CN106339548A - Tolerance automatic distribution method and high-reliability load switch design method - Google Patents

Abstract

The invention relates to a tolerance automatic distribution method and a high-reliability load switch design method. According to the invention, firstly, a function relationship of a key factor center value and a key factor tolerance and reliability is established, and a function relationship of the key factor tolerance and processing feasibility as well as processing cost is established; then a key factor tolerance automatic distribution multi-objective optimization model is constructed, and a weight coefficient of each objective function is determined; the key factor tolerance of a load switch is automatically distributed; finally, according to a tolerance distribution result, comprehensive indexes of reliability, processing feasibility and processing cost of a product are calculated; if the calculated comprehensive indexes meet set threshold requirements, the design method is completed. By the evaluation method, the problem of low processing design reliability of the load switch product under the influence of key factor fluctuation caused by processing dispersibility is solved.

Description

A kind of tolerance auto-allocation method and highly reliable on-load switch method for designing

Technical field

The invention belongs to on-load switch product reliability design technical field, particularly to a kind of tolerance auto-allocation method And highly reliable on-load switch method for designing.

Background technology

In on-load switch product design production process, its output characteristics is determined by technical specification.Ideally, product Output characteristics will be consistent with technical specification.However, product is in life cycle management, due to the shadow by various interference factors Ring, lead to its output to assume distribution characteristicss, and the central value being distributed also may deviate technical specification.Tolerance Allocation can ensure While properties of product, product anti-interference is improved by control design case tolerance.Therefore, carried out using tolerance auto-allocation method On-load switch reliability Allowance Design is significant to improving product reliability.

Tolerance Allocation is under the jurisdiction of Reliability Distribution category, and the more equal distribution method of application, redistribution procedure, agree divide at present Join method etc. and belong to System Reliability Assignment Method, the premise of its application is the reliability between whole system and unit or subsystem Property block diagram distinct, inefficacy relation determine.Though on-load switch can be considered system, its mechanical, electrical, magnetic-coupled feature makes its whole machine lose Necessary connection is had no between effect and each component failures.According to the feature of on-load switch, Tolerance Allocation is carried out to it and belongs to non-linear Multi-objective optimization question.Simulated annealing is based on Monte Carlo iterations solution strategies, takes when scanning for solution domain space " balance " strategy, both using existing information, can enter new subspace so that certain probability changes current search direction again, can be Limits ground approaching overall situation approximate optimal solution, is highly suitable for the automatic assigning process of tolerance of on-load switch.

Content of the invention

It is an object of the invention to provide a kind of tolerance auto-allocation method, in order to solve tolerance distribution and tolerance automatically The inaccurate problem of allocation result, present invention also offers a kind of method for designing of highly reliable on-load switch, for solving Under the key factor influence of fluctuations that processing dispersibility leads to, the problem of the reliability fabrication design of on-load switch product.

For realizing the problems referred to above, the technical scheme is that

A kind of automatic distribution method of on-load switch tolerance based on simulated annealing, comprises the following steps:

(1) determine key factor and its initial value participating in Tolerance Allocation；

(2) produce object function initial value, and determine initial temperature；

(3) lower the temperature；

(4) random disturbance produce each key factor value now, and before and after calculation perturbation object function increment δ；

(5) if δ is more than zero, accept the value of each key factor after disturbance, if δ is less than zero, with exp (δ/bt_k) Probability accept the value of each key factor after disturbance；

(6) judge whether Markov process is stable, if Markov process is unstable, execution step (4)-(5), if Markov process is stable, and Tolerance Allocation scheme also meets the threshold requirement of setting, then terminate searching process, exports Tolerance Allocation As a result, if Tolerance Allocation scheme is unsatisfactory for the threshold requirement setting, execution step (3)-(6), until Tolerance Allocation scheme is full Foot requires.

A kind of highly reliable on-load switch method for designing, the method comprises the following steps:

S1 sets up the functional relationship of key factor central value and tolerance and reliability, setting up key factor tolerance can with processing Row and the functional relationship of processing cost；

S2 builds key factor tolerance and automatically distributes Model for Multi-Objective Optimization；

S3 determines the weight coefficient of each object function；

S4 carries out the automatic distribution of tolerance based on simulated annealing principle, and obtaining makes object function meet given threshold requirement Tolerance Allocation scheme；

S5 calculates the aggregative indicator of the reliability, Machinability Evaluation and processing cost of product according to Tolerance Allocation result；

If the aggregative indicator that s6 calculates is unsatisfactory for the threshold requirement setting, execution step s2-s5, if calculate is comprehensive Close index and meet the threshold requirement setting, then complete this method for designing.

Wherein, step s4, on-load switch tolerance auto-allocation method, specifically comprise the following steps that

(1) determine key factor and its initial value participating in Tolerance Allocation；

(2) produce object function initial value, and determine initial temperature；

(3) lower the temperature；

(4) random disturbance produce each key factor value now, and before and after calculation perturbation object function increment δ；

(5) if δ is more than zero, accept the value of each key factor after disturbance, if δ is less than zero, with exp (δ/bt_k) Probability accept the value of each key factor after disturbance；

(6) judge whether Markov process is stable, if Markov process is unstable, execution step (4)-(5), if Markov process is stable, and Tolerance Allocation scheme also meets the threshold requirement of setting, then terminate searching process, exports Tolerance Allocation As a result, if Tolerance Allocation scheme is unsatisfactory for the threshold requirement setting, execution step (3)-(6), until Tolerance Allocation scheme is full Foot requires.

In described step s1, key factor central value and tolerance with the functional relationship of reliability are:

$r=\underset{i=1}{\overset{n}{\π}}{r}_i=\underset{i=1}{\overset{n}{\π}}p(x_i>y_i)=\underset{i=1}{\overset{n}{\π}}\underset{-\mathrm{\∞}}{\overset{\mathrm{\∞}}{\&integral;}}\[\underset{y}{\overset{\mathrm{\∞}}{\&integral;}}{f}_i\left(x\right)dx\]g_i\left(y\right)dy$

Wherein, r_iExpression key factor i ∈ (1,2 ..., intensity n) is x_i, stress be y_iCorresponding reliability；f_i(x) For characterizing the probability density function of key factor i central value and tolerance respective intensities；g_iY () is the corresponding stress of key factor i Probability density function；

Key factor tolerance with the functional relationship of Machinability Evaluation and processing cost is:

$\left\{\begin{array}{c}{f}_{cap}\left(x\right)=100(w_{ca1}\frac{x_1-x_{d1}}{x_{u1}-x_{d1}}+w_{ca2}\frac{x_2-x_{d2}}{x_{u2}-x_{d2}}+...+w_{can}\frac{x_n-x_{dn}}{x_{un}-x_{dn}})\\ f_{\cos t}\left(x\right)=100(w_{co1}\frac{x_{u1}-x_1}{x_{u1}-x_{d1}}+w_{co2}\frac{x_{\mathrm{u2}}-x_{d2}}{x_{u2}-x_{d2}}+...+w_{con}\frac{x_{un}-x_n}{x_{un}-x_{dn}})\end{array}\right.$

Wherein, f_cap(x) and f_costThe corresponding Machinability Evaluation target of (x) respectively n key factor tolerance x and processing Cost objective；x、x_u、x_dIt is respectively the next characteristic key factor tolerance, tolerance limit and the lower variation of tolerance of relay；w_ca、w_coPoint Cloth is for the weight occupied by each key factor when Machinability Evaluation function and cost function.

In described step s2, tolerance is automatically distributed Model for Multi-Objective Optimization and is:

$\left\{\begin{array}{c}x=(x_1,x_2,...x_n)\\ \max f\left(x\right)\\ s.t.f\left(x\right)=w{\[r_1\left(x\right),r_2\left(x\right),...,r_m\left(x\right),f_{cap}\left(x\right),f_{\cos t}\left(x\right)\]}^t\\ r_j\left(x\right)>r_{j\_req},(j=1,2,...,m)\\ x_i>0,(i=1,2,...,n)\end{array}\right.$

In formula, w=[w₁,w₂,...,w_m+2] be each object function weight coefficient；r_jX () represents that course of action j is corresponding Reliability objectives；r_{j_req}X () represents the corresponding reliability index requirements of course of action j.

A kind of beneficial effect of tolerance auto-allocation method of the present invention is: the present invention solve tolerance automatically distribution and The inaccurate problem of Tolerance Allocation result.

The beneficial effect of a kind of highly reliable on-load switch method for designing of the present invention is: the technical side that the present invention provides Case, while considering reality processing ability and processing cost, has very big to the reliability design technology of on-load switch Raising, complete the design of highly reliable on-load switch product.

By the actual requirement of on-load switch reliability Allowance Design, set up its tolerance and automatically distribute multiple-objection optimization mould Type, by the analysis to each target significance level, determines the weight of object function, and by tolerance, distribution obtains dividing of tolerance automatically Join.

Brief description

Fig. 1 is a kind of method for designing flow chart of highly reliable on-load switch；

Fig. 2 is a kind of tolerance auto-allocation method flow chart.

Specific embodiment

A kind of automatic distribution method of on-load switch tolerance of the present invention, this distribution method is based on simulated annealing and carries out tolerance Automatic distribution, obtain and make object function meet the Tolerance Allocation scheme of given threshold requirement, specifically comprise the following steps that a, determine ginseng Key factor and its initial value with Tolerance Allocation；B, generation object function initial value, and determine initial temperature；C, cooling；D, random Disturbance produce each key factor value now, and before and after calculation perturbation object function increment δ；E, judge that whether δ is more than 0, It is to accept the value of each key factor after disturbance, otherwise with exp (δ/bt_k) probability accept each key factor after disturbance Value；F, judge that Markov process is whether stable, be then execution " g ", otherwise execute " d "；G, judge that Tolerance Allocation scheme is No satisfaction requires, if being unsatisfactory for requiring, repeats " c "～" f ", until Tolerance Allocation scheme meets requirement, terminates optimizing Journey, exports the automatic allocation result of tolerance.

A kind of method for designing of the highly reliable on-load switch of the present invention, the specifically comprising the following steps that of the method

1 be based on Reliability Distinguishing criterion and " stress-intensity " interference theory set up key factor central value and its tolerance with Functional relationship between on-load switch production reliability；Set up reflection Machinability Evaluation and processing cost and each key factor tolerance Functional relationship；

2 set up the tolerance many mesh of distribution automatically including the object functions such as reliability index, Machinability Evaluation and processing cost Mark Optimized model；

3 set up hierarchical chart describes the associated relation between targets at different levels and related keyword factor；By step analysis Method determines the weight coefficient of the object functions such as reliability index, Machinability Evaluation and processing cost；

4 Model for Multi-Objective Optimization set up with step 2, as object function, carry out on-load switch pass by simulated annealing The automatic distribution of key factor tolerance；

5 according to step 4 optimize after on-load switch key factor central value and tolerance allocation result, by step 1 institute The functional relationship stated calculates production reliability, Machinability Evaluation and processing cost；Judge reliability, Machinability Evaluation and be processed into Whether this aggregative indicator meets the threshold requirement of setting, if be unsatisfactory for, repeated execution of steps 2～step 5, until described Aggregative indicator meets the threshold requirement setting, then complete the design of the on-load switch of high reliability based on simulated annealing.

In step 1, described Reliability Distinguishing criterion mathematic(al) representation is:

$\left\{\begin{array}{c}{f}_{contact}-f_{contact-req}>0\\ min(t_c-t_{c-req},t_b-t_{b-req})>0\\ min(v_{c-b}-v_{c-b-req},v_{b-b}-v_{b-b-req})>0\\ \max (v_{c-c}-v_{c-c-req},v_{c-a}-v_{c-a-req},v_{b-c}-v_{b-c-req},v_{b-c}-v_{b-c-req})<0\end{array}\right.$

Wherein, f_contact、t_c、t_bIt is respectively contact force during adhesive, pickup time and release time；v_c-b、v_c-c、v_c-a、 v_b-b、v_b-c、v_b-aIt is respectively adhesive and separation initial velocity, contact impact velocity and the armature of release process contact and yoke collides Speed.All expressions particular requirement value of lower marker tape " req "；

In step 1, key factor central value and tolerance with the functional relationship of reliability are:

$r=\underset{i=1}{\overset{n}{\&pi;}}{r}_i=\underset{i=1}{\overset{n}{\&pi;}}p(x_i>y_i)=\underset{i=1}{\overset{n}{\&pi;}}\underset{-\mathrm{\&infin;}}{\overset{\mathrm{\&infin;}}{\&integral;}}\&lsqb;\underset{y}{\overset{\mathrm{\&infin;}}{\&integral;}}{f}_i\left(x\right)dx\&rsqb;g_i\left(y\right)dy$

Wherein, r_iExpression key factor i ∈ (1,2 ..., intensity n) is x_i, stress be y_iCorresponding reliability；f_i(x) For characterizing the probability density function of key factor i central value and tolerance respective intensities；g_iY () is the corresponding stress of key factor i Probability density function；

Key factor tolerance with the functional relationship of Machinability Evaluation and processing cost is:

$\left\{\begin{array}{c}{f}_{cap}\left(x\right)=100(w_{ca1}\frac{x_1-x_{d1}}{x_{u1}-x_{d1}}+w_{ca2}\frac{x_2-x_{d2}}{x_{u2}-x_{d2}}+...+w_{can}\frac{x_n-x_{dn}}{x_{un}-x_{dn}})\\ f_{\cos t}\left(x\right)=100(w_{co1}\frac{x_{u1}-x_1}{x_{u1}-x_{d1}}+w_{co2}\frac{x_{\mathrm{u2}}-x_{d2}}{x_{u2}-x_{d2}}+...+w_{con}\frac{x_{un}-x_n}{x_{un}-x_{dn}})\end{array}\right.$

Wherein, f_cap(x) and f_costThe corresponding Machinability Evaluation target of (x) respectively n key factor tolerance x and processing Cost objective；x、x_u、x_dIt is respectively the next characteristic key factor tolerance, tolerance limit and the lower variation of tolerance of relay；w_ca、w_coPoint Cloth is for the weight occupied by each key factor when Machinability Evaluation function and cost function.

Tolerance is automatically distributed Model for Multi-Objective Optimization and is in step 2:

$\left\{\begin{array}{c}x=(x_1,x_2,...x_n)\\ \max f\left(x\right)\\ s.t.f\left(x\right)=w{\&lsqb;r_1\left(x\right),r_2\left(x\right),...,r_m\left(x\right),f_{cap}\left(x\right),f_{\cos t}\left(x\right)\&rsqb;}^t\\ r_j\left(x\right)>r_{j\_req},(j=1,2,...,m)\\ x_i>0,(i=1,2,...,n)\end{array}\right.$

Wherein, w=[w₁,w₂,...,w_m+2] be each object function weight coefficient；r_jX () represents that course of action j is corresponding Reliability objectives；r_{j_req}X () represents the corresponding reliability index requirements of course of action j.

In step 3, determine reliability in terms of contact breaking speed, contact impact velocity and armature impact velocity three Degree requires；By tuning parameter and machining parameter determination Machinability Evaluation and processing cost function；Set up hierarchical chart to retouch State the associated relation between targets at different levels and related keyword factor；Determine that reliability index, processing are feasible by analytic hierarchy process (AHP) Property relative weighting with the object function such as processing cost；By important scale matrix eigenvalue, the solution of characteristic vector and each After the product calculation of the sub- weight of level, obtain each reliability objectives, Machinability Evaluation target and processing cost target with regard to catalogue Target absolute weight.

In step 4, a, the key factor determining participation Tolerance Allocation and its initial value；B, generation object function are initial Value, and determine initial temperature；C, cooling；D, random disturbance produce each key factor value now, and object function before and after calculation perturbation Increment δ；E, judge that whether δ is more than 0, be to accept the value of each key factor after disturbance, otherwise with exp (δ/bt_k) Probability accepts the value of each key factor after disturbance；F, judge that Markov process is whether stable, be then execution " g ", otherwise hold Row " d "；G, judging whether Tolerance Allocation scheme meets requirement, if being unsatisfactory for requiring, repeating " c "～" f ", until tolerance Allocative decision meets requirement, terminates searching process, exports the automatic allocation result of tolerance.

Claims (7)

1. a kind of on-load switch tolerance auto-allocation method it is characterised in that: methods described is based on simulated annealing being held The automatic distribution of difference, obtains the Tolerance Allocation scheme making object function meet given threshold requirement.

2. a kind of on-load switch tolerance auto-allocation method according to claim 1, specifically comprises the following steps that

(1) determine key factor and its initial value participating in Tolerance Allocation；

(2) produce object function initial value, and determine initial temperature；

(3) lower the temperature；

(4) random disturbance produce each key factor value now, and before and after calculation perturbation object function increment δ；

(5) if δ is more than zero, accept the value of each key factor after disturbance, if δ is less than zero, with exp (δ/bt_k) general Rate accepts the value of each key factor after disturbance；

(6) judge whether Markov process is stable, if Markov process is unstable, execution step (4)-(5), Ruo Maer Can husband's process stabilization, Tolerance Allocation scheme also meets the threshold requirement of setting, then terminate searching process, output Tolerance Allocation knot Really, if Tolerance Allocation scheme is unsatisfactory for the threshold requirement setting, execution step (3)-(6), until Tolerance Allocation scheme meets Require.

3. a kind of method for designing of highly reliable on-load switch is it is characterised in that specifically comprise the following steps that

S1 sets up the functional relationship of key factor central value and tolerance and reliability, sets up key factor tolerance and Machinability Evaluation And the functional relationship of processing cost；

S2 builds key factor tolerance and automatically distributes Model for Multi-Objective Optimization；

S3 determines the weight coefficient of each object function；

S4. carry out the automatic distribution of tolerance based on simulated annealing, obtain the appearance making object function meet given threshold requirement Difference allocative decision；

S5 calculates the aggregative indicator of the reliability, Machinability Evaluation and processing cost of product according to Tolerance Allocation result；

If the aggregative indicator that s6 calculates is unsatisfactory for the threshold requirement setting, execution step s2-s5, if the synthesis calculating refers to Mark meets the threshold requirement setting, then complete this method for designing.

4. a kind of method for designing of highly reliable on-load switch according to claim 3 is it is characterised in that be based on simulation The on-load switch tolerance auto-allocation method of annealing method, specifically comprises the following steps that

(1) determine key factor and its initial value participating in Tolerance Allocation；

(2) produce object function initial value, and determine initial temperature；

(3) lower the temperature；

(4) random disturbance produce each key factor value now, and before and after calculation perturbation object function increment δ；

(5) if δ is more than zero, accept the value of each key factor after disturbance, if δ is less than zero, with exp (δ/bt_k) general Rate accepts the value of each key factor after disturbance；

(6) judge whether Markov process is stable, if Markov process is unstable, execution step (4)-(5), Ruo Maer Can husband's process stabilization, Tolerance Allocation scheme also meets the threshold requirement of setting, then terminate searching process, output Tolerance Allocation knot Really, if Tolerance Allocation scheme is unsatisfactory for the threshold requirement setting, execution step (3)-(6), until Tolerance Allocation scheme meets Require.

5. a kind of method for designing of highly reliable on-load switch according to claim 3 is it is characterised in that key factor Central value and tolerance with the functional relationship of reliability are:

$r=\underset{i=1}{\overset{n}{\&pi;}}{r}_i=\underset{i=1}{\overset{n}{\&pi;}}p(x_i>y_i)=\underset{i=1}{\overset{n}{\&pi;}}\underset{-\mathrm{\&infin;}}{\overset{\mathrm{\&infin;}}{\&integral;}}\&lsqb;\underset{y}{\overset{\mathrm{\&infin;}}{\&integral;}}{f}_i\left(x\right)dx\&rsqb;g_i\left(y\right)dy$

Wherein, r_iExpression key factor i ∈ (1,2 ..., intensity n) is x_i, stress be y_iCorresponding reliability；f_iX () is to characterize Key factor i central value and the probability density function of tolerance respective intensities；g_iY () is close for the probability of the corresponding stress of key factor i Degree function.

6. method for designing, key factor tolerance and the processing of a kind of highly reliable on-load switch according to claim 3 The functional relationship of feasibility and processing cost is:

$\left\{\begin{array}{c}{f}_{cap}\left(x\right)=100(w_{ca1}\frac{x_1-x_{d1}}{x_{u1}-x_{d1}}+w_{ca2}\frac{x_2-x_{d2}}{x_{u2}-x_{d2}}+...+w_{can}\frac{x_n-x_{dn}}{x_{un}-x_{dn}})\\ f_{\cos t}\left(x\right)=100(w_{co1}\frac{x_{u1}-x_1}{x_{u1}-x_{d1}}+w_{co2}\frac{x_{u2}-x_2}{x_{u2}-x_{d2}}+...+w_{con}\frac{x_{un}-x_n}{x_{un}-x_{dn}})\end{array}\right.$

Wherein, f_cap(x) and f_costThe corresponding Machinability Evaluation target of (x) respectively n key factor tolerance x and processing cost mesh Mark；x、x_u、x_dIt is respectively the next characteristic key factor tolerance, tolerance limit and the lower variation of tolerance of relay；w_ca、w_coIt is distributed as pin To the weight occupied by each key factor when Machinability Evaluation function and cost function.

7. a kind of highly reliable on-load switch according to any one of claim 3-6 method for designing it is characterised in that: Described tolerance distributes Model for Multi-Objective Optimization automatically:

$\left\{\begin{array}{c}x=(x_1,x_2,...x_n)\\ \max f\left(x\right)\\ s.t.f\left(x\right)=w{\&lsqb;r_1\left(x\right),r_2\left(x\right),...,r_m\left(x\right),f_{cap}\left(x\right),f_{\cos t}\left(x\right)\&rsqb;}^t\\ r_j\left(x\right)>r_{j\_req}(j=1,2,...,m)\\ x_i>0(i=1,2,...,n)\end{array}\right.$

Wherein, w=[w₁,w₂,...,w_m+2] be each object function weight coefficient；r_jX () represents the corresponding reliability of course of action j Property target；r_{j_req}X () represents the corresponding reliability index requirements of course of action j.