;
and calculating the average value of the three experimental results, and taking the average value as the final experimental result.
The method comprises the following steps of collecting cutting force and surface roughness in a cutting experiment, and calculating the material removal rate to obtain an experiment data set, wherein the experiment data set specifically comprises the following steps:
collecting cutting force of a cutter for processing the stainless steel material and surface roughness of the stainless steel material in the cutting experiment to obtain cutting force data and surface roughness data;
calculating the material removal rate of the stainless steel material after the machining is finished by using a formula MRR (maximum ratio of R) 1000Vdf to obtain material removal rate data; wherein MRR represents the material removal rate in mm3Min; v represents the cutting speed in m/min; d represents the depth of cut in mm; f represents the feed rate in mm/r;
obtaining an experimental data set according to the cutting force data, the surface roughness data and the material removal rate data;
the cutting force is collected through a Kistler dynamometer, the Kistler dynamometer obtains cutting force signals in three directions, the average value of a stable section is taken as a cutting force result in each direction, and the resultant force in the three directions is calculated as final cutting force data;
the surface roughness was measured by a Mahr roughness meter, which takes the mean of the plateaus as the final surface roughness data.
The normalizing treatment is carried out on the experimental data set, then the result of the normalizing treatment is input into a support vector machine, and correlation models of cutting force, surface roughness and material removal rate are respectively established, and the method specifically comprises the following steps:
respectively carrying out normalization pretreatment on all data in the experimental data set, and normalizing all data to be in a [0, 1] interval;
selecting a radial basis function as a kernel function of a support vector machine model, and determining optimal parameters C and sigma of the support vector machine classification model through a cross verification method and a grid search method, wherein C represents a penalty coefficient, and sigma represents a width parameter;
respectively establishing a cutting force support vector machine model and a surface roughness support vector machine model by utilizing the optimal parameters C and sigma through a network open source code LIBSVM3.22 of the support vector machine;
obtaining a material removal rate formula model through the calculation formula of the material removal rate;
the correlation model includes the cutting force support vector machine model, the surface roughness support vector machine model, and the material removal rate formula model.
And respectively carrying out normalization processing on all data in the experimental data set by adopting a normalization formula as follows:
wherein N is*For the normalization result, N is the initial value, NmaxAnd NminThe maximum value and the minimum value of the initial value are respectively.
Optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto frontier and a corresponding cutting parameter combination, and specifically comprising the following steps:
introducing the cutting force support vector machine model, the surface roughness support vector machine model and the material removal rate formula model into an NSPSO intelligent optimization algorithm;
generating an initial population randomly within the speed and position range of the particles by utilizing the NSPSO intelligent optimization algorithm;
carrying out normalization pretreatment on the position of the initial population, and calculating the fitness of each particle in the initial population based on an objective function;
updating the current population according to a speed and position updating formula of an NSPSO intelligent optimization algorithm, and merging the child and the parent to perform non-dominated solution set sorting; the child represents the updated particle swarm, and the parent represents the particle swarm before updating;
and generating a new current generation population, updating the individual extreme value and the population extreme value, and repeating iterative updating to obtain a final Pareto front edge and an optimal population.
The current population is updated according to an update formula of the speed and the position of the NSPSO intelligent optimization algorithm, wherein the update formula is as follows:
wherein v isi，jAnd pi，jRespectively representing the speed and the position of the ith particle in the jth generation; v. ofi，j+1And pi，j+1Respectively representing the speed and the position of the ith particle in the j +1 th generation; c. C1And c2Represents the acceleration factor, r1And r2Is a random number between 0 and 1; pi，bestIs the historically optimal individual for the ith particle, GbestThe method comprises the following steps of (1) selecting a global optimal individual randomly in a first-layer non-dominated solution set;
ω represents an inertia factor, expressed as:
wherein j represents the current algebra, and N is the maximum iteration number;
the generating a new current generation population, updating individual extreme values and population extreme values, and repeating iterative updating to obtain a final Pareto front edge and an optimal population specifically comprises the following steps:
sequentially selecting N individuals ranked in the front to become new parents according to the ranking level of the non-dominated solution set ranking;
when the number of the selected non-dominated solution sets on a certain layer overflows, sorting the individuals on the current layer from large to small according to the congestion distance, sequentially taking out the individuals, putting the individuals into the next parent generation until the upper limit of the population is reached, and entering the next updating;
and repeating the updating steps until the maximum iteration times are reached to obtain a final Pareto front edge and an optimal population, wherein the final Pareto front edge and the optimal population are the optimal cutting parameter combination of the multi-objective optimization.
The invention also provides a multi-objective optimization system for stainless steel processing, which comprises the following components:
the parameter extreme value acquisition module is used for acquiring extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material;
the orthogonal table creating module is used for creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth;
the cutting experiment implementation module is used for carrying out a cutting experiment according to the L9 orthogonal table;
the experimental data generation module is used for acquiring cutting force and surface roughness in the cutting experiment, and calculating the material removal rate to generate an experimental data set;
the correlation model creating module is used for carrying out normalization processing on the experimental data set, then inputting the normalization processing result into a support vector machine, and establishing a correlation model of cutting force, surface roughness and material removal rate;
and the optimization obtaining parameter module is used for optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto front edge and a corresponding cutting parameter combination.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the cutting force is used for representing energy consumption, the smaller the numerical value of the cutting force is, the smaller the machining amount of the cutter in machining a workpiece is, and the smaller the machining amount is, the lower the power consumption, the cutter loss and the like are, namely, the lower the energy consumption is; the quality of the part is represented by the surface roughness, and the smaller the surface roughness value is, the smoother the surface of the workpiece is, and the higher the quality of the workpiece is; the production efficiency is represented by a material removal rate, and the smaller the material removal rate value is, the smaller the amount of cut of the workpiece is, the shorter the machining cycle is, and the higher the production efficiency is. Cutting parameters are directly designed according to cutting force, surface roughness and material removal rate, the optimal cutting parameters are input into machining center equipment, production efficiency and part quality can be directly improved, energy consumption is obviously reduced, the cutting process is controlled in a high-quality and high-efficiency mode, and various optimized parameter combinations are provided to meet different requirements.
The method utilizes the cutting force to characterize the energy consumption, solves the problems that the traditional method is difficult to establish a mechanism model aiming at the energy consumption and design the cutting parameters according to the energy consumption, meets the requirement of current sustainable manufacturing, gets rid of the constraint of the traditional experience model, can accurately establish the relationship between a complex target and input parameters, and is not limited by dimension.
According to the invention, a correlation model of cutting force, surface roughness and material removal rate is established through a support vector machine-SVR, and the whole process can be realized through a network open source code LIBSVM3.22 of the support vector machine. The correlation model is optimized by using an NSPSO intelligent optimization algorithm, and the obtained cutting parameters are more accurate and the machining precision is effectively improved by multiple times of normalization processing and iterative updating.
The invention designs an experimental scheme by using an L9 orthogonal table, wherein L represents the orthogonal table, 9 represents the line number of the orthogonal table, namely the required experimental frequency, fully utilizes the characteristics of uniform dispersibility and orderliness of the L9 orthogonal table, replaces the artificially and subjectively specified experimental frequency, and ensures that the experimental scheme is more balanced, the experimental frequency is less, the efficiency is high, and the experimental result is more accurate and reliable.
According to the invention, a series of optimal parameter combinations obtained through a multi-objective optimization algorithm can be directly selected for actual operators as required to guide the stainless steel processing process and the debugging work of a processing center.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a schematic flow diagram of a multi-objective optimization method for stainless steel processing;
FIG. 2 is a schematic diagram of a multi-objective optimization system for stainless steel processing;
FIG. 3 is a fitting result of a resultant cutting force and a surface roughness based on a support vector machine model;
FIG. 4 is a schematic diagram of an optimization flow of the NSPSO intelligent optimization algorithm;
FIG. 5 is a graph of the Pareto front particle three-dimensional spatial distribution with respect to cutting force, surface roughness, and material removal rate.
Detailed Description
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 derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a multi-objective optimization method and a multi-objective optimization system for stainless steel processing, which start from extreme values of cutting parameters such as feed rate, cutting speed and cutting depth, design a cutting experimental scheme based on an L9 orthogonal table, utilize target parameters of cutting force, surface integrity and material removal rate to characterize energy consumption by cutting force, characterize part quality by surface integrity and production efficiency by material removal rate, establish a correlation model of the three and introduce the correlation model into a multi-objective algorithm NSPSO for optimization, obtain a final Pareto front of multi-objective optimization and a corresponding cutting parameter combination through iterative updating, thereby obtaining a cutting parameter data set with high precision, high quality and high efficiency, considering energy consumption, production quality and efficiency, not only effectively reducing the production requirement, realizing energy conservation and environmental protection, and meeting the current sustainable manufacturing trend, and the cutting precision can be obviously improved, so that the product quality and the production efficiency are improved, and the development of enterprises and cutting processes is facilitated.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Examples
As shown in fig. 1, the present embodiment shows a multi-objective optimization method for stainless steel processing, comprising the steps of:
s1: selecting a stainless steel material to be processed, and acquiring extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material;
the extreme values are the maximum value and the minimum value of the feed rate, the cutting speed and the cutting depth.
Wherein the feed rate (Vf) represents the speed at which a reference point on the tool moves along a tool trajectory relative to the workpiece; the cutting speed (VC) represents the instantaneous speed of a point on the cutting edge of the tool in the main direction of motion with respect to the surface to be machined; the cutting depth (AP) represents the perpendicular distance between the machined surface and the surface to be machined when cutting the workpiece. For greater simplicity and clarity, the present application uses f to denote feed rate in mm/r, V to denote cutting speed in m/min, and the letter d to characterize depth to denote cutting depth in mm.
The maximum and minimum values for the feed rate, cutting speed and depth of cut are typically determined by the nominal parameters of the machining system and the tool manual. In practice, the maximum and minimum values of the feed rate, the cutting speed and the cutting depth are usually determined by an operator by combining the rated parameters of the machining equipment and the tool parameters in a tool manual after considering the stability and realizability of the machining system, because the parameters of the machining equipment are different and the parameters of the tool for machining stainless steel are also different.
The tool used in this embodiment is a cemented carbide tool, which is of the designation CNMG120404-BF, and it should be noted that the machining tool should be determined according to the actual machining condition, and should not be limited to this tool of this embodiment, for example, various tools having tool shanks such as HSK and BT, or all tools made of other materials, can be applied to the technical solution of the present invention.
For the sake of clarity, the present embodiment selects a set of extreme data of feed rate, cutting speed and cutting depth, namely, feed rate f: 0.05-0.15mm/r, cutting speed V: 40-120m/min, cutting depth d: 0.4-2.0mm, but the data should not be regarded as the limiting condition of the invention, and the actual extreme value data should be determined according to the actual processing situation. When different machining centers or tools are used, different extreme values can be used, but the method and principle do not change.
It should be noted that, in the multi-objective optimization method for stainless steel processing proposed in this embodiment, the material of the selected blank is stainless steel AISI304, but the material of the blank to be processed should not be limited to only stainless steel, but may also be applied to blanks made of various materials such as nodular cast iron and chromium. All materials can be suitable for the multi-objective optimization method provided that the materials conform to the machining modes of milling, boring, drilling and tapping and the like.
S2: creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth;
the specific process of creating the L9 orthogonal table comprises the following steps:
taking the feeding rate, the cutting speed and the cutting depth as three factors, and numbering the three factors as (i) - (iii);
respectively selecting the minimum value, the middle value and the maximum value of the feed rate, the cutting speed and the cutting depth from the extreme value intervals of the feed rate, the cutting speed and the cutting depth, taking the minimum value, the middle value and the maximum value as three levels, and respectively numbering the three levels to be 1-3;
creating an L9 orthogonal table according to the three factors and the three levels; the L9 orthogonal table is a table with 3 columns and 9 rows, the number of columns represents factors, the number of rows represents experiment times, and the number of the experiment times is one to nine;
the number of 9 experiments was grouped by the number of levels of 3 so that each 3 experiments was 1 group, and the level number of each factor was filled in the L9 orthogonal table according to the filling method of the orthogonal table.
L9 is an orthogonal table commonly used in Taguchi design method, wherein "L" represents the orthogonal table, "9" represents the number of rows of the orthogonal table, i.e. the required experiment times, and L9 orthogonal table has the characteristics of uniform dispersion and order comparability. The following two characteristics must be satisfied when creating orthogonal tables: 1) the number of occurrences of different numbers in each column is equal. The characteristic shows that the probability of each level of each factor participating in the experiment is completely the same as that of each level of other factors, thereby ensuring that the interference of the levels of other factors is eliminated to the maximum extent in each level, effectively comparing the experiment results and finding out the optimal experiment condition. 2) In any two columns of its transversely organized pairs, each pair appears equally many times. This feature ensures that the experimental points are evenly distributed among the complete combination of factors and levels and are therefore highly representative.
In this embodiment, with the above method of creating the L9 orthogonal table, first, the feed rate f: 0.05-0.15mm/r, cutting speed V: 40-120m/min, cutting depth d: the data set of 0.4-2.0mm is summarized in the three-factor three-level experiment as shown in Table 1:
TABLE 1
The data in level 1 are the minimum values of the feed rate, the cutting speed and the cutting depth, the data in level 3 correspond to the maximum values of the feed rate, the cutting speed and the cutting depth, and the data in level 2 are the median values calculated according to the maximum values and the minimum values of the feed rate, the cutting speed and the cutting depth.
An L9 orthogonal table created based on the above three factors and three levels is shown in table 2, the L9 orthogonal table in table 2 is a table of 3 columns and 9 rows (here, one column of the test number and one row of the factor are not included in the number of columns and the number of rows), the number of columns represents the factor, the number of rows represents the number of experiments, and the number of the experiment numbers is one to nine;
TABLE 2
In this embodiment, the process of filling levels 1 to 3 into the L9 orthogonal table only needs to satisfy the two aforementioned characteristics that must be satisfied when creating the orthogonal table, and is not described herein again. Moreover, it should be noted that table 2 of this embodiment is only one of the filling methods, and the L9 orthogonal table is not unique, and the L9 orthogonal table that meets the two characteristics of the orthogonal table has the same function, and can achieve the effects of reducing the number of experiments and improving the accuracy and efficiency of the experiments.
S3: performing a cutting experiment according to the L9 orthogonal table;
in this embodiment, nine sets of experiments are performed according to the experimental information in the L9 orthogonal table, each set of experiments is repeated three times, then an average value of the results of the three experiments is calculated, and the average value is used as a final experiment result, and the specific experiment results are shown in table 3:
TABLE 3
It should be noted that the cutting force in this embodiment is also called a total cutting force, as shown in table 3, the cutting force is obtained by using a Kistler dynamometer in three directions of x, y, and z, an average value of the stable segments is taken as a final cutting force component in each direction, the total cutting force is calculated as a final cutting force according to the cutting force components in the three directions, and Fx, Fy, and Fz correspond to the force components in the three directions of x, y, and z, respectively.
Usually, when designing a cutting experiment, an operator often depends on subjective consciousness to design the experiment times at will, once the size precision of a finished workpiece processed by a set of cutting parameters does not reach the standard, the next experiment debugging is continued, and the experiment is stopped until the precision of each size of the finished workpiece is qualified. A large number of experiments are not planned, so that time and labor are wasted, and the cutting cost is wasted. In the embodiment, the L9 orthogonal table is used for designing the experimental scheme, the characteristics of uniform dispersibility and uniformity of the L9 orthogonal table are fully utilized, artificial subjective regulation of experimental times is replaced, the experimental times are reduced, the influence of experimental variables on results is obtained as far as possible, the experimental scheme is more balanced, the experimental efficiency is higher, and the experimental results are more accurate and reliable.
S4: collecting cutting force and surface roughness in the cutting experiment, and simultaneously calculating the material removal rate to obtain an experiment data set;
the experimental data set in this embodiment includes cutting parameters including cutting depth, cutting speed, and feed rate, and target parameters including cutting force, surface roughness, and material removal rate.
The cutting force and the surface roughness are acquired by an instrument in the experimental process, and the material removal rate is calculated by using a formula and specifically comprises the following steps:
collecting cutting force of a cutter for processing the stainless steel material and surface roughness of the stainless steel material in the cutting experiment to obtain cutting force data and surface roughness data;
calculating the material removal rate of the stainless steel material after the machining is finished by using a formula MRR (maximum ratio of R) 1000Vdf to obtain material removal rate data;
wherein MRR represents the material removal rate in mm3Min; v represents the cutting speed in m/min; d represents the depth of cut in mm; f represents the feed rate in mm/r;
obtaining an experimental data set according to the cutting force data, the surface roughness data and the material removal rate data;
the cutting force was collected by a Kistler dynamometer in units of N. The Kistler dynamometer obtains cutting force signals in three directions, the average value of the stable section is taken as a cutting force result in each direction, and the resultant force in the three directions is calculated to serve as final cutting force data;
the surface roughness is measured by a Mahr roughness meter in μm. The Mahr coarseness gauge takes the mean of the plateaus as the final surface roughness data.
The Kistler dynamometer and the Mahr roughness gauge are common measurement tools in the machining field, and therefore, the Kistler dynamometer measures the cutting force and the Mahr roughness gauge measures the surface roughness in a specific process, which is not described herein in detail, and other brands and types of dynamometers or roughness gauges can be used.
S5: normalizing the experimental data set, inputting the result of the normalization into a support vector machine, and establishing a correlation model of cutting force, surface roughness and material removal rate;
in this embodiment, the experimental data set is normalized to avoid interference caused by different dimensions of three different dimensions, namely, cutting force, surface roughness and material removal rate, and a specific normalization formula is as follows:
wherein N is*For the normalization result, N is the initial value, NmaxAnd NminThe maximum value and the minimum value of the initial value are respectively.
In this embodiment, a correlation model of the cutting force, the surface roughness, and the material removal rate is established, and the specific process is as follows:
respectively carrying out normalization pretreatment on all data in the experimental data set, and normalizing all data to be in a [0, 1] interval;
selecting a radial basis function as a kernel function of a support vector machine model, and determining optimal parameters C and sigma of the support vector machine classification model through a cross verification method and a grid search method, wherein C represents a penalty coefficient, and sigma represents a width parameter; the optimal parameters for the final determination of the cutting force in this embodiment are: c3.7321, σ 0.134, the optimum parameters of the surface roughness are: c1 and σ 0.7071, where the values of C and σ are determined by the input experimental data set, so the values of C and σ are not unique, depending on the actual production situation;
respectively establishing a cutting force support vector machine model and a surface roughness support vector machine model by utilizing the optimal parameters C and sigma through a network open source code LIBSVM3.22 of the support vector machine;
obtaining a material removal rate formula model through a calculation formula MRR of the material removal rate being 1000 Vdf;
the correlation model includes the cutting force support vector machine model, the surface roughness support vector machine model, and the material removal rate formula model.
In the embodiment, a cutting force support vector machine model and a surface roughness support vector machine model are established by using a support vector machine model, a material removal rate formula model is established by using a material removal rate formula, and a correlation model of the cutting force, the surface roughness and the material removal rate is established to link the cutting force, the surface roughness and the material removal rate, so that the cutting force represents energy consumption, the smaller the cutting force value is, the smaller the machining amount of a cutter in machining a workpiece is, the smaller the machining amount is, the lower the power consumption and the cutter loss are, and the lower the energy consumption is; the quality of the part is represented by the surface roughness, and the smaller the surface roughness value is, the smoother the surface of the workpiece is, and the higher the quality of the workpiece is; the production efficiency is represented by a material removal rate, and the smaller the material removal rate value is, the smaller the amount of cut of the workpiece is, the shorter the machining cycle is, and the higher the production efficiency is.
The cutting force is used for representing energy consumption, the problems that a mechanism model is difficult to establish aiming at the energy consumption and cutting parameters are difficult to design according to the energy consumption in the traditional method are solved, the constraint of a traditional experience model is broken away, the relation between a complex target and input parameters can be accurately established, the method is not limited by dimensions, the method accords with the major trend of sustainable development of the current manufacturing industry, the energy conservation and the environmental protection are realized, and meanwhile, the service lives of processing equipment and a cutter can be effectively prolonged.
S6: and optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto frontier and a corresponding cutting parameter combination.
In this embodiment, the association model is optimized by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto frontier and a corresponding cutting parameter combination, and the specific process is as follows:
introducing the cutting force support vector machine model, the surface roughness support vector machine model and the material removal rate formula model into an NSPSO intelligent optimization algorithm; generating an initial population randomly within the speed and position range of the particles by utilizing the NSPSO intelligent optimization algorithm; carrying out normalization pretreatment on the position of the initial population, and calculating the fitness of each particle in the initial population based on an objective function; updating the current population according to a speed and position updating formula of an NSPSO intelligent optimization algorithm, and merging the child and the parent to perform non-dominated solution set sorting; the child represents the updated particle swarm, and the parent represents the particle swarm before updating; and generating a new current generation population, updating the individual extreme value and the population extreme value, and repeating iterative updating to obtain a final Pareto front edge and an optimal population.
The method for optimizing the material removal rate includes the following steps of introducing the cutting force support vector machine model, the surface roughness support vector machine model and the material removal rate formula model into an NSPSO intelligent optimization algorithm, and specifically includes the following steps:
and inputting basic parameters of an NSPSO intelligent optimization algorithm into the cutting force support vector machine model, the surface roughness support vector machine model and the material removal rate formula model. The basic parameters are shown in table 4, each particle in table 4 includes cutting speed, cutting depth, and feeding speed, and the particle position range is an extreme value range of the cutting speed, the cutting depth, and the feeding speed.
The NSPSO intelligent optimization algorithm is also called as a particle swarm optimization algorithm, and the principle is that each particle in a particle swarm represents a possible solution of a problem, and the intelligence of problem solution is realized through the simple behavior of individual particles and the information interaction in the swarm.
The particle dimension is three-dimensional and comprises two characteristics of position and speed, wherein the position represents cutting speed, cutting depth and feeding speed respectively, and the speed represents the corresponding particle evolution direction. An initial population is randomly generated within the range of the speed and the position of the particles, and the position of the initial population is normalized. The initial population is generated by taking random numbers, namely at a feed rate: 0.05-0.15 mm/r; cutting speed: 40-120 m/min; cutting depth: and randomly generating 100 groups as an initial population within the range of 0.4-2.0mm, and carrying out normalization processing on the initial population to finally obtain 100 particles, wherein each particle comprises three dimensions, namely cutting speed, cutting depth and feed rate, and each dimension belongs to the range of [0, 1] after normalization.
TABLE 4
The calculating the fitness of each particle in the initial population based on an objective function specifically includes:
the objective functions of the cutting force and the surface roughness are calculated by respective support vector machine models, the process is realized through LIBSVM3.22, the normalized particle position matrix is used as input data and is input into a code of the LIBSVM, the cutting force corresponding to each particle can be calculated by giving the optimal C and sigma values of a cutting force support vector regression machine, and similarly, the optimal C and sigma values of the surface roughness support vector machine model are given, and the surface roughness corresponding to each particle can be calculated. The cutting force and the surface roughness are respectively the first dimension and the second dimension of the fitness, and the material removal rate MRR of the fitness in the third dimension is calculated by a material removal rate formula.
Updating the current population according to a speed and position updating formula of an NSPSO intelligent optimization algorithm, and merging the child and the parent to perform non-dominated solution set ordering; the child represents the updated particle swarm, and the parent represents the particle swarm before updating, which specifically includes:
the speed matrix and the position matrix of 100 particles are sequentially substituted into an updating formula, when updating for the first time, the particle swarm to be updated is an initial population, in the later updating, the particle swarm to be updated is a current latest population, a new position matrix and a new speed matrix can be obtained after updating, and the updating formula is as follows:
wherein v isi，jAnd pi，jRespectively representing the speed and the position of the ith particle in the jth generation; v. ofi，j+1And pi，j+1Respectively representing the speed and the position of the ith particle in the j +1 th generation; c. C1And c2Represents the acceleration factor, r1And r2Is a random number between 0 and 1; pi，bestIs the historically optimal individual for the ith particle, GbestThe method comprises the following steps of (1) selecting a global optimal individual randomly in a first-layer non-dominated solution set;
ω represents an inertia factor, expressed as:
wherein j represents the current algebra, and N is the maximum iteration number.
The generating a new current generation population, updating individual extreme values and population extreme values, and repeating iterative updating to obtain a final Pareto front edge and an optimal population specifically comprises the following steps:
sequentially selecting N individuals ranked in the front to become new parents according to the ranking level of the non-dominated solution set ranking;
when the number of the selected non-dominated solution sets on a certain layer overflows, sorting the individuals on the current layer from large to small according to the congestion distance, sequentially taking out the individuals sorted according to the congestion distance, placing the individuals into the next parent generation until the upper limit of the population is reached, and entering the next updating;
and repeating the updating steps until the maximum iteration times are reached to obtain a final Pareto front edge and an optimal population, wherein the final Pareto front edge and the optimal population are the optimal cutting parameter combination of the multi-objective optimization.
After updating, obtaining new position matrixes and speed matrixes of 100 particles, wherein the 100 particles are the current latest child particle group, the particle group before updating is called a parent, calculating the fitness corresponding to each particle according to the position matrix of the child individual according to the step of calculating the fitness, similarly, the fitness of each particle also comprises three dimensions of cutting force, surface roughness and material removal rate, merging the child and the parent for non-dominated solution set sorting, and for convenience of understanding, introducing related concepts first.
Dominating: in the multi-objective optimization, if each target of an individual a is better than that of an individual b, the individual a dominates the individual b or the individual b is dominated by the individual a, and if the targets of the individual a and the individual b are good and bad, the individual a and the individual b have no domination and dominated relationship and belong to the same level.
Non-dominated solution set: in a population consisting of a group of individuals, if an individual a is not dominated by any other individual, the individual a is called a non-dominated solution of the population, and the set of all the non-dominated solutions is the non-dominated solution set of the population.
Non-dominated solution set ordering: the non-dominant solution set of the whole population is called a first-layer non-dominant solution set, after the first-layer non-dominant solution set is excluded from the population, the non-dominant solution set is selected from the remaining individuals, called a second-layer non-dominant solution set, the second-layer non-dominant solution set is excluded from the population, the third-layer non-dominant solution set is selected from the remaining individuals, and so on until no dominant relationship exists among the remaining individuals in the population, the solution sets belong to the same level, and the individuals are the non-dominant solution set at the bottommost layer. The process of dividing the whole population into multiple layers of non-dominant solution sets according to the method is called non-dominant solution set ordering.
Ranking level: the ranking level of the non-dominant solution set to which the individual belongs is the ranking level, and the smaller the ranking level is, the higher the non-dominant level to which the individual belongs is, the better the individual is.
In this embodiment, the children and the parents are merged according to the above non-dominated solution set sorting process to perform non-dominated solution set sorting, and then the first 100 individuals are sequentially selected to become new parents according to a small-to-large sorting level, where the small-to-large sorting level is the order sorting of the individuals from superiority to inferiority. It should be noted here that the number of parent individuals to be selected may be any number, as long as the selection in the sorting order is ensured, but the number is not too small, otherwise, the number of non-dominated solutions cannot overflow, and is not too large, otherwise, the selection difficulty is increased.
In the embodiment, the relevance model of the cutting force, the surface roughness and the material removal rate is established through the support vector machine-SVR, and the whole process can be realized through the network open source code LIBSVM3.22 of the support vector machine, so that the method is convenient, quick and easy to realize. The correlation model is optimized by using an NSPSO intelligent optimization algorithm, and normalization processing and iterative updating are carried out for multiple times, so that the obtained cutting parameters are more accurate, the processing precision is effectively improved, and the product quality is ensured.
According to the invention, the optimal cutting parameter combination is obtained according to the cutting force, the surface roughness and the material removal rate, and the cutting parameter combination is input into the machining center equipment, so that the quality and the production efficiency of parts can be obviously improved, the energy consumption is obviously reduced, the cutting process is controlled with high quality and high efficiency, and various optimized parameter combinations are provided to adapt to different requirements.
FIG. 2 is a schematic diagram of a multi-objective optimization system for stainless steel processing. As shown in the figure, the embodiment further provides a multi-objective optimization system for stainless steel processing, which specifically includes:
the parameter extreme value acquisition module is used for acquiring extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material;
the orthogonal table creating module is used for creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth;
the cutting experiment implementation module is used for carrying out a cutting experiment according to the L9 orthogonal table;
the experimental data generation module is used for acquiring cutting force and surface roughness in the cutting experiment, and calculating the material removal rate to generate an experimental data set;
the correlation model creating module is used for carrying out normalization processing on the experimental data set, then inputting the normalization processing result into a support vector machine, and establishing a correlation model of cutting force, surface roughness and material removal rate;
and the optimization obtaining parameter module is used for optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto front edge and a corresponding cutting parameter combination.
In the cutting experiments of this example, the tool is preferably a cemented carbide tool of the brand CNMG 120404-BF. It should be noted that the machining tool is determined according to the actual machining situation, and should not be limited to this brand of tool in this embodiment, for example, all tools having various tool shanks such as HSK, BT, etc. or tools made of other materials can be applied to the technical solution of the present invention, and all the tools should fall within the protection scope of the present invention.
Fig. 3 is a diagram showing the fitting results of the resultant cutting force and the surface roughness based on a support vector machine model. Fig. 3 shows a comparison between the model prediction result obtained by the support vector machine and the actual measurement result after 9 experiments. As can be seen from the figure, the support vector machine can effectively fit the input parameter combination, the cutting force resultant force and the surface roughness, wherein the predicted result of the surface roughness and the experimental result have higher goodness of fit, and therefore, the predicted performance of the surface roughness is better.
Fig. 4 is a schematic diagram of the optimization flow of the NSPSO intelligent optimization algorithm. Each particle comprises cutting speed, cutting depth and feeding speed, an initial population is randomly generated in the speed and position range of the particle, and the position of the initial population is normalized. And then calculating the fitness of each particle based on an objective function, wherein the objective functions of the cutting force and the surface roughness are support vector regression machines, and the material removal rate is calculated by a formula. After the fitness is calculated, updating the current population according to a speed and position updating formula of an NSPSO algorithm, wherein the updating formula is as follows:
wherein v isi，jAnd pi，jRespectively representing the speed and the position of the ith particle in the jth generation; v. ofi，j+1And pi，j+1Respectively representing the speed and the position of the ith particle in the j +1 th generation; c. C1And c2Represents the acceleration factor, r1And r2Is a random number between 0 and 1; pi，bestIs the historically optimal individual for the ith particle, GbestThe method comprises the following steps of (1) selecting a global optimal individual randomly in a first-layer non-dominated solution set;
ω represents an inertia factor, expressed as:
wherein j represents the current algebra, and N is the maximum iteration number.
After the updating is finished, the fitness of the new generation of individuals is calculated, the offspring and the parents are merged for non-dominated solution set sorting, the first 100 individuals are sequentially selected to become the new parents according to the sorting level, when the number of the non-dominated solution sets in a certain layer is selected to overflow, the individuals in the current layer are sorted from large to small according to the crowding distance, the individuals are sequentially taken out and placed into the next parent, and the next updating is carried out until the upper limit of the population is reached. And repeating the updating steps until the maximum iteration times are reached to obtain the final Pareto front edge and the optimal population.
Fig. 5 is a Pareto front particle three-dimensional spatial distribution diagram about cutting force, surface roughness and material removal rate, each particle represents a target value corresponding to a set of cutting parameters, and through the Pareto front, an operator can select a suitable cutting parameter combination according to actual target requirements to achieve the purpose of optimizing processing.
Table 5 shows the combinations of 20 sets of cutting parameters and their corresponding target values. As can be seen from table 5, the four sets of data with experimental numbers 1, 6, 8 and 14 have relatively small values of cutting force, surface roughness and material removal rate, the cutting force represents energy consumption (the smaller the value of cutting force, the smaller the amount of machining, the less the tool performs work on the workpiece, the less power consumption, tool loss and the like, the lower the energy consumption), the surface roughness represents product quality (the smaller the value of surface roughness, the smoother the surface of the workpiece, and the higher the quality of the workpiece), and the material removal rate represents production efficiency (the smaller the value of material removal rate, the less the amount of machining the workpiece, the higher the production efficiency), which means that the four sets of data with experimental numbers 1, 6, 8 and 14 have less machining amount, higher the quality of the workpiece, and lower the energy consumption, and therefore, the four sets of data with experimental numbers 1, 6, 8 and 14 have lower energy consumption, the cutting parameter combinations corresponding to the cutting speed, the cutting depth and the feeding rate are the cutting parameter combinations with low energy consumption, high product quality and high production efficiency, wherein the cutting parameter combination of the 1 st group is optimal.
Therefore, the cutting parameter combination obtained by the invention can effectively improve the production efficiency and the part quality and reduce the energy consumption.
TABLE 5
In summary, the system for the multi-objective optimization method for stainless steel processing provided by the invention starts with extreme values of cutting parameters such as feed rate, cutting speed and cutting depth, designs a cutting experimental scheme based on an L9 orthogonal table, utilizes cutting force, surface integrity and material removal rate to characterize energy consumption by cutting force, characterize part quality by surface integrity, characterize production efficiency by material removal rate, establishes a correlation model of the three, introduces the correlation model into a multi-objective algorithm NSPSO for optimization, obtains a final Pareto front edge of multi-objective optimization and a corresponding cutting parameter combination through iterative updating, thereby obtaining a cutting parameter combination with high precision, high quality and high efficiency, gives consideration to energy consumption, production quality and efficiency, not only effectively reduces energy consumption required by production, realizes energy conservation and environmental protection, and can meet the current trend of sustainable manufacturing, the cutting precision can be obviously improved, so that the product quality and the production efficiency are improved, and various cutting parameter combinations can be used for standardizing process design, so that operation guidance is provided for stainless steel turning, and the development of enterprises and cutting processes is facilitated.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (10)
1. A method for multi-objective optimization for stainless steel processing, comprising:
obtaining extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material;
creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth;
performing a cutting experiment according to the L9 orthogonal table;
collecting cutting force and surface roughness in the cutting experiment, and simultaneously calculating the material removal rate to obtain an experiment data set;
normalizing the experimental data set, inputting the result of the normalization into a support vector machine, and establishing a correlation model of cutting force, surface roughness and material removal rate;
and optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto frontier and a corresponding cutting parameter combination.
2. The multi-objective optimization method for stainless steel machining according to claim 1, wherein the obtaining extreme values of feed rate, cutting speed and cutting depth of the selected stainless steel material comprises:
and determining the maximum value and the minimum value of the feed rate, the cutting speed and the cutting depth when the tool machines the stainless steel material according to the rated parameters of the machining system and a tool manual.
3. The multi-objective optimization method for stainless steel machining according to claim 1, wherein the creating of the L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth specifically comprises:
taking the feeding rate, the cutting speed and the cutting depth as three factors, and numbering the three factors as (i) - (iii);
respectively selecting the minimum value, the middle value and the maximum value of the feed rate, the cutting speed and the cutting depth from the extreme value intervals of the feed rate, the cutting speed and the cutting depth, taking the minimum value, the middle value and the maximum value as three levels, and respectively numbering the three levels to be 1-3;
creating an L9 orthogonal table according to the three factors and the three levels; the L9 orthogonal table is a table with 3 columns and 9 rows, the number of columns represents factors, the number of rows represents experiment times, and the number of the experiment times is one to nine;
the number of 9 experiments was grouped by the number of levels of 3 so that each 3 experiments was 1 group, and the level number of each factor was filled in the L9 orthogonal table according to the filling method of the orthogonal table.
4. The multi-objective optimization method for stainless steel machining according to claim 1, wherein the cutting experiment according to the L9 orthogonal table specifically comprises:
carrying out nine groups of experiments according to the experimental information in the L9 orthogonal table, wherein each group of experiments is repeated three times;
and calculating the average value of the three experimental results, and taking the average value as the final experimental result.
5. The multi-objective optimization method for stainless steel machining according to claim 1, wherein the collecting of the cutting force and the surface roughness in the cutting experiment and the calculating of the material removal rate simultaneously obtain an experimental data set specifically comprises:
collecting cutting force of a cutter for processing the stainless steel material and surface roughness of the stainless steel material in the cutting experiment to obtain cutting force data and surface roughness data;
calculating the material removal rate of the stainless steel material after the machining is finished by using a formula MRR (maximum ratio of R) 1000Vdf to obtain material removal rate data; wherein MRR represents the material removal rate in mm3Min; v represents the cutting speed in m/min; d represents the depth of cut in mm; f represents the feed rate in mm/r;
obtaining an experimental data set according to the cutting force data, the surface roughness data and the material removal rate data;
the cutting force is collected through a Kistler dynamometer, the Kistler dynamometer obtains cutting force signals in three directions, the average value of a stable section is taken as a cutting force result in each direction, and the resultant force in the three directions is calculated as final cutting force data;
the surface roughness was measured by a Mahr roughness meter, which takes the mean of the plateaus as the final surface roughness data.
6. The multi-objective optimization method for stainless steel processing according to claim 1, wherein the normalizing the experimental data set and then inputting the normalized results into a support vector machine to respectively establish correlation models of cutting force, surface roughness and material removal rate comprises:
respectively carrying out normalization pretreatment on all data in the experimental data set, and normalizing all data to be in a [0, 1] interval;
selecting a radial basis function as a kernel function of a support vector machine model, and determining optimal parameters C and sigma of the support vector machine classification model through a cross verification method and a grid search method, wherein C represents a penalty coefficient, and sigma represents a width parameter;
respectively establishing a cutting force support vector machine model and a surface roughness support vector machine model by utilizing the optimal parameters C and sigma through a network open source code LIBSVM3.22 of the support vector machine;
obtaining a material removal rate formula model through the calculation formula of the material removal rate;
the correlation model includes the cutting force support vector machine model, the surface roughness support vector machine model, and the material removal rate formula model.
7. The multi-objective optimization method for stainless steel processing of claim 6, wherein the normalization of all data in the experimental data set is performed separately using the following normalization formula:
wherein N is*For the normalization result, N is the initial value, NmaxAnd NminThe maximum value and the minimum value of the initial value are respectively.
8. The multi-objective optimization method for stainless steel processing according to claim 6, wherein the association model is optimized by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto front and a corresponding cutting parameter combination, and the method specifically comprises the following steps:
introducing the cutting force support vector machine model, the surface roughness support vector machine model and the material removal rate formula model into an NSPSO intelligent optimization algorithm;
generating an initial population randomly within the speed and position range of the particles by utilizing the NSPSO intelligent optimization algorithm;
carrying out normalization pretreatment on the position of the initial population, and calculating the fitness of each particle in the initial population based on an objective function;
updating the current population according to a speed and position updating formula of an NSPSO intelligent optimization algorithm, and merging the child and the parent to perform non-dominated solution set sorting; the child represents the updated particle swarm, and the parent represents the particle swarm before updating;
and generating a new current generation population, updating the individual extreme value and the population extreme value, and repeating iterative updating to obtain a final Pareto front edge and an optimal population.
9. The multi-objective optimization method for stainless steel processing according to claim 8, wherein the current population is updated according to an update formula of speed and position of NSPSO intelligent optimization algorithm, the update formula is as follows:
wherein v isi，jAnd pi，jRespectively representing the speed and the position of the ith particle in the jth generation; v. ofi，j+1And pi，j+1Respectively representing the speed and the position of the ith particle in the j +1 th generation; c. C1And c2Represents the acceleration factor, r1And r2Is a random number between 0 and 1; pi，bestIs the historically optimal individual for the ith particle, GbestThe method comprises the following steps of (1) selecting a global optimal individual randomly in a first-layer non-dominated solution set;
ω represents an inertia factor, expressed as:
wherein j represents the current algebra, and N is the maximum iteration number;
the generating a new current generation population, updating individual extreme values and population extreme values, and repeating iterative updating to obtain a final Pareto front edge and an optimal population specifically comprises the following steps:
sequentially selecting N individuals ranked in the front to become new parents according to the ranking level of the non-dominated solution set ranking;
when the number of the selected non-dominated solution sets on a certain layer overflows, sorting the individuals on the current layer from large to small according to the congestion distance, sequentially taking out the individuals, putting the individuals into the next parent generation until the upper limit of the population is reached, and entering the next updating;
and repeating the updating steps until the maximum iteration times are reached to obtain a final Pareto front edge and an optimal population, wherein the final Pareto front edge and the optimal population are the optimal cutting parameter combination of the multi-objective optimization.
10. A multi-objective optimization system for stainless steel processing, comprising:
the parameter extreme value acquisition module is used for acquiring extreme values of the feed rate, the cutting speed and the cutting depth of the selected stainless steel material;
the orthogonal table creating module is used for creating an L9 orthogonal table according to the extreme values of the feed rate, the cutting speed and the cutting depth;
the cutting experiment implementation module is used for carrying out a cutting experiment according to the L9 orthogonal table;
the experimental data generation module is used for acquiring cutting force and surface roughness in the cutting experiment, and calculating the material removal rate to generate an experimental data set;
the correlation model creating module is used for carrying out normalization processing on the experimental data set, then inputting the normalization processing result into a support vector machine, and establishing a correlation model of cutting force, surface roughness and material removal rate;
and the optimization obtaining parameter module is used for optimizing the correlation model by using an NSPSO intelligent optimization algorithm to obtain a multi-objective optimized Pareto front edge and a corresponding cutting parameter combination.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010847400.1A CN111948977B (en) | 2020-08-21 | 2020-08-21 | Multi-objective optimization method and system for stainless steel processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010847400.1A CN111948977B (en) | 2020-08-21 | 2020-08-21 | Multi-objective optimization method and system for stainless steel processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111948977A true CN111948977A (en) | 2020-11-17 |
| CN111948977B CN111948977B (en) | 2021-08-10 |
Family
ID=73359279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010847400.1A Active CN111948977B (en) | 2020-08-21 | 2020-08-21 | Multi-objective optimization method and system for stainless steel processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111948977B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112925278A (en) * | 2021-01-29 | 2021-06-08 | 重庆大学 | Multi-target hobbing process parameter optimization and decision method |
| CN113052207A (en) * | 2021-03-10 | 2021-06-29 | 中国特种设备检测研究院 | Stainless steel environment cracking judgment method and system based on Hybrid idea |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106980877A (en) * | 2017-03-15 | 2017-07-25 | 陕西中爆安全网科技有限公司 | A kind of Prediction of Blasting Vibration method based on particle cluster algorithm Support Vector Machines Optimized |
| CN108153918A (en) * | 2018-02-05 | 2018-06-12 | 北京工业大学 | Method is recommended based on the personalized film of multi-objective optimization algorithm that adaptive quadrature intersects |
| CN109145477A (en) * | 2018-08-31 | 2019-01-04 | 西北工业大学 | A kind of cutting parameter optimization method based on SPH cutting Model |
| CN109839895A (en) * | 2019-01-24 | 2019-06-04 | 温州大学 | A kind of method that cutter geometrical structure parameter and working process parameter optimize jointly |
| CN110579971A (en) * | 2019-10-25 | 2019-12-17 | 福州大学 | multi-objective cutting parameter optimization method for green manufacturing |
-
2020
- 2020-08-21 CN CN202010847400.1A patent/CN111948977B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106980877A (en) * | 2017-03-15 | 2017-07-25 | 陕西中爆安全网科技有限公司 | A kind of Prediction of Blasting Vibration method based on particle cluster algorithm Support Vector Machines Optimized |
| CN108153918A (en) * | 2018-02-05 | 2018-06-12 | 北京工业大学 | Method is recommended based on the personalized film of multi-objective optimization algorithm that adaptive quadrature intersects |
| CN109145477A (en) * | 2018-08-31 | 2019-01-04 | 西北工业大学 | A kind of cutting parameter optimization method based on SPH cutting Model |
| CN109839895A (en) * | 2019-01-24 | 2019-06-04 | 温州大学 | A kind of method that cutter geometrical structure parameter and working process parameter optimize jointly |
| CN110579971A (en) * | 2019-10-25 | 2019-12-17 | 福州大学 | multi-objective cutting parameter optimization method for green manufacturing |
Non-Patent Citations (1)
| Title |
|---|
| 李鹏宇: "面向能效的数控铣削加工工艺参数优化模型及方法研究", 《中国优秀硕士学位论文全文数据库工程科技I辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112925278A (en) * | 2021-01-29 | 2021-06-08 | 重庆大学 | Multi-target hobbing process parameter optimization and decision method |
| CN112925278B (en) * | 2021-01-29 | 2023-09-15 | 重庆大学 | Multi-target gear hobbing process parameter optimization and decision method |
| CN113052207A (en) * | 2021-03-10 | 2021-06-29 | 中国特种设备检测研究院 | Stainless steel environment cracking judgment method and system based on Hybrid idea |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111948977B (en) | 2021-08-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107193258B (en) | Numerical control processing technology route and cutting parameter integrated optimization method towards energy consumption | |
| CN110052713B (en) | Composite manufacturing process for increasing and decreasing materials of parts | |
| CN103699055B (en) | Aircraft structure intelligent numerical control machining prgraming system and method | |
| Bouacha et al. | Hard turning behavior improvement using NSGA-II and PSO-NN hybrid model | |
| CN103390082B (en) | The sane excellent method of completing the square of a kind of gang tool geometric accuracy | |
| CN104267693A (en) | Method for optimizing cutting parameters considering machining energy efficiency | |
| Kotliar et al. | Ensuring the economic efficiency of enterprises by multi-criteria selection of the optimal manufacturingprocess | |
| CN111948977B (en) | Multi-objective optimization method and system for stainless steel processing | |
| Pawanr et al. | Multi-objective optimization of machining parameters to minimize surface roughness and power consumption using TOPSIS | |
| CN105446264B (en) | The machine tool accuracy optimum design method of feature based | |
| CN106843153B (en) | The reusable NC technology mapping method of process oriented scheme | |
| CN106228298A (en) | Processing scheme method for optimizing under cloud manufacturing environment | |
| CN108920806B (en) | Heavy machine tool reliability distribution method based on trapezoidal fuzzy number and sorting method | |
| CN104318022A (en) | Method for predicting workpiece surface roughness and increasing cutting efficiency | |
| Krimpenis et al. | CNC micromilling properties and optimization using genetic algorithms | |
| CN110210079A (en) | A kind of lathe supporting member quality matches method towards complete machine dynamic characteristic | |
| CN112861433A (en) | Product low-carbon design method based on multi-level integrated framework | |
| Krimpenis et al. | Rough milling optimisation for parts with sculptured surfaces using genetic algorithms in a Stackelberg game | |
| CN116466651A (en) | Numerical control machining process parameter optimization method and system based on hybrid heuristic algorithm | |
| CN113868860A (en) | Part self-adaptive cost estimation method based on process knowledge | |
| Wang et al. | Optimization of the tool selection based on big data | |
| CN111666720A (en) | Large ship engine cylinder body milling method based on improved genetic algorithm | |
| Forse | Statistical tolerancing for optics | |
| CN117032072B (en) | Geometric error distribution method for five-axis ultra-precision machine tool | |
| CN110083126A (en) | A kind of die industry machining period assessment system, computation model and calculation method |
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 |