CN115329475A - Part preparation method and equipment based on partitioned multistage cryogenic treatment - Google Patents

Part preparation method and equipment based on partitioned multistage cryogenic treatment Download PDF

Info

Publication number
CN115329475A
CN115329475A CN202210837481.6A CN202210837481A CN115329475A CN 115329475 A CN115329475 A CN 115329475A CN 202210837481 A CN202210837481 A CN 202210837481A CN 115329475 A CN115329475 A CN 115329475A
Authority
CN
China
Prior art keywords
cryogenic treatment
cryogenic
microstructure
stage
neural network
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
Application number
CN202210837481.6A
Other languages
Chinese (zh)
Other versions
CN115329475B (en
Inventor
王新云
丁华平
唐学峰
牛勇
邓磊
龚攀
张茂
金俊松
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.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
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 Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202210837481.6A priority Critical patent/CN115329475B/en
Publication of CN115329475A publication Critical patent/CN115329475A/en
Application granted granted Critical
Publication of CN115329475B publication Critical patent/CN115329475B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/10Geometric CAD
    • G06F30/17Mechanical parametric or variational design
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/20Design optimisation, verification or simulation
    • G06F30/27Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • G06N3/084Backpropagation, e.g. using gradient descent
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • G06N3/086Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/12Computing arrangements based on biological models using genetic models
    • G06N3/126Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00Details relating to CAD techniques
    • G06F2111/10Numerical modelling

Landscapes

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

Abstract

本发明属于深冷处理相关技术领域,其公开了一种基于分区多级深冷处理的零件制备方法及设备,包括以下步骤:(1)通过神经网络建立多级深冷工艺参数‑微观组织‑性能之间的非线性映射关系;(2)将目标零件划分为多个子区域,反推出各个子区域所需的多级深冷处理参数,进而获得目标零件内部对应的一系列深冷工艺参数分布组合;(3)对分区多级深冷处理过程进行高通量数值模拟,以确定目标零件内部不同区域的最佳深冷处理参数组合;(4)对待处理零件进行分区独立式的深冷处理,以得到所述目标零件。本发明解决了现有技术中对零件深冷处理的适用性不强、工艺复杂的技术问题。

Figure 202210837481

The invention belongs to the technical field of cryogenic treatment, and discloses a method and equipment for preparing parts based on sub-regional multi-stage cryogenic treatment, comprising the following steps: (1) establishing a multi-stage cryogenic process parameter-microstructure-performance relationship through a neural network (2) Divide the target part into multiple sub-regions, invert the multi-level cryogenic treatment parameters required by each sub-region, and then obtain a series of corresponding cryogenic process parameter distribution combinations inside the target part; ( 3) Carry out high-throughput numerical simulation of the partitioned multi-stage cryogenic treatment process to determine the optimal combination of cryogenic treatment parameters for different areas inside the target part; (4) Perform partition-independent cryogenic treatment on the part to be treated to obtain the target Components. The invention solves the technical problems in the prior art that the applicability to the cryogenic treatment of the parts is not strong and the process is complicated.

Figure 202210837481

Description

基于分区多级深冷处理的零件制备方法及设备Part preparation method and equipment based on partitioned multi-stage cryogenic treatment

技术领域technical field

本发明属于深冷处理相关技术领域,更具体地,涉及一种基于分区多级深冷处理的零件制备方法及设备。The invention belongs to the technical field related to cryogenic treatment, and more specifically relates to a part preparation method and equipment based on multi-stage cryogenic treatment in different regions.

背景技术Background technique

轻量化结构零件,如整体叶盘、涡轮盘等能够实现减重约30%,提升发动机的推重比和使用可靠性。整体叶盘、涡轮盘零件沿径向具有较大的温度梯度和应力梯度,不同区域对材料性能的要求有差异,叶片强调具有优良的耐高周疲劳性能,盘体强调具有高的高温蠕变抗力和损伤容限性能。涡轮盘的盘辐工作温度相对较低,细晶组织更符合盘辐高的屈服强度和低周疲劳性能要求;盘缘温度相对较高,粗晶组织具有高的蠕变和损伤容限性能,能够避免榫槽可能出现的微裂纹。为了进一步发挥整体叶盘等零件的性能潜力,根据零件不同区域实际服役环境,选用恰当的合金材料及组织状态,为此科学家们提出了双性能整体叶盘、涡轮盘的设计思路,突破传统热加工技术追求均一组织的惯性思维,发展了一系列具有梯度微观组织特征的单合金双性能或者双合金的梯度结构零件。Lightweight structural parts, such as blisks and turbine disks, can reduce weight by about 30%, improving the engine's thrust-to-weight ratio and reliability. The overall blisk and turbine disk parts have large temperature gradients and stress gradients in the radial direction, and different regions have different requirements for material properties. The blades emphasize excellent high-cycle fatigue resistance, and the disk body emphasizes high-temperature creep Resistance and damage tolerance performance. The working temperature of the spokes of the turbine disk is relatively low, and the fine-grained structure is more in line with the high yield strength and low cycle fatigue performance requirements of the spokes; the edge temperature is relatively high, and the coarse-grained structure has high creep and damage tolerance performance. Possible microcracks in the tongue and groove can be avoided. In order to further develop the performance potential of parts such as blisks, appropriate alloy materials and microstructures should be selected according to the actual service environment in different regions of the parts. For this reason, scientists have proposed the design idea of dual-performance blisks and turbine disks, breaking through the traditional thermal Processing technology pursues the inertial thinking of uniform organization, and has developed a series of single-alloy dual-performance or double-alloy gradient structure parts with gradient microstructure characteristics.

目前梯度结构金属零件,特别是具有代表性的双性能整体叶盘、涡轮盘等零件的制造方法主要有焊接法、分区控温锻造和分区控温热处理。焊接法可实现异种材料的连接,但是其最大的问题是连接区域往往会成为整个构件的薄弱环节,这对于强调高可靠性和长寿命的航空发动机高速转动部件来说是个重要隐患。较之于分区控温锻造工艺,采用分区控温热处理工艺,在叶片与盘体之间更容易形成稳定可控的温度梯度,从而得到所需要的双重组织,而且工艺过程操作相对容易,一致性好。经传统热处理后,材料仍会存在某些不足,如淬火后组织不稳定、较高的热应力与组织应力、组织不均匀等,均会恶化材料性能进而影响材料的使用寿命。一般来说,通过单一热处理工艺难以解决此类问题,而深冷处理作为热处理的一种重要附加工艺,能够有效地对材料热处理后的性能进行二次优化,对延长材料的服役寿命具有显著效果。At present, the manufacturing methods of gradient structure metal parts, especially the representative dual-performance integral blisks and turbine disks, mainly include welding method, zoned temperature controlled forging and zoned temperature controlled heat treatment. The welding method can realize the connection of dissimilar materials, but its biggest problem is that the connection area often becomes the weak link of the whole component, which is an important hidden danger for high-speed rotating parts of aero-engines that emphasize high reliability and long life. Compared with the zoned temperature control forging process, the zoned temperature controlled heat treatment process is easier to form a stable and controllable temperature gradient between the blade and the disc, so as to obtain the required double structure, and the process operation is relatively easy and consistent. it is good. After traditional heat treatment, the material still has some shortcomings, such as unstable structure after quenching, high thermal stress and structural stress, uneven structure, etc., which will deteriorate the material performance and affect the service life of the material. Generally speaking, it is difficult to solve such problems through a single heat treatment process. As an important additional process of heat treatment, cryogenic treatment can effectively re-optimize the performance of materials after heat treatment, and has a significant effect on prolonging the service life of materials.

深冷处理是指把材料(或工件)置于一定的低温(-130℃以下)环境下,通过控制深冷处理温度、升降温速率以及保温时间、处理次数等工艺参数对材料进行处理,使材料的微观组织发生不同程度的不可逆转变,从而达到改善综合性能的目的。研究表明,深冷处理对模具钢、高温合金、钛合金、硬质合金、非晶合金和高熵合金等材料均有显著的改性效果,具有促进残余奥氏体转化、细化晶粒、提高位错密度、促进孪晶和亚晶组织生成、促进相转变和碳化物析出,并促进织构生成等作用。Cryogenic treatment refers to placing materials (or workpieces) in a certain low temperature (below -130°C) environment, and processing materials by controlling process parameters such as cryogenic treatment temperature, heating and cooling rate, holding time, and treatment times, so that the material's Different degrees of irreversible changes occur in the microstructure, so as to achieve the purpose of improving the comprehensive performance. Studies have shown that cryogenic treatment has a significant modification effect on materials such as die steel, superalloys, titanium alloys, hard alloys, amorphous alloys and high-entropy alloys, and can promote the transformation of retained austenite, refine grains, and improve Dislocation density, promote the formation of twins and subgrain structures, promote phase transformation and carbide precipitation, and promote texture formation.

对于梯度结构零件,不同部位的组织要求、甚至合金类型不同,对深冷处理的需求存在较大的差异。对于这类零件的一般深冷处理方式是采用隔热材料包裹住某一阶段不进行深冷处理的区域,往往需要多次工序、工序复杂且效果不好。现有深冷处理工艺和装置均只适用于单一材料、均匀零件,难以满足双性能叶盘、涡轮盘等具有梯度结构零件不同区域的深冷需求。For parts with gradient structure, different parts have different organizational requirements and even different alloy types, so the requirements for cryogenic treatment are quite different. The general cryogenic treatment method for such parts is to use heat insulating material to wrap the area that does not undergo cryogenic treatment at a certain stage, which often requires multiple processes, the process is complicated and the effect is not good. Existing cryogenic treatment processes and devices are only suitable for single-material, uniform parts, and it is difficult to meet the cryogenic requirements of different regions of parts with gradient structures such as dual-performance blisks and turbine disks.

发明内容Contents of the invention

针对现有技术的以上缺陷或改进需求,本发明提供了一种基于分区多级深冷处理的零件制备方法及设备,该制备方法通过控制多级深冷处理过程中样品内部的温度梯度场分布来获得具有相应组织特征、性能分布的零件,其能够通过多级深冷处理过程中的温度、保温时间等工艺参数优化来实现对零件不同区域的差异化深冷,从而获得目标组织和性能分布,以解决现有技术中对零件深冷处理的适用性不强、工艺复杂的技术问题。In view of the above defects or improvement needs of the prior art, the present invention provides a part preparation method and equipment based on partitioned multi-stage cryogenic treatment. The preparation method obtains by controlling the temperature gradient field distribution inside the sample during the multi-stage cryogenic treatment process. Parts with corresponding tissue characteristics and performance distribution can achieve differentiated cryogenic treatment of different regions of the part through the optimization of process parameters such as temperature and holding time in the multi-stage cryogenic treatment process, so as to obtain the target tissue and performance distribution to solve the problem. In the prior art, the applicability to cryogenic treatment of parts is not strong and the technical problems are complicated.

为实现上述目的,按照本发明的一个方面,提供了一种基于分区多级深冷处理的零件制备方法,该制备方法主要包括以下步骤:In order to achieve the above object, according to one aspect of the present invention, a method for preparing parts based on subregional multi-stage cryogenic treatment is provided, the preparation method mainly includes the following steps:

(1)通过多级深冷处理实验获得目标零件的材料成分的组织-性能演化规律,并通过神经网络建立多级深冷工艺参数-微观组织-性能之间的非线性映射关系;(1) Obtain the microstructure-property evolution law of the material composition of the target part through multi-stage cryogenic treatment experiments, and establish the non-linear mapping relationship between multi-stage cryogenic process parameters-microstructure-property through neural network;

(2)根据目标零件的组织性能分布要求将目标零件划分为多个子区域,结合得到的非线性映射关系反推出各个子区域所需的多级深冷处理参数,进而获得采用多级深冷处理目标零件时目标零件内部对应的一系列深冷工艺参数分布组合;(2) Divide the target part into multiple sub-regions according to the distribution requirements of the target part's organization and performance, and deduce the multi-level cryogenic treatment parameters required for each sub-region by combining the obtained nonlinear mapping relationship, and then obtain the multi-level cryogenic treatment target part A series of cryogenic process parameter distribution combinations corresponding to the inside of the target part;

(3)结合得到的深冷工艺参数分布组合对分区多级深冷处理过程进行高通量数值模拟,将数值模拟时温度分布或者微观组织分布最接近目标零件的组织性能分布要求所对应的深冷处理工艺参数组合确定为目标零件内部不同区域的最佳深冷处理参数组合;(3) Combining the obtained cryogenic process parameter distribution combination to carry out high-throughput numerical simulation on the partitioned multi-stage cryogenic treatment process, the temperature distribution or microstructure distribution in the numerical simulation is closest to the cryogenic treatment corresponding to the microstructure and performance distribution requirements of the target part The combination of process parameters is determined as the best combination of cryogenic treatment parameters for different regions inside the target part;

(4)结合目标零件的几何划分区域和最佳深冷处理参数组合对待处理零件进行分区独立式的深冷处理,以得到所述目标零件。(4) Combining the geometrically divided regions of the target part and the optimal cryogenic treatment parameter combination to perform partitioned and independent cryogenic treatment on the part to be treated to obtain the target part.

进一步地,目标零件的材料成分的组织-性能演化规律的获取包括以下步骤:以目标零件的材料为初始实验对象,在预先设置的深冷处理参数范围内,设置正交实验或者单因素试验,在不同深冷处理工艺参数条件下对所述初始试验对象进行多级深冷处理,并对深冷处理得到的样品进行微观组织表征和性能测试,获得不同深冷处理参数条件下样品中微观组织特征参数和性能数据,进而获得多级深冷处理工艺参数-微观组织-性能数据库。Furthermore, the acquisition of the structure-property evolution law of the material composition of the target part includes the following steps: taking the material of the target part as the initial experimental object, within the preset cryogenic treatment parameter range, setting up an orthogonal experiment or a single factor test, in Perform multi-stage cryogenic treatment on the initial test object under different cryogenic treatment process parameters, and perform microstructure characterization and performance testing on the samples obtained from cryogenic treatment, and obtain microstructural characteristic parameters and performance data of samples under different cryogenic treatment parameters , and then obtain the multi-stage cryogenic treatment process parameter-microstructure-property database.

进一步地,多级深冷处理是指从起始处理温度到最低处理温度之间经历若干次保温、升降温过程的工艺;微观组织特征参数包括晶粒尺寸、位错密度、残余奥氏体体积分数、碳化物相的尺寸和体积分数中的一种或者多种;所述多级深冷处理工艺参数包括温度、保温时间、升降温速率、处理次数中的一种或者多种。Furthermore, multi-stage cryogenic treatment refers to a process that undergoes several heat preservation and heating and cooling processes from the initial treatment temperature to the lowest treatment temperature; the microstructure characteristic parameters include grain size, dislocation density, retained austenite volume fraction , one or more of the size and volume fraction of the carbide phase; the multi-stage cryogenic treatment process parameters include one or more of temperature, holding time, heating and cooling rate, and treatment times.

进一步地,所建立的多级深冷工艺参数-微观组织-性能之间的非线性映射关系包含深冷处理工艺参数-微观组织关系模型、微观组织-性能关系模型、以及深冷处理工艺参数-性能关系模型。Further, the established nonlinear mapping relationship between multi-stage cryogenic process parameters-microstructure-performance includes cryogenic treatment process parameters-microstructure relationship model, microstructure-performance relationship model, and cryogenic treatment process parameter-performance relationship Model.

进一步地,每种关系模型均采用含有多个隐含层的BP神经网络模型来描述,其中深冷工艺参数-微观组织非线性映射关系的建立包含如下步骤:Further, each relationship model is described by a BP neural network model containing multiple hidden layers, and the establishment of the cryogenic process parameter-microstructure nonlinear mapping relationship includes the following steps:

以多级深冷处理实验获得的深冷工艺参数的集合为输入,以获得的微观组织特征参数集合为输出,构建并训练一个含有多隐含层的BP神经网络模型,给各隐含层和输出层选择合适的激励函数;利用遗传算法优化所述BP神经网络初始的权值和阈值,得到最优个体的权值和阈值;将得到的最优个体的权值和阈值赋值给所述BP神经网络模型,训练过程中使用误差逆传播算法对各隐含层的权值和阈值进行更新,直到代价函数J小于设定精度或者达到最大迭代次数则训练结束。Taking the set of cryogenic process parameters obtained from multi-level cryogenic treatment experiments as input, and the obtained set of microstructure characteristic parameters as output, a BP neural network model with multiple hidden layers is constructed and trained, and each hidden layer and output layer selects an appropriate activation function; optimizes the initial weights and thresholds of the BP neural network using a genetic algorithm to obtain the weights and thresholds of the optimal individual; assigns the weights and thresholds of the obtained optimal individual to the BP neural network In the network model, the error backpropagation algorithm is used to update the weights and thresholds of each hidden layer during the training process, until the cost function J is less than the set accuracy or reaches the maximum number of iterations, then the training ends.

进一步到底,隐含层的激励函数均选择logistic函数,输出层的激励函数选择线性函数g(x)=x。Further to the end, the activation function of the hidden layer is selected as a logistic function, and the activation function of the output layer is selected as a linear function g(x)=x.

进一步地,采用遗传算法优化所述BP神经网络初始的权值和阈值,具体包括如下步骤:Further, using genetic algorithm to optimize the initial weights and thresholds of the BP neural network, specifically includes the following steps:

S1:首先根据所述BP神经网络模型的拓扑图确定神经网络的权值和阈值的个数,遵循如下公式:S1: First, determine the number of weights and thresholds of the neural network according to the topological diagram of the BP neural network model, following the following formula:

Figure BDA0003749236780000041
Figure BDA0003749236780000041

其中Num为权值和阈值的总个数,i表示第i层神经元,Hi为第i层神经元的节点数;Where N um is the total number of weights and thresholds, i represents the i-th layer of neurons, H i is the number of nodes of the i-th layer of neurons;

S2:采用实数编码方式对神经网络的权值阈值进行编码操作,初始化种群,初始的权值阈值在(-1,1)间随机取值,设置种群的适应度函数为F1;S2: Use the real number encoding method to encode the weight threshold of the neural network, initialize the population, the initial weight threshold is randomly selected between (-1,1), and set the fitness function of the population to F1;

Figure BDA0003749236780000042
Figure BDA0003749236780000042

其中F1为适应度值,ρ1为使用初始的权值和阈值的神经网络预测的增强相成分,η1为使用初始的权值和阈值的神经网络预测的增强相体积分数,δ1为使用初始的权值和阈值的神经网络预测的增强相平均尺寸,ρ0为增强相成分期望值,η0为增强相体积分数期望值,δ0为增强相平均尺寸期望值;Among them, F1 is the fitness value, ρ1 is the enhanced phase component predicted by the neural network using the initial weight and threshold, η1 is the enhanced phase volume fraction predicted by the neural network using the initial weight and threshold, and δ1 is the initial weight The average size of the enhanced phase predicted by the neural network of value and threshold, ρ0 is the expected value of the enhanced phase composition, η0 is the expected value of the volume fraction of the enhanced phase, and δ0 is the expected value of the average size of the enhanced phase;

S3:计算种群中所有个体的适应度值,并使用轮盘赌算法进行选择操作,从父代中挑选适应度高的个体产生下一代个体,每个个体被选中的概率遵循下公式:S3: Calculate the fitness value of all individuals in the population, and use the roulette algorithm to select individuals with high fitness from the parent generation to generate the next generation of individuals. The probability of each individual being selected follows the following formula:

Figure BDA0003749236780000051
Figure BDA0003749236780000051

其中,pk为第k个个体被选中的概率,Fk为第k个个体的适应度值,K为种群中个体的总数;Among them, p k is the probability that the kth individual is selected, F k is the fitness value of the kth individual, and K is the total number of individuals in the population;

S4:对种群中的个体进行交叉操作,设定交叉概率为pc,产生一个随机数若小于交叉概率,则进行交叉操作,交叉时随机选择两个个体并随机选择交叉位,遵循以下公式进行交叉操作:S4: Perform a crossover operation on the individuals in the population, set the crossover probability as pc, generate a random number that is less than the crossover probability, then perform the crossover operation, randomly select two individuals and randomly select the crossover position during the crossover, and perform crossover according to the following formula operate:

Figure BDA0003749236780000052
Figure BDA0003749236780000052

其中,akj是第k个染色体在j位上的实数,alj是第l个染色体在j位上的实数,b为(0,1)间的随机数;Among them, a kj is the real number of the k-th chromosome at the j-position, a lj is the real number of the l-th chromosome at the j-position, and b is a random number between (0,1);

S5:对种群中的个体进行变异操作,设定变异概率为pm,产生一个随机数若小于变异概率,则进行变异操作,变异时随机选择一个个体并随机选择变异位,遵循以下公式进行变异操作:S5: Perform a mutation operation on the individuals in the population, set the mutation probability to pm, generate a random number that is less than the mutation probability, then perform the mutation operation, randomly select an individual and randomly select the mutation bit during mutation, and perform the mutation operation according to the following formula :

Figure BDA0003749236780000053
Figure BDA0003749236780000053

Figure BDA0003749236780000054
Figure BDA0003749236780000054

其中,aij是第i个染色体在j位上的实数,g为当前迭代次数,f(g)是变异因子,Gmax为最大迭代次数,amax是aij取值的上限,amin是aij取值的下限,r和r'为(0,1)间的随机数;Among them, a ij is the real number of the i-th chromosome at position j, g is the current iteration number, f(g) is the variation factor, G max is the maximum iteration number, a max is the upper limit of the value of a ij , and a min is The lower limit of the value of a ij , r and r' are random numbers between (0,1);

S6:循环步骤S3至S5直至得到满意的适应度值或达到限定的迭代次数,输出最优的个体即适应度值最大的个体。S6: Repeat steps S3 to S5 until a satisfactory fitness value is obtained or a limited number of iterations is reached, and the optimal individual is output, that is, the individual with the largest fitness value.

按照本发明的另一个方面,提供了一种基于分区多级深冷处理的零件制备设备,该制备设备采用如上所述的基于分区多级深冷处理的零件制备方法来制备零件的;所述设备形成有多个独立的深冷处理腔室,每个深冷处理腔室能够独立对待制备零件的一段区域进行深冷处理。According to another aspect of the present invention, there is provided a kind of parts preparation equipment based on subregional multi-stage cryogenic treatment, the preparation equipment adopts the above-mentioned parts preparation method based on subregional multi-stage cryogenic treatment to prepare parts; the equipment forms There are multiple independent cryogenic treatment chambers, and each cryogenic treatment chamber can independently perform cryogenic treatment on a section of the part to be prepared.

进一步地,每个深冷处理腔室分别设置有独立的电阻加热丝、温度传感器、冷却介质的进出管及移动机构,使得每个深冷处理腔室均能够单独工作以对目标零件的局部区域进行深冷处理。Further, each cryogenic treatment chamber is provided with an independent resistance heating wire, a temperature sensor, a cooling medium inlet and outlet pipe, and a moving mechanism, so that each cryogenic treatment chamber can work independently to deep-dry a local area of the target part. Cold treatment.

进一步地,所述设备还包括多个隔热板,且所述设备形成有收容腔,多个所述隔热板间隔设置在所述收容腔内以将所述收容腔分割成多个独立的所述深冷处理腔室;所述隔热板还用于承载待制备的零件;每个隔热板连接一个所述移动机构,所述移动机构用于带动所述隔热板进行移动,以改变隔热板两侧的深冷处理腔室的空间大小。Further, the device also includes a plurality of heat insulation boards, and the device is formed with a storage chamber, and a plurality of the heat insulation boards are arranged at intervals in the storage chamber to divide the storage chamber into a plurality of independent The cryogenic treatment chamber; the heat shield is also used to carry the parts to be prepared; each heat shield is connected to one of the moving mechanisms, and the movement mechanism is used to drive the heat shield to move to change The space size of the cryogenic processing chamber on both sides of the heat shield.

总体而言,通过本发明所构思的以上技术方案与现有技术相比,本发明提供的基于分区多级深冷处理的零件制备方法及设备主要具有以下有益效果:Generally speaking, compared with the prior art through the above technical solutions conceived by the present invention, the part preparation method and equipment based on subregional multi-stage cryogenic treatment provided by the present invention mainly have the following beneficial effects:

1.所述制备方法利用金属材料在深冷处理时的微观组织、性能演化直接取决于深冷处理中温度场分布这一事实依据,根据目标零件中微观组织和性能要求,反推出利用多级深冷处理目标零件时所需的梯度温度场分布,并通过数值模拟确定出能够获得该目标梯度温度场分布的几何划分方案,最终采用反推确的多级深冷处理参数和几何划分方案,调整深冷处理装置各腔室空间,设置各腔室的深冷处理程序,获得目标零件。1. The preparation method uses the fact that the microstructure and performance evolution of metal materials during cryogenic treatment directly depends on the temperature field distribution in cryogenic treatment, and according to the microstructure and performance requirements of the target part, it is inversely deduced that the use of multi-stage cryogenic treatment The gradient temperature field distribution required for the target part, and the geometric division scheme that can obtain the target gradient temperature field distribution is determined through numerical simulation, and finally the back-calculated multi-stage cryogenic treatment parameters and geometric division scheme are used to adjust the cryogenic treatment device For each chamber space, set the cryogenic treatment program for each chamber to obtain the target parts.

2.本发明采用多级降温方式,相比于传统直接浸入液氮等冷却介质等方式,对零件产生的热冲击小,处理过程不容易产生裂纹和变形,适用于薄壁零件和脆性材料;同时针对零件的不同区域的深冷处理需求,采用分区深冷处理方式,一次深冷处理即可完成整个零件的微观组织和力学性能调控,显著简化了该类零件的深冷处理工序,提高了处理效率,且不同区域之间的微观组织、力学性能过渡良好,能够避免零件服务过程中的界面应力集中,提高了整体服务寿命。2. The present invention adopts a multi-stage cooling method. Compared with the traditional method of directly immersing in liquid nitrogen and other cooling media, the thermal shock to the parts is small, and the processing process is not easy to produce cracks and deformations, and is suitable for thin-walled parts and brittle materials; At the same time, according to the cryogenic treatment requirements of different regions of the parts, the sub-regional cryogenic treatment method is adopted, and the microstructure and mechanical properties of the entire part can be adjusted in one cryogenic treatment, which significantly simplifies the cryogenic treatment process of this type of parts and improves the processing efficiency. The transition of microstructure and mechanical properties between different regions is good, which can avoid the interface stress concentration during the part service process and improve the overall service life.

3.本发明首先对目标零件的材料成分进行多级深冷处理的正交实验或者单因素实验,获得不同深冷处理参数下其微观组织、性能数据,然后借助于这些实验数据利用神经网络技术建立深冷处理参数比如温度、保温时间、升降温速率、处理次数与微观组织参数如晶粒尺寸、位错密度、残余奥氏体体积分数、碳化物等析出相体积分数等,性能参数如屈服强度、室温塑性、断裂强度、抗弯强度、硬度等的复杂非线性映射关系,能够全面反映深冷处理参数,微观组织-性能之间的复杂影响规律,帮助反推所需的梯度温度场、优化工艺参数,减少预实验的次数,实现零件分区多级深冷处理工艺的快速、智能制定。3. The present invention first carries out multi-stage cryogenic treatment orthogonal experiments or single factor experiments on the material composition of the target part, obtains its microstructure and performance data under different cryogenic treatment parameters, and then uses neural network technology to establish deep Cold treatment parameters such as temperature, holding time, heating and cooling rate, treatment times and microstructure parameters such as grain size, dislocation density, retained austenite volume fraction, carbide and other precipitated phase volume fraction, etc., performance parameters such as yield strength, room temperature The complex nonlinear mapping relationship of plasticity, fracture strength, flexural strength, hardness, etc., can fully reflect the cryogenic treatment parameters, the complex influence law between microstructure and performance, and help reverse the required gradient temperature field and optimize process parameters. Reduce the number of pre-tests, and realize the fast and intelligent formulation of multi-stage cryogenic treatment processes for parts partitions.

4.该设备将传统深冷装置的单一腔室改为多腔室设计,各腔室独立控温、空间灵活可调且各腔室之间设置隔热层,能够满足单一合金梯度组织或者多合金组织组成的复杂形状零件的深冷需求,显著简化了深冷处理工序,深冷处理效果较好。4. The equipment changes the single chamber of the traditional cryogenic device into a multi-chamber design. The temperature of each chamber is controlled independently, the space is flexible and adjustable, and the heat insulation layer is set between each chamber, which can meet the requirements of single alloy gradient structure or multi-chamber structure. The cryogenic requirement of complex shape parts composed of alloy structure significantly simplifies the cryogenic treatment process, and the cryogenic treatment effect is better.

附图说明Description of drawings

图1是本发明提供的一种基于分区多级深冷处理的零件制备方法的流程示意图;Fig. 1 is a schematic flow chart of a part preparation method based on partitioned multi-stage cryogenic treatment provided by the present invention;

图2是本发明一个实施例提供的基于分区多级深冷处理的零件制备设备的结构示意图;Fig. 2 is a schematic structural view of parts preparation equipment based on subregional multi-stage cryogenic treatment provided by an embodiment of the present invention;

图3是图2中的基于分区多级深冷处理的零件制备设备的剖视图;Fig. 3 is a cross-sectional view of the parts preparation equipment based on subregional multi-stage cryogenic treatment in Fig. 2;

图4是图2中的基于分区多级深冷处理的零件制备设备的局部示意图;Fig. 4 is a partial schematic diagram of the parts preparation equipment based on subregional multi-stage cryogenic treatment in Fig. 2;

图5是图2中的基于分区多级深冷处理的零件制备设备处于开合状态时的示意图;Fig. 5 is a schematic diagram of the parts preparation equipment based on subregional multi-stage cryogenic treatment in Fig. 2 when it is in an open and closed state;

图6是本发明另一个实施例提供的基于分区多级深冷处理的零件制备设备的结构示意图;Fig. 6 is a schematic structural diagram of parts preparation equipment based on zoned multi-stage cryogenic treatment provided by another embodiment of the present invention;

图7是图6中的基于分区多级深冷处理的零件制备设备的剖视图;Fig. 7 is a cross-sectional view of the parts preparation equipment based on subregional multi-stage cryogenic treatment in Fig. 6;

图8是图6中的基于分区多级深冷处理的零件制备设备的局部示意图;Fig. 8 is a partial schematic diagram of the parts preparation equipment based on zoned multi-stage cryogenic treatment in Fig. 6;

图9是图6中的基于分区多级深冷处理的零件制备设备处于开合状态时的示意图。Fig. 9 is a schematic diagram of the parts preparation equipment based on multi-stage cryogenic treatment in Fig. 6 when it is in an open and closed state.

在所有附图中,相同的附图标记用来表示相同的元件或结构,其中:1-上管线室,2-上腔室,3-下腔室,4-下管线室,5-连接机构,6-液氮管,7-移动机构,8-隔热层,9-电阻加热丝,10-热电偶,11-板状梯度结构零件,12-热电偶信号线,13-温度控制系统,14-隔热板,15-盘形梯度结构零件。In all the drawings, the same reference numerals are used to denote the same elements or structures, wherein: 1-upper pipeline chamber, 2-upper chamber, 3-lower chamber, 4-lower pipeline chamber, 5-connection mechanism , 6-liquid nitrogen tube, 7-moving mechanism, 8-heat insulation layer, 9-resistance heating wire, 10-thermocouple, 11-plate gradient structure parts, 12-thermocouple signal line, 13-temperature control system, 14-insulation plate, 15-disc-shaped gradient structural parts.

具体实施方式Detailed ways

为了使本发明的目的、技术方案及优点更加清楚明白,以下结合附图及实施例,对本发明进行进一步详细说明。应当理解,此处所描述的具体实施例仅仅用以解释本发明,并不用于限定本发明。此外,下面所描述的本发明各个实施方式中所涉及到的技术特征只要彼此之间未构成冲突就可以相互组合。In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not constitute a conflict with each other.

请参阅图1,本发明提供了一种基于分区多级深冷处理的零件制备方法,其能够实现具有零件的微观组织灵活设置、性能分区快速优化,不同区域之间组织性能连续性更好,多级处理产生的热冲击小,处理过程中不易发生零件开裂,适用的零件类型和材料范围广。其中,该制备方法可以制备梯度结构及均匀结构零件。Please refer to Fig. 1, the present invention provides a part preparation method based on multi-stage cryogenic treatment in different regions, which can realize the flexible setting of the microstructure of parts, rapid optimization of performance partitions, better continuity of tissue performance between different regions, and multiple The thermal shock generated by the grade treatment is small, and the parts are not easy to crack during the treatment process, and the applicable parts types and materials are wide. Among them, the preparation method can prepare gradient structure and uniform structure parts.

本发明提供的基于分区多级深冷处理的零件制备方法主要包括以下步骤:The method for preparing parts based on partitioned multi-stage cryogenic treatment provided by the present invention mainly includes the following steps:

步骤一,通过多级深冷处理实验获得目标零件的材料成分的组织-性能演化规律,并通过神经网络建立多级深冷工艺参数-微观组织-性能之间的非线性映射关系。Step 1: Obtain the microstructure-property evolution law of the material composition of the target part through multi-stage cryogenic treatment experiments, and establish the nonlinear mapping relationship between multi-stage cryogenic process parameters-microstructure-property through neural network.

通过多级深冷处理实验获得目标零件的材料成分的组织-性能演化规律,具体包括以下步骤:以目标零件的材料为初始实验对象,在预先设置的深冷处理参数范围内,设置正交实验或者单因素试验,在不同深冷处理工艺参数条件下对所述初始试验对象进行多级深冷处理,并对深冷处理得到的样品进行微观组织表征和性能测试,获得不同深冷处理参数条件下样品中微观组织特征参数和性能数据,进而获得多级深冷处理工艺参数-微观组织-性能数据库。Obtain the microstructure-property evolution law of the material composition of the target part through the multi-stage cryogenic treatment experiment, which specifically includes the following steps: taking the material of the target part as the initial experimental object, within the preset cryogenic treatment parameter range, set up an orthogonal experiment or a single In the factor test, the initial test object is subjected to multi-stage cryogenic treatment under different cryogenic treatment process parameters, and the microstructure characterization and performance test are carried out on the samples obtained from the cryogenic treatment, so as to obtain the microstructural characteristics of the samples under different cryogenic treatment parameter conditions Parameter and performance data, and then obtain multi-stage cryogenic treatment process parameter-microstructure-performance database.

多级深冷处理是指从起始处理温度到最低处理温度之间经历若干次保温、升降温过程的工艺。微观组织特征参数包括晶粒尺寸、位错密度、残余奥氏体体积分数、碳化物相的尺寸和体积分数中的一种或者多种。所述多级深冷处理工艺参数包括温度、保温时间、升降温速率、处理次数中的一种或者多种。所述性能数据包括材料的力学性能,如屈服强度、断裂强度,电磁学性能如电导率、磁导率、热导率中的一种或者多种。Multi-stage cryogenic treatment refers to a process that undergoes several heat preservation, heating and cooling processes from the initial treatment temperature to the lowest treatment temperature. The characteristic parameters of microstructure include one or more of grain size, dislocation density, volume fraction of retained austenite, size and volume fraction of carbide phase. The multi-stage cryogenic treatment process parameters include one or more of temperature, holding time, heating and cooling rate, and treatment times. The performance data includes mechanical properties of the material, such as yield strength and fracture strength, and electromagnetic properties such as one or more of electrical conductivity, magnetic permeability, and thermal conductivity.

通过X射线衍射分析(XRD)确定深冷处理样品中相组成及位错密度,通过金相显微镜或者扫描电镜、EBSD等手段获得深冷处理样品中晶粒尺寸、残余奥氏体体积分数、碳化物相的尺寸和体积分数等。Determine the phase composition and dislocation density in cryogenically treated samples by X-ray diffraction analysis (XRD), and obtain the grain size, retained austenite volume fraction, and carbide phase in cryogenically treated samples by means of metallographic microscopy, scanning electron microscopy, and EBSD. size and volume fraction etc.

所建立的目标零件的材料成分的深冷处理工艺参数-微观组织-性能之间的非线性映射关系,包含深冷处理工艺参数-微观组织关系模型、微观组织-性能关系模型、深冷处理工艺参数-性能关系模型等3种关系模型。The established nonlinear mapping relationship between the cryogenic treatment process parameters-microstructure-performance of the material composition of the target part, including the cryogenic treatment process parameters-microstructure relationship model, the microstructure-performance relationship model, the cryogenic treatment process parameter-performance 3 relational models including relational model.

每种关系模型均采用含有多个隐含层的BP神经网络模型来描述,其中深冷工艺参数-微观组织非线性映射关系的建立包含如下步骤:Each relational model is described by a BP neural network model containing multiple hidden layers, and the establishment of the cryogenic process parameter-microstructure nonlinear mapping includes the following steps:

以所述多级深冷处理实验获得的深冷工艺参数的集合为输入,以获得的微观组织特征参数集合为输出,构建并训练一个含有多隐含层的BP神经网络模型,给各隐含层和输出层选择合适的激励函数;利用遗传算法优化所述BP神经网络初始的权值和阈值,得到最优个体的权值和阈值;将得到的最优个体的权值和阈值赋值给所述BP神经网络模型,训练过程中使用误差逆传播算法对各隐含层的权值和阈值进行更新,直到代价函数J小于设定精度或者达到最大迭代次数则训练结束。The set of cryogenic process parameters obtained by the multi-stage cryogenic treatment experiment is input, and the obtained microstructure characteristic parameter set is output, and a BP neural network model containing multiple hidden layers is constructed and trained to give each hidden layer and the output layer to select a suitable excitation function; utilize the genetic algorithm to optimize the initial weight and threshold of the BP neural network to obtain the weight and threshold of the optimal individual; assign the weight and threshold of the optimal individual obtained to the In the BP neural network model, the error back propagation algorithm is used to update the weights and thresholds of each hidden layer during the training process, and the training ends until the cost function J is less than the set accuracy or reaches the maximum number of iterations.

所述隐含层的激励函数均选择logistic函数,输出层的激励函数选择线性函数g(x)=x。The activation functions of the hidden layer all select the logistic function, and the activation functions of the output layer select the linear function g(x)=x.

采用遗传算法优化所述BP神经网络初始的权值和阈值,具体包括如下步骤:Optimizing the initial weights and thresholds of the BP neural network using a genetic algorithm specifically includes the following steps:

S1:首先根据所述BP神经网络模型的拓扑图确定神经网络的权值和阈值的个数,遵循如下公式:S1: First, determine the number of weights and thresholds of the neural network according to the topological diagram of the BP neural network model, following the following formula:

Figure BDA0003749236780000101
Figure BDA0003749236780000101

其中Num为权值和阈值的总个数,i表示第i层神经元,Hi为第i层神经元的节点数;Where N um is the total number of weights and thresholds, i represents the i-th layer of neurons, H i is the number of nodes of the i-th layer of neurons;

S2:采用实数编码方式对神经网络的权值阈值进行编码操作,初始化种群,初始的权值阈值在(-1,1)间随机取值,设置种群的适应度函数为F1;S2: Use the real number encoding method to encode the weight threshold of the neural network, initialize the population, the initial weight threshold is randomly selected between (-1,1), and set the fitness function of the population to F1;

Figure BDA0003749236780000102
Figure BDA0003749236780000102

其中F1为适应度值,ρ1为使用初始的权值和阈值的神经网络预测的增强相成分,η1为使用初始的权值和阈值的神经网络预测的增强相体积分数,δ1为使用初始的权值和阈值的神经网络预测的增强相平均尺寸,ρ0为增强相成分期望值,η0为增强相体积分数期望值,δ0为增强相平均尺寸期望值;Among them, F1 is the fitness value, ρ1 is the enhanced phase component predicted by the neural network using the initial weight and threshold, η1 is the enhanced phase volume fraction predicted by the neural network using the initial weight and threshold, and δ1 is the initial weight The average size of the enhanced phase predicted by the neural network of value and threshold, ρ0 is the expected value of the enhanced phase composition, η0 is the expected value of the volume fraction of the enhanced phase, and δ0 is the expected value of the average size of the enhanced phase;

S3:计算种群中所有个体的适应度值,并使用轮盘赌算法进行选择操作,从父代中挑选适应度高的个体产生下一代个体,每个个体被选中的概率遵循下公式:S3: Calculate the fitness value of all individuals in the population, and use the roulette algorithm to select individuals with high fitness from the parent generation to generate the next generation of individuals. The probability of each individual being selected follows the following formula:

Figure BDA0003749236780000103
Figure BDA0003749236780000103

其中,pk为第k个个体被选中的概率,Fk为第k个个体的适应度值,K为种群中个体的总数;Among them, p k is the probability that the kth individual is selected, F k is the fitness value of the kth individual, and K is the total number of individuals in the population;

S4:对种群中的个体进行交叉操作,设定交叉概率为pc,产生一个随机数若小于交叉概率,则进行交叉操作,交叉时随机选择两个个体并随机选择交叉位,遵循以下公式进行交叉操作:S4: Perform a crossover operation on the individuals in the population, set the crossover probability as pc, generate a random number that is less than the crossover probability, then perform the crossover operation, randomly select two individuals and randomly select the crossover position during the crossover, and perform crossover according to the following formula operate:

Figure BDA0003749236780000111
Figure BDA0003749236780000111

其中,akj是第k个染色体在j位上的实数,alj是第l个染色体在j位上的实数,b为(0,1)间的随机数;Among them, a kj is the real number of the k-th chromosome at the j-position, a lj is the real number of the l-th chromosome at the j-position, and b is a random number between (0,1);

S5:对种群中的个体进行变异操作,设定变异概率为pm,产生一个随机数若小于变异概率,则进行变异操作,变异时随机选择一个个体并随机选择变异位,遵循以下公式进行变异操作:S5: Perform a mutation operation on the individuals in the population, set the mutation probability to pm, generate a random number that is less than the mutation probability, then perform the mutation operation, randomly select an individual and randomly select the mutation bit during mutation, and perform the mutation operation according to the following formula :

Figure BDA0003749236780000112
Figure BDA0003749236780000112

Figure BDA0003749236780000113
Figure BDA0003749236780000113

其中,aij是第i个染色体在j位上的实数,g为当前迭代次数,f(g)是变异因子,Gmax为最大迭代次数,amax是aij取值的上限,amin是aij取值的下限,r和r'为(0,1)间的随机数;Among them, a ij is the real number of the i-th chromosome at position j, g is the current iteration number, f(g) is the variation factor, G max is the maximum iteration number, a max is the upper limit of the value of a ij , and a min is The lower limit of the value of a ij , r and r' are random numbers between (0,1);

S6:循环步骤S3至S5直至得到满意的适应度值或达到限定的迭代次数,输出最优的个体即适应度值最大的个体。S6: Repeat steps S3 to S5 until a satisfactory fitness value is obtained or a limited number of iterations is reached, and the optimal individual is output, that is, the individual with the largest fitness value.

步骤二,根据目标零件的组织性能分布要求将目标零件划分为多个子区域,结合得到的非线性映射关系反推出各个子区域所需的多级深冷处理参数,进而获得采用多级深冷处理目标零件时目标零件内部对应的一系列深冷工艺参数分布组合。Step 2: Divide the target part into multiple sub-regions according to the distribution requirements of the tissue performance of the target part, combine the obtained nonlinear mapping relationship to deduce the multi-level cryogenic treatment parameters required by each sub-region, and then obtain the multi-level cryogenic treatment target part A series of cryogenic process parameter distribution combinations corresponding to the inside of the target part.

步骤三,结合得到的深冷工艺参数分布组合对分区多级深冷处理过程进行高通量数值模拟,将数值模拟时温度分布或者微观组织分布最接近目标零件的组织性能分布要求所对应的深冷处理工艺参数组合确定为目标零件内部不同区域的最佳深冷处理参数组合。Step 3: Combining the obtained cryogenic process parameter distribution combination to carry out high-throughput numerical simulation on the partitioned multi-stage cryogenic treatment process, the temperature distribution or microstructure distribution in the numerical simulation is closest to the cryogenic treatment corresponding to the structure and performance distribution requirements of the target part The combination of process parameters is determined as the optimal combination of cryogenic treatment parameters for different regions inside the target part.

数值模拟时所采用商用数值模拟软件如Deform、Abaqus、Ansys或者Comsol对所述多级深冷处理过程进行传热模拟或传热-微观组织耦合模拟。In numerical simulation, commercial numerical simulation software such as Deform, Abaqus, Ansys or Comsol is used to perform heat transfer simulation or heat transfer-microstructure coupling simulation on the multi-stage cryogenic treatment process.

步骤四,结合目标零件的几何划分区域和最佳深冷处理参数组合对待处理零件进行分区独立式的深冷处理,以得到所述目标零件。Step 4: Combining the geometrically divided regions of the target part and the optimal cryogenic treatment parameter combination to perform partitioned and independent cryogenic treatment on the part to be treated, so as to obtain the target part.

以下以几个实施例来对本发明进行进一步的详细说明。The present invention will be further described in detail with several embodiments below.

实施例1Example 1

目标零件为高温合金GH2132双性能板状零件,零件中间部位工作温度低,要求组织晶粒尺寸细小,保证足够的强度、蠕变抗力和耐疲劳抗力,边缘部位承受的工作温度高,要求粗晶粒组织,保证室温塑性、蠕变抗力和抗疲劳扩展性能。因此,要求目标板状零件不同区域具有不同晶粒尺寸的显微组织结构,以获得相应的力学性能。根据零件实际服役工况,将目标梯度结构零件分为4个特征区域,中心部位晶粒为超细晶组织,平均晶粒小于1μm,边缘部位为粗晶粒组织,平均晶粒尺寸为100μm,中间区域的晶粒尺寸逐渐过渡。同时中心部位的维氏硬度不低于380HV,室温抗拉强度不低于800MPa,室温拉伸塑性不低于20%,边缘区域维氏硬度不低于300HV,室温抗拉强度不低于500MPa,室温拉伸塑性不低于40%。其分区多级深冷处理方法,包括如下步骤:The target part is a high-temperature alloy GH2132 dual-performance plate-shaped part. The middle part of the part has a low working temperature and requires a small grain size to ensure sufficient strength, creep resistance and fatigue resistance. The edge part bears high working temperature and requires coarse grain size. Granular structure, to ensure room temperature plasticity, creep resistance and fatigue expansion resistance. Therefore, different regions of the target plate-shaped part are required to have microstructures with different grain sizes in order to obtain corresponding mechanical properties. According to the actual service conditions of the parts, the target gradient structure parts are divided into 4 characteristic areas, the central part of the grain is ultra-fine grain structure, the average grain size is less than 1 μm, the edge part is coarse grain structure, the average grain size is 100 μm, The grain size in the middle region gradually transitions. At the same time, the Vickers hardness of the central part is not less than 380HV, the tensile strength at room temperature is not less than 800MPa, the tensile plasticity at room temperature is not less than 20%, the Vickers hardness of the edge area is not less than 300HV, and the tensile strength at room temperature is not less than 500MPa. Tensile plasticity at room temperature is not less than 40%. Its divisional multi-stage cryogenic treatment method includes the following steps:

(1)通过多级深冷处理实验获得高温合金GH2132的组织-性能演化规律,多级深冷处理工艺参数选择为温度,两段式温度曲线,包含温度T1和T2,微观组织参数选择为晶粒尺寸,性能参数选择为维氏硬度,抗拉强度和室温拉伸塑性。并通过BP神经网络建立多级深冷工艺参数-微观组织-性能之间的非线性映射关系。( 1 ) The microstructure-property evolution law of the superalloy GH2132 was obtained through multi-stage cryogenic treatment experiments. Particle size, performance parameters are selected as Vickers hardness, tensile strength and tensile plasticity at room temperature. And the non-linear mapping relationship between multi-level cryogenic process parameters-microstructure-performance is established through BP neural network.

深冷处理温度设置为-50℃,-75℃,-100℃,-125℃,-150℃,-175℃,-196℃,保温时间统一为1h,升降温速率50℃/min,处理次数为1。通过金相磨抛后拍摄金相照片确定多级深冷处理后的平均晶粒尺寸,通过ImageJ图像软件统计晶粒尺寸分布和平均晶粒尺寸。维氏硬度通过美国威尔逊(Wilson Hardness)公司的430SVD维氏硬度计进行测试,加载载荷为1kg,每个试样取5个点进行硬度测试,将个数据最大值和最小值去掉后求平均值作为该试样的硬度。拉伸棒料为直径10mm的标准拉伸试样,具体参照GB/T228-2010《金属材料室温拉伸试验方法》的要求进行室温准静态拉伸实验并统计室温拉伸强度和室温拉伸塑性数据。The cryogenic treatment temperature is set to -50°C, -75°C, -100°C, -125°C, -150°C, -175°C, -196°C, the holding time is uniformly 1h, the heating and cooling rate is 50°C/min, and the number of treatments is 1. The average grain size after multi-stage cryogenic treatment was determined by taking metallographic photos after metallographic grinding and polishing, and the grain size distribution and average grain size were counted by ImageJ image software. The Vickers hardness is tested by the 430SVD Vickers hardness tester of Wilson Hardness Company in the United States. The loading load is 1kg. Five points are taken for each sample for hardness testing. The maximum and minimum values of each data are removed and the average value is calculated. as the hardness of the sample. The tensile bar is a standard tensile sample with a diameter of 10mm. Specifically, refer to the requirements of GB/T228-2010 "Metallic Materials Tensile Test Method at Room Temperature" to conduct quasi-static tensile tests at room temperature and calculate the room temperature tensile strength and room temperature tensile plasticity. data.

通过BP神经网络建立温度,晶粒尺寸和维氏硬度、室温抗拉强度、室温拉伸塑性之间的非线性映射关系,以预实验获得的实验数据作为训练样本,直至达到理想的预测精度。The nonlinear mapping relationship between temperature, grain size and Vickers hardness, room temperature tensile strength and room temperature tensile plasticity is established through BP neural network, and the experimental data obtained in the pre-experiment is used as the training sample until the ideal prediction accuracy is achieved.

(2)根据零件实际服役工况,将目标梯度结构板状零件分为4个特征区域,结合获得的非线性映射关系,反推出各子区域所需的多级深冷处理参数,进而获得采用多级深冷处理目标零件时其内部的一系列深冷工艺参数分布组合。(2) According to the actual service conditions of the parts, the target gradient structural plate parts are divided into four characteristic areas, combined with the obtained nonlinear mapping relationship, the multi-level cryogenic treatment parameters required for each sub-area are reversely deduced, and then the multi-level cryogenic treatment parameters required by the multi-level sub-area are obtained. A series of cryogenic process parameter distribution combinations inside when the target part is treated in the first-level cryogenic treatment.

(3)结合步骤(2)获得目标零件内部一系列深冷工艺参数组合,对分区多级深冷处理过程进行高通量数值模拟,数值模拟采用comsol软件。将数值模拟时温度分布或者微观组织分布最接近目标零件的组织性能分布要求所对应的深冷处理工艺参数组合确定为目标零件内部不同区域的最佳深冷处理参数组合。本实施例中,各个子区域的多级深冷处理的参数为区域I:T1为-175℃,T2为-196℃,区域Ⅱ:T1为-150℃,T2为-175℃,区域Ⅲ:T1为-125℃,T2为-150℃,区域Ⅳ:T1为-100℃,T2为-125℃。升降温速率统一为50℃/min,每个温度阶段的保温时间为1h。(3) Combining step (2) to obtain a series of cryogenic process parameter combinations inside the target part, and perform high-throughput numerical simulation on the partitioned multi-stage cryogenic treatment process, using comsol software for numerical simulation. The combination of cryogenic treatment process parameters corresponding to the temperature distribution or microstructure distribution closest to the target part's tissue performance distribution requirements during numerical simulation is determined as the best combination of cryogenic treatment parameters for different regions inside the target part. In this embodiment, the parameters of multi-stage cryogenic treatment in each sub-region are as follows: Region I : T1 is -175°C, T2 is -196°C, Region II : T1 is -150°C, T2 is -175 °C, Zone III : T1 is -125°C, T2 is -150 °C, Zone IV : T1 is -100°C, T2 is -125 °C. The heating and cooling rate is uniformly 50°C/min, and the holding time of each temperature stage is 1h.

(4)结合步目标梯度结构板状零件几何划分区域和步骤(3)确定的最佳深冷处理参数组合,调整多腔室分区深冷装置的各腔室几何空间,各腔室按照步骤(3)确定的最佳深冷处理参数进行分区独立式的深冷处理,进而获得所述目标梯度结构板状零件。(4) In conjunction with the optimal cryogenic treatment parameter combination determined in the step target gradient structure plate-shaped part geometric division area and step (3), adjust the geometric space of each chamber of the multi-chamber partition cryogenic device, and each chamber is according to the step (3) ) to determine the optimal cryogenic treatment parameters to carry out subarea independent cryogenic treatment, and then obtain the target gradient structure plate-shaped part.

实施例2Example 2

目标零件为高温合金GH2132双性能盘形零件,同样的零件中间部位工作温度低,要求组织晶粒尺寸细小,保证足够的强度、蠕变抗力和耐疲劳抗力,边缘部位承受的工作温度高,要求粗晶粒组织,保证室温塑性、蠕变抗力和抗疲劳扩展性能。因此,要求目标板状零件不同区域具有不同晶粒尺寸的显微组织结构,以获得相应的力学性能。根据零件实际服役工况,将目标梯度结构零件分为3个特征区域,中心部位晶粒为超细晶组织,平均晶粒小于1μm,边缘部位为粗晶粒组织,平均晶粒尺寸为100μm,中间区域的晶粒尺寸逐渐过渡。同时中心部位的维氏硬度不低于380HV,室温抗拉强度不低于800MPa,室温拉伸塑性不低于20%,边缘区域维氏硬度不低于300HV,室温抗拉强度不低于500MPa,室温拉伸塑性不低于40%。The target part is a high-temperature alloy GH2132 dual-performance disc-shaped part. The middle part of the same part has a low working temperature, and the grain size of the structure is required to be small to ensure sufficient strength, creep resistance and fatigue resistance. The working temperature of the edge part is high, requiring Coarse grain structure ensures room temperature plasticity, creep resistance and fatigue expansion resistance. Therefore, different regions of the target plate-shaped part are required to have microstructures with different grain sizes in order to obtain corresponding mechanical properties. According to the actual service conditions of the parts, the target gradient structure parts are divided into three characteristic areas. The central part of the grain is ultra-fine grain structure, the average grain size is less than 1 μm, and the edge part is coarse grain structure, with an average grain size of 100 μm. The grain size in the middle region gradually transitions. At the same time, the Vickers hardness of the central part is not less than 380HV, the tensile strength at room temperature is not less than 800MPa, the tensile plasticity at room temperature is not less than 20%, the Vickers hardness of the edge area is not less than 300HV, and the tensile strength at room temperature is not less than 500MPa. Tensile plasticity at room temperature is not less than 40%.

预实验获得实验数据样本,并利用数据样本对神经网络进行训练直至达到理想的预测精度。根据神经网络建立的工艺参数-组织-性能之间的非线性映射关系反推各个区域的多级深冷处理参数。结合高通量数值模拟确定最佳的工艺参数组合;利用适用于盘形零件的分区多级深冷处理装置对目标梯度结构零件进行深冷处理,处理结束后取出零件即可获得具有梯度晶粒组织的目标梯度结构盘形零件。The pre-experiment obtains the experimental data samples, and uses the data samples to train the neural network until the ideal prediction accuracy is achieved. According to the nonlinear mapping relationship between process parameters-tissue-performance established by the neural network, the multi-level cryogenic treatment parameters of each region are reversed. Combining with high-throughput numerical simulation to determine the optimal combination of process parameters; using a partitioned multi-stage cryogenic treatment device suitable for disc-shaped parts to carry out cryogenic treatment on the target gradient structure parts, and taking out the parts after the treatment can obtain the gradient grain structure. Object gradient structure disk-shaped part.

本发明还提供了一种基于分区多级深冷处理的零件制备设备,所述设备采用如上所述的基于分区多级深冷处理的零件制备方法来制备零件的。所述设备形成有多个独立的深冷处理腔室,每个深冷处理腔室能够独立对待制备零件的一段区域进行深冷处理。The present invention also provides a parts preparation equipment based on subregional multi-stage cryogenic treatment, which adopts the above-mentioned part preparation method based on subregional multi-stage cryogenic treatment to prepare parts. The equipment is formed with a plurality of independent cryogenic treatment chambers, and each cryogenic treatment chamber can independently perform cryogenic treatment on a section of the part to be prepared.

请参阅图2、图3、图4及图5,本发明一个实施方式提供的基于分区多级深冷处理的零件制备设备,所述设备包括上管线室1、上腔室2、下腔室3、下管线室4、连接机构5、液氮管6、移动机构7、隔热层8、电阻加热丝9、热电偶10、热电偶信号线12、温度控制系统13及隔热板14。Please refer to Fig. 2, Fig. 3, Fig. 4 and Fig. 5, an embodiment of the present invention provides parts preparation equipment based on partitioned multi-stage cryogenic treatment, the equipment includes an upper pipeline chamber 1, an upper chamber 2, and a lower chamber 3 , lower pipeline chamber 4, connection mechanism 5, liquid nitrogen tube 6, moving mechanism 7, heat insulation layer 8, resistance heating wire 9, thermocouple 10, thermocouple signal line 12, temperature control system 13 and heat shield 14.

所述上管线室1、所述上腔室2、所述下腔室3及所述下管线室4依次通过连接机构5自上而下连接到一起,所述上管线室1与所述上腔室2固定连接,所述下腔室3与所述下管线室4固定连接到一起,所述上腔室2与所述下腔室3的一侧转动连接。所述上管线室1及所述下管线室4均为箱体,所述上腔室2及所述下腔室3为矩形框,且两者的结构及尺寸分别对应,以实现闭合。The upper pipeline chamber 1, the upper chamber 2, the lower chamber 3 and the lower pipeline chamber 4 are sequentially connected together from top to bottom through a connecting mechanism 5, and the upper pipeline chamber 1 and the upper pipeline chamber The chamber 2 is fixedly connected, the lower chamber 3 and the lower pipeline chamber 4 are fixedly connected together, and the upper chamber 2 is rotatably connected to one side of the lower chamber 3 . Both the upper pipeline chamber 1 and the lower pipeline chamber 4 are boxes, and the upper chamber 2 and the lower chamber 3 are rectangular frames, and the structures and sizes of the two are corresponding to realize closure.

所述上腔室2形成有第一收容腔,所述下腔室3形成有第二收容腔,所述上腔室2通过转动而抵靠在所述下腔室3上时,所述第一收容腔与所述第二收容腔相连通而形成一个收容腔。所述第一收容腔内间隔设置有多个隔热板14,所述第二收容腔也间隔设置有多个隔热板14,所述第一收容腔内的隔热板14的数量及位置与所述第二收容腔内的隔热板14的数量及位置分别对应,以将所述收容腔分割成多个独立的深冷处理腔室。相对的两个隔热板14分别开设有收容槽,两个收容槽相对所形成的卡槽的形状及尺寸分别与待制备的零件对应的区域的形状及尺寸相对应,所述卡槽用于卡紧零件的对应区域,以方便对该区域独立进行深冷处理。本实施方式中,待制备的零件为板状梯度结构零件11。The upper chamber 2 is formed with a first receiving chamber, and the lower chamber 3 is formed with a second receiving chamber. When the upper chamber 2 is turned against the lower chamber 3, the second A storage cavity communicates with the second storage cavity to form a storage cavity. A plurality of heat insulation boards 14 are arranged at intervals in the first storage chamber, and a plurality of heat insulation boards 14 are also arranged at intervals in the second storage chamber. The number and positions of the heat insulation boards 14 in the first storage chamber The number and position of the heat insulation boards 14 in the second storage chamber correspond respectively to divide the storage chamber into a plurality of independent cryogenic treatment chambers. The opposite two heat shields 14 are provided with receiving grooves respectively, and the shapes and sizes of the draw-in grooves formed by the two receiving grooves are respectively corresponding to the shapes and sizes of the regions corresponding to the parts to be prepared, and the draw-in grooves are used for Clamp the corresponding area of the part to facilitate independent cryogenic treatment of this area. In this embodiment, the part to be prepared is a plate-shaped gradient structure part 11 .

所述液氮管6分为进管、出管及支管,所述进管的数量及所述出管的数量相同,且均与所述深冷处理腔室的数量相同。多个所述进管的一端分别伸入所述上管线室1后连接于所述支管,所述支管呈n型,其两端分别穿过所述上管线室1的底板后伸入对应的深冷处理腔室。多个所述出管的一端伸入所述下管线室4后连接于所述支管,所述支管的两端分别穿过所述下管线室4的底板后进入所述深冷处理腔室,如此所述进管、所述深冷处理腔室及所述出管相连通。The liquid nitrogen pipe 6 is divided into an inlet pipe, an outlet pipe and a branch pipe, and the number of the inlet pipe and the outlet pipe are the same, and both are the same as the number of the cryogenic treatment chambers. One ends of the plurality of inlet pipes extend into the upper pipeline chamber 1 and then are connected to the branch pipes. The branch pipes are n-shaped, and their two ends respectively pass through the bottom plate of the upper pipeline chamber 1 and then extend into the corresponding Cryogenic processing chamber. One end of a plurality of outlet pipes stretches into the lower pipeline chamber 4 and is connected to the branch pipe, and the two ends of the branch pipe respectively pass through the bottom plate of the lower pipeline chamber 4 and then enter the cryogenic treatment chamber, so The inlet pipe, the cryogenic treatment chamber and the outlet pipe are connected.

每个深冷处理腔室的两端分别对称设置有电阻加热丝9及热电偶10。每个深冷处理腔室的内壁铺设有隔热层8,隔热层8可以为真空隔热板、气凝胶毡、硅酸铝纤维等;内壁之间用密封垫片隔绝不同深冷处理腔室的冷却介质,内壁与零件接触部位使用随形垫片,保证各腔室之间的密封性,可以为橡胶垫片、石棉垫片等导热性能差、密封效果好的垫片。Both ends of each cryogenic treatment chamber are respectively symmetrically provided with resistance heating wires 9 and thermocouples 10 . The inner wall of each cryogenic treatment chamber is laid with a thermal insulation layer 8, and the thermal insulation layer 8 can be a vacuum insulation board, airgel felt, aluminum silicate fiber, etc.; different cryogenic treatment chambers are isolated with sealing gaskets between the inner walls For the cooling medium, conformal gaskets are used at the contact parts between the inner wall and the parts to ensure the sealing between the chambers. Rubber gaskets, asbestos gaskets and other gaskets with poor thermal conductivity and good sealing effect can be used.

每个隔热板14分别连接有移动机构7,所述移动机构7用于支撑及带动所述隔热板14移动,以改变隔热板14两侧的深冷处理腔室的空间。其中,所述热电偶10通过热电偶信号线12连接于所述温度控制系统13,所述温度控制系统13还连接于所述电阻加热丝9、冷却介质供应系统及连接于所述冷却介质供应系统的进管。Each heat shield 14 is respectively connected with a moving mechanism 7 for supporting and driving the heat shield 14 to move so as to change the space of the cryogenic treatment chamber on both sides of the heat shield 14 . Wherein, the thermocouple 10 is connected to the temperature control system 13 through the thermocouple signal line 12, and the temperature control system 13 is also connected to the resistance heating wire 9, the cooling medium supply system and the cooling medium supply system. The inlet pipe of the system.

本实施方式,所述冷却介质可以为液氮、液氦或者其与酒精等介质的混合液体或气体,所述冷却介质通过冷却管通入深冷处理腔室内;温度传感器可以为热电偶,对称安装在待深冷处理的梯度零件两端,当然还可以为其他类型温度传感器,深冷处理腔室温度的确定为两个温度传感器的温度测量值的平均值,温度控制系统通过调节对应的电阻加热丝的电加热功率和冷却介质的流量来控制对应的深冷处理腔室的温度;所述移动机构负责调节腔室之间的隔热内壁的位置,灵活调节每个腔室的空间大小,以及负责零件在深冷装置内部的支撑和运动。In this embodiment, the cooling medium can be liquid nitrogen, liquid helium, or a mixed liquid or gas thereof with alcohol and other media, and the cooling medium is passed into the cryogenic processing chamber through a cooling pipe; the temperature sensor can be a thermocouple, symmetrically installed At both ends of the gradient part to be cryogenically treated, other types of temperature sensors can also be used. The temperature of the cryogenically treated chamber is determined as the average value of the temperature measurements of the two temperature sensors. The temperature control system adjusts the corresponding resistance heating wire. The temperature of the corresponding cryogenic treatment chamber is controlled by the electric heating power and the flow rate of the cooling medium; the moving mechanism is responsible for adjusting the position of the heat-insulating inner wall between the chambers, flexibly adjusting the space size of each chamber, and being responsible for the parts in the Support and movement inside a cryogenic unit.

请参阅图6、图7、图8及图9,本发明另一个实施方式提供的基于分区多级深冷处理的零件制备设备与上面实施方式提供的基于分区多级深冷处理的零件制备设备基本相同,不同点在于隔热板的形状及液氮管的设置。本实施方式待制备的零件为盘形梯度结构零件15,隔热板对应的为由弧形片形成,弧形片相背的两端重叠且连接成圆形后分别设置在所述第一收容腔及所述第二收容腔内,位于第一收容腔内的隔热板同轴设置,位于第二收容腔内的隔热板也同轴设置,如此将所述收容腔分割成多个环形的深冷处理腔室。其中通过改变弧形片相背的两端非重叠区域大小来改变对应的深冷处理腔室的空间大小。出管及进管分别通过一字型的支管与对应的深冷处理腔室相连通。Please refer to Fig. 6, Fig. 7, Fig. 8 and Fig. 9, the parts preparation equipment based on partitioned multi-stage cryogenic treatment provided by another embodiment of the present invention is basically the same as the part preparation equipment based on partitioned multi-stage cryogenic treatment provided in the above embodiment , the difference lies in the shape of the heat shield and the setting of the liquid nitrogen tube. The part to be prepared in this embodiment is a disc-shaped gradient structure part 15, and the corresponding heat shield is formed by an arc-shaped sheet. In the cavity and the second storage cavity, the heat insulation plate located in the first storage cavity is coaxially arranged, and the heat insulation plate located in the second storage cavity is also coaxially arranged, so that the storage cavity is divided into a plurality of annular cryogenic processing chamber. Wherein, the space size of the corresponding cryogenic processing chamber is changed by changing the size of the non-overlapping area at opposite ends of the arc-shaped sheet. The outlet pipe and the inlet pipe are respectively connected with the corresponding cryogenic treatment chamber through inline branch pipes.

本领域的技术人员容易理解,以上所述仅为本发明的较佳实施例而已,并不用以限制本发明,凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等,均应包含在本发明的保护范围之内。It is easy for those skilled in the art to understand that the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, All should be included within the protection scope of the present invention.

Claims (10)

1.一种基于分区多级深冷处理的零件制备方法,其特征在于,该方法包括以下步骤:1. A method for preparing parts based on subregional multistage cryogenic treatment, characterized in that the method may further comprise the steps: (1)通过多级深冷处理实验获得目标零件的材料成分的组织-性能演化规律,并通过神经网络建立多级深冷工艺参数-微观组织-性能之间的非线性映射关系;(1) Obtain the microstructure-property evolution law of the material composition of the target part through multi-stage cryogenic treatment experiments, and establish the non-linear mapping relationship between multi-stage cryogenic process parameters-microstructure-property through neural network; (2)根据目标零件的组织性能分布要求将目标零件划分为多个子区域,结合得到的非线性映射关系反推出各个子区域所需的多级深冷处理参数,进而获得采用多级深冷处理目标零件时目标零件内部对应的一系列深冷工艺参数分布组合;(2) Divide the target part into multiple sub-regions according to the distribution requirements of the target part's organization and performance, and deduce the multi-level cryogenic treatment parameters required for each sub-region by combining the obtained nonlinear mapping relationship, and then obtain the multi-level cryogenic treatment target part A series of cryogenic process parameter distribution combinations corresponding to the inside of the target part; (3)结合得到的深冷工艺参数分布组合对分区多级深冷处理过程进行高通量数值模拟,将数值模拟时温度分布或者微观组织分布最接近目标零件的组织性能分布要求所对应的深冷处理工艺参数组合确定为目标零件内部不同区域的最佳深冷处理参数组合;(3) Combining the obtained cryogenic process parameter distribution combination to carry out high-throughput numerical simulation on the partitioned multi-stage cryogenic treatment process, the temperature distribution or microstructure distribution in the numerical simulation is closest to the cryogenic treatment corresponding to the microstructure and performance distribution requirements of the target part The combination of process parameters is determined as the best combination of cryogenic treatment parameters for different regions inside the target part; (4)结合目标零件的几何划分区域和最佳深冷处理参数组合对待处理零件进行分区独立式的深冷处理,以得到所述目标零件。(4) Combining the geometrically divided regions of the target part and the optimal cryogenic treatment parameter combination to perform partitioned and independent cryogenic treatment on the part to be treated to obtain the target part. 2.如权利要求1所述的基于分区多级深冷处理的零件制备方法,其特征在于:目标零件的材料成分的组织-性能演化规律的获取包括以下步骤:以目标零件的材料为初始实验对象,在预先设置的深冷处理参数范围内,设置正交实验或者单因素试验,在不同深冷处理工艺参数条件下对所述初始试验对象进行多级深冷处理,并对深冷处理得到的样品进行微观组织表征和性能测试,获得不同深冷处理参数条件下样品中微观组织特征参数和性能数据,进而获得多级深冷处理工艺参数-微观组织-性能数据库。2. The part preparation method based on partitioned multi-stage cryogenic treatment as claimed in claim 1, characterized in that: the acquisition of the organization-performance evolution law of the material composition of the target part comprises the following steps: taking the material of the target part as the initial experimental object , within the pre-set range of cryogenic treatment parameters, an orthogonal experiment or a single factor test is set up, and the initial test object is subjected to multi-stage cryogenic treatment under different cryogenic treatment process parameters, and the microstructure of the sample obtained from the cryogenic treatment is analyzed. Characterization and performance testing, to obtain microstructure characteristic parameters and performance data in samples under different cryogenic treatment parameters, and then obtain a multi-level cryogenic treatment process parameter-microstructure-performance database. 3.如权利要求2所述的基于分区多级深冷处理的零件制备方法,其特征在于:多级深冷处理是指从起始处理温度到最低处理温度之间经历若干次保温、升降温过程的工艺;微观组织特征参数包括晶粒尺寸、位错密度、残余奥氏体体积分数、碳化物相的尺寸和体积分数中的一种或者多种;所述多级深冷处理工艺参数包括温度、保温时间、升降温速率、处理次数中的一种或者多种。3. The method for preparing parts based on partitioned multi-stage cryogenic treatment as claimed in claim 2, characterized in that: multi-stage cryogenic treatment refers to the process of several times of heat preservation and heating and cooling processes from the initial treatment temperature to the lowest treatment temperature. Process; microstructure characteristic parameters include one or more of grain size, dislocation density, retained austenite volume fraction, carbide phase size and volume fraction; the multi-stage cryogenic treatment process parameters include temperature, heat preservation One or more of time, heating and cooling rate, and processing times. 4.如权利要求1所述的基于分区多级深冷处理的零件制备方法,其特征在于:所建立的多级深冷工艺参数-微观组织-性能之间的非线性映射关系包含深冷处理工艺参数-微观组织关系模型、微观组织-性能关系模型、以及深冷处理工艺参数-性能关系模型。4. The method for preparing parts based on partitioned multi-stage cryogenic treatment as claimed in claim 1, characterized in that: the established multi-stage cryogenic process parameters-microstructure-performance nonlinear mapping relationship includes cryogenic treatment process parameters -Microstructure relationship model, microstructure-performance relationship model, and cryogenic treatment process parameter-performance relationship model. 5.如权利要求4所述的基于分区多级深冷处理的零件制备方法,其特征在于:每种关系模型均采用含有多个隐含层的BP神经网络模型来描述,其中深冷工艺参数-微观组织非线性映射关系的建立包含如下步骤:5. The method for preparing parts based on partitioned multi-stage cryogenic treatment as claimed in claim 4, wherein: each relational model is described by a BP neural network model containing a plurality of hidden layers, wherein the cryogenic process parameter- The establishment of the nonlinear mapping relationship of the microstructure includes the following steps: 以多级深冷处理实验获得的深冷工艺参数的集合为输入,以获得的微观组织特征参数集合为输出,构建并训练一个含有多隐含层的BP神经网络模型,给各隐含层和输出层选择合适的激励函数;利用遗传算法优化所述BP神经网络初始的权值和阈值,得到最优个体的权值和阈值;将得到的最优个体的权值和阈值赋值给所述BP神经网络模型,训练过程中使用误差逆传播算法对各隐含层的权值和阈值进行更新,直到代价函数J小于设定精度或者达到最大迭代次数则训练结束。Taking the set of cryogenic process parameters obtained from multi-level cryogenic treatment experiments as input, and the obtained set of microstructure characteristic parameters as output, a BP neural network model with multiple hidden layers is constructed and trained, and each hidden layer and output layer selects an appropriate activation function; optimizes the initial weights and thresholds of the BP neural network using a genetic algorithm to obtain the weights and thresholds of the optimal individual; assigns the weights and thresholds of the obtained optimal individual to the BP neural network In the network model, the error backpropagation algorithm is used to update the weights and thresholds of each hidden layer during the training process, until the cost function J is less than the set accuracy or reaches the maximum number of iterations, then the training ends. 6.如权利要求5所述的基于分区多级深冷处理的零件制备方法,其特征在于:隐含层的激励函数均选择logistic函数,输出层的激励函数选择线性函数g(x)=x。6. The part preparation method based on partitioned multi-stage cryogenic treatment as claimed in claim 5, characterized in that: the activation function of the hidden layer is selected from the logistic function, and the activation function of the output layer is selected from the linear function g(x)=x. 7.如权利要求5所述的基于分区多级深冷处理的零件制备方法,其特征在于:采用遗传算法优化所述BP神经网络初始的权值和阈值,具体包括如下步骤:7. The method for preparing parts based on subregional multistage cryogenic treatment as claimed in claim 5, characterized in that: adopt genetic algorithm to optimize the initial weight and threshold of the BP neural network, specifically comprising the steps of: S1:首先根据所述BP神经网络模型的拓扑图确定神经网络的权值和阈值的个数,遵循如下公式:S1: First, determine the number of weights and thresholds of the neural network according to the topological diagram of the BP neural network model, following the following formula:
Figure FDA0003749236770000031
Figure FDA0003749236770000031
其中Num为权值和阈值的总个数,i表示第i层神经元,Hi为第i层神经元的节点数;Where N um is the total number of weights and thresholds, i represents the i-th layer of neurons, H i is the number of nodes of the i-th layer of neurons; S2:采用实数编码方式对神经网络的权值阈值进行编码操作,初始化种群,初始的权值阈值在(-1,1)间随机取值,设置种群的适应度函数为F1;S2: Use the real number encoding method to encode the weight threshold of the neural network, initialize the population, the initial weight threshold is randomly selected between (-1,1), and set the fitness function of the population to F1;
Figure FDA0003749236770000032
Figure FDA0003749236770000032
其中F1为适应度值,ρ1为使用初始的权值和阈值的神经网络预测的增强相成分,η1为使用初始的权值和阈值的神经网络预测的增强相体积分数,δ1为使用初始的权值和阈值的神经网络预测的增强相平均尺寸,ρ0为增强相成分期望值,η0为增强相体积分数期望值,δ0为增强相平均尺寸期望值;Among them, F1 is the fitness value, ρ1 is the enhanced phase component predicted by the neural network using the initial weight and threshold, η1 is the enhanced phase volume fraction predicted by the neural network using the initial weight and threshold, and δ1 is the initial weight The average size of the enhanced phase predicted by the neural network of value and threshold, ρ0 is the expected value of the enhanced phase composition, η0 is the expected value of the volume fraction of the enhanced phase, and δ0 is the expected value of the average size of the enhanced phase; S3:计算种群中所有个体的适应度值,并使用轮盘赌算法进行选择操作,从父代中挑选适应度高的个体产生下一代个体,每个个体被选中的概率遵循下公式:S3: Calculate the fitness value of all individuals in the population, and use the roulette algorithm to select individuals with high fitness from the parent generation to generate the next generation of individuals. The probability of each individual being selected follows the following formula:
Figure FDA0003749236770000033
Figure FDA0003749236770000033
其中,pk为第k个个体被选中的概率,Fk为第k个个体的适应度值,K为种群中个体的总数;Among them, p k is the probability that the kth individual is selected, F k is the fitness value of the kth individual, and K is the total number of individuals in the population; S4:对种群中的个体进行交叉操作,设定交叉概率为pc,产生一个随机数若小于交叉概率,则进行交叉操作,交叉时随机选择两个个体并随机选择交叉位,遵循以下公式进行交叉操作:S4: Perform a crossover operation on the individuals in the population, set the crossover probability as pc, generate a random number that is less than the crossover probability, then perform the crossover operation, randomly select two individuals and randomly select the crossover position during the crossover, and perform crossover according to the following formula operate:
Figure FDA0003749236770000034
Figure FDA0003749236770000034
其中,akj是第k个染色体在j位上的实数,alj是第l个染色体在j位上的实数,b为(0,1)间的随机数;Among them, a kj is the real number of the k-th chromosome at the j-position, a lj is the real number of the l-th chromosome at the j-position, and b is a random number between (0,1); S5:对种群中的个体进行变异操作,设定变异概率为pm,产生一个随机数若小于变异概率,则进行变异操作,变异时随机选择一个个体并随机选择变异位,遵循以下公式进行变异操作:S5: Perform a mutation operation on the individuals in the population, set the mutation probability to pm, generate a random number that is less than the mutation probability, then perform the mutation operation, randomly select an individual and randomly select the mutation bit during mutation, and perform the mutation operation according to the following formula :
Figure FDA0003749236770000041
Figure FDA0003749236770000041
Figure FDA0003749236770000042
Figure FDA0003749236770000042
其中,aij是第i个染色体在j位上的实数,g为当前迭代次数,f(g)是变异因子,Gmax为最大迭代次数,amax是aij取值的上限,amin是aij取值的下限,r和r'为(0,1)间的随机数;Among them, a ij is the real number of the i-th chromosome at position j, g is the current iteration number, f(g) is the variation factor, G max is the maximum iteration number, a max is the upper limit of the value of a ij , and a min is The lower limit of the value of a ij , r and r' are random numbers between (0,1); S6:循环步骤S3至S5直至得到满意的适应度值或达到限定的迭代次数,输出最优的个体即适应度值最大的个体。S6: Repeat steps S3 to S5 until a satisfactory fitness value is obtained or a limited number of iterations is reached, and the optimal individual is output, that is, the individual with the largest fitness value.
8.一种基于分区多级深冷处理的零件制备设备,其特征在于:所述设备采用权利要求1-7任一项所述的基于分区多级深冷处理的零件制备方法来制备零件的;所述设备形成有多个独立的深冷处理腔室,每个深冷处理腔室能够独立对待制备零件的一段区域进行深冷处理。8. A parts preparation equipment based on partitioned multi-stage cryogenic treatment, characterized in that: said equipment adopts the part preparation method based on partitioned multi-stage cryogenic treatment according to any one of claims 1-7 to prepare parts; The above equipment is formed with a plurality of independent cryogenic treatment chambers, and each cryogenic treatment chamber can independently perform cryogenic treatment on a section of the part to be prepared. 9.如权利要求8所述的基于分区多级深冷处理的零件制备设备,其特征在于:每个深冷处理腔室分别设置有独立的电阻加热丝、温度传感器、冷却介质的进出管及移动机构,使得每个深冷处理腔室均能够单独工作以对目标零件的局部区域进行深冷处理。9. The parts preparation equipment based on partitioned multi-stage cryogenic treatment as claimed in claim 8, characterized in that: each cryogenic treatment chamber is provided with independent resistance heating wires, temperature sensors, cooling medium inlet and outlet pipes and moving mechanisms , enabling each cryogenic chamber to work independently to cryogenically treat a localized area of the target part. 10.如权利要求9所述的基于分区多级深冷处理的零件制备设备,其特征在于:所述设备还包括多个隔热板,且所述设备形成有收容腔,多个所述隔热板间隔设置在所述收容腔内以将所述收容腔分割成多个独立的所述深冷处理腔室;所述隔热板还用于承载待制备的零件;每个隔热板连接一个所述移动机构,所述移动机构用于带动所述隔热板进行移动,以改变隔热板两侧的深冷处理腔室的空间大小。10. The parts preparation equipment based on partitioned multi-stage cryogenic treatment as claimed in claim 9, characterized in that: the equipment also includes a plurality of heat insulation boards, and the equipment is formed with a storage chamber, and a plurality of the heat insulation plates Plates are arranged at intervals in the storage cavity to divide the storage cavity into a plurality of independent cryogenic treatment chambers; the heat insulation plates are also used to carry the parts to be prepared; each heat insulation plate is connected to a The moving mechanism is used to drive the heat shield to move, so as to change the space size of the cryogenic treatment chamber on both sides of the heat shield.
CN202210837481.6A 2022-07-15 2022-07-15 Part preparation method and equipment based on partitioned multi-stage cryogenic treatment Active CN115329475B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210837481.6A CN115329475B (en) 2022-07-15 2022-07-15 Part preparation method and equipment based on partitioned multi-stage cryogenic treatment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210837481.6A CN115329475B (en) 2022-07-15 2022-07-15 Part preparation method and equipment based on partitioned multi-stage cryogenic treatment

Publications (2)

Publication Number Publication Date
CN115329475A true CN115329475A (en) 2022-11-11
CN115329475B CN115329475B (en) 2023-04-25

Family

ID=83918203

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210837481.6A Active CN115329475B (en) 2022-07-15 2022-07-15 Part preparation method and equipment based on partitioned multi-stage cryogenic treatment

Country Status (1)

Country Link
CN (1) CN115329475B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120168039A1 (en) * 2009-07-03 2012-07-05 Snecma Cryogenic treatment of martensitic steel with mixed hardening
CN104133969A (en) * 2014-05-12 2014-11-05 北京航空航天大学 Method for calculating adsorption capture area of liquid helium cryogenic pump of vacuum plume experimental system
RU163939U1 (en) * 2016-02-24 2016-08-20 Федеральное государственное бюджетное образовательное учреждение высшего профессионального образования "Южно-Уральский государственный университет" (национальный исследовательский университет) (ФГБОУ ВПО "ЮУрГУ" (НИУ)) EJECTION COOLER COOLER IN COMBINED ENGINES
US20180028250A1 (en) * 2016-07-26 2018-02-01 Csa Medical, Inc. Method & apparatus to perform cryotherapy

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120168039A1 (en) * 2009-07-03 2012-07-05 Snecma Cryogenic treatment of martensitic steel with mixed hardening
CN104133969A (en) * 2014-05-12 2014-11-05 北京航空航天大学 Method for calculating adsorption capture area of liquid helium cryogenic pump of vacuum plume experimental system
RU163939U1 (en) * 2016-02-24 2016-08-20 Федеральное государственное бюджетное образовательное учреждение высшего профессионального образования "Южно-Уральский государственный университет" (национальный исследовательский университет) (ФГБОУ ВПО "ЮУрГУ" (НИУ)) EJECTION COOLER COOLER IN COMBINED ENGINES
US20180028250A1 (en) * 2016-07-26 2018-02-01 Csa Medical, Inc. Method & apparatus to perform cryotherapy

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
WU ZHISHENG等: "Effect of deep cryogenic treatment on electrode life and microstructure for spot welding hot dip galvanized steel", 《MATERIALS & DESIGN》 *
管晓光等: "深冷处理对WE43A异质焊补接头组织及性能影响研究", 《电焊机》 *

Also Published As

Publication number Publication date
CN115329475B (en) 2023-04-25

Similar Documents

Publication Publication Date Title
Gatsos et al. Review on computational modeling of process–microstructure–property relationships in metal additive manufacturing
CN112935251B (en) A kind of preparation method of amorphous alloy gradient composite material
WO2021057606A1 (en) Method for laser impact strengthening biphase titanium alloy additive component
US6660110B1 (en) Heat treatment devices and method of operation thereof to produce dual microstructure superalloy disks
CN114160796B (en) Heat treatment process method for preparing turbine disk and turbine disk
CN111753452B (en) Energy field assisted intelligent multi-point forming method and system for amorphous alloy part
Xiao et al. Surface roughness prediction model of GH4169 superalloy abrasive belt grinding based on multilayer perceptron (MLP)
Huang et al. Deformation behavior, microstructure evolution, phase transformation and plastic instability origin of powder metallurgy Al0. 8Co0. 5Cr1. 5CuFeNi alloy during high temperature deformation
CN110684938A (en) Method for predicting dynamic recrystallization grain size of metal or alloy material under variable-strain-rate working condition
CN115329475A (en) Part preparation method and equipment based on partitioned multistage cryogenic treatment
CN109540891A (en) The method of high temperature alloy heat treatment visual Simulation
CN109002628A (en) A kind of elevated temperature strength prediction technique for the martensite steel degenerated based on microstructure
Furrer et al. Optimizing the heat treatment of Ni-based superalloy turbine discs
Oumarou et al. A dynamic process model for predicting the performance of horizontal anode baking furnaces
CN115029544B (en) Heat treatment deformation control method of thin-walled ring parts based on simulation prediction
CN114166605B (en) A method for simulating and predicting the core microstructure and properties of large-size CrMo steel components
Wang et al. Microstructure evolution and surface cracking behavior of superheavy forgings during hot forging
Chen et al. Study of the microstructure evolution of alloy structural steel and inhomogeneity effect of the microscale pulsed currents during current-assisted plane strain compressions by modeling a novel cellular automata method
Liu et al. Study on Hot Deformation Behavior and Processing Maps of Cobalt‐Based Superalloy GH5188
Zouari et al. Multipass forging of Inconel 718 in the delta-Supersolvus domain: assessing and modeling microstructure evolution
CN114169225A (en) Method for optimizing machining sequence of aluminum alloy component based on computer simulation and computer equipment
CN114015862A (en) Vacuum annealing and performance uniformity control method for nuclear zirconium alloy pipe
CN103993114A (en) Method for controlling large-scale blast-furnace hot blast stove
CN110751991B (en) Method for predicting dissolution fraction of delta phase of nickel-based alloy containing Nb under time-varying working condition
CN208617944U (en) Heat treatment device for rapid electromagnetic energy aging of aluminum alloys

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