CN106845729A - A kind of electronic product rack assembly work based on gray theory determines method - Google Patents

Abstract

A kind of electronic product rack assembly work based on gray theory determines method, it is characterised in that：Comprise the following steps：Step 1：Obtain the factor to affect of influence rack assembly work：According to the requirement of rack assembly technology and assembly technology, the main factor to affect of analyzing influence rack assembly work；Step 2：Set up the rack assembly work computation model based on gray theory：Rack assembly work quota historical data is chosen, is set up and is based on GM（1, N）The rack assembly work computation model of model；Step 3：The cabinet design information of quota is treated in acquisition, using based on GM（1, N）The rack assembly work computation model of model calculates the rack man-hour for treating quota；Step 4：Modeling data is updated, modeling essential information is screened, it is ensured that set up the correctness of computation model.The present invention not only increases the accuracy of hour norm, and also improves the efficiency of quota.

Description

A kind of electronic product rack assembly work based on gray theory determines method

Technical field

The invention belongs to machine-building hour norm field, more particularly to electronic product rack assembly work quota field, Specifically a kind of electronic product rack assembly work based on gray theory determines method.

Background technology

With manufacturing industry and expanding economy, the precision that electronic product rack assembly work is formulated is proposed more with efficiency Requirement high, the hour norm is the essential information for realizing scientific management production, is to assist enterprise's plan of arranging production, real-time tracking The important side that the foundation of production status is production technology optimization, further taps the production potential, promote labor productivity to improve Method, be also enterprise product quotation, calculate human cost, evaluation economic benefit important evidence, therefore formulate man-hour precision and Efficiency has obtained extensive concern and research.

The production hour of product is by tact-time T_Single-pieceWith quasi- knot time T_{Quasi- knot}Composition.Wherein, single products Manufacturing time is made up of activity duration, operation allowance two parts again.

Activity duration：It is used for the time for completing single products production or consumption required for manufacture part.Activity duration is again By basic process time T_{Substantially}With auxiliary time of production T_AuxiliaryTwo parts are constituted, and basic process time is that workman enters according to technological requirement Row processing, the time required for making effective object that physics or chemical change to occur；Auxiliary time of production is to ensure to complete Time necessary to basic processing operation required for back work.

The operation allowance：I.e. because of place of arranging production, because of other works in operator's completion production operation or part processing Make to need to temporarily cease operation, because individual physiological demand and eliminate work produce it is tired needed for the temporary transient rest and reorganization time.

The quasi- knot time：When preparing before being produced needed for production single-piece or Bulk product and terminating required afterwards Between, as prepared production material, being familiar with the time that production drawing, adjusting device etc. are consumed.

To meet the demand of enterprise's customized production, should as far as possible shorten the time for formulating assembly work, improve and formulate assembling The precision in man-hour, therefore enterprise scien` can be helped to manage the hour norm, by the scientific management to the hour norm, improve enterprise Production efficiency, reduce enterprise cost of labor.During rack assembly work is formulated, intelligent algorithm plays crucial work With.

But present hour norm intelligent algorithm is required for substantial amounts of data as support, shortage solution data volume is few, The few hour norm problem of information content.It is considered herein that data volume is few when rack assembly work is formulated, inherence lacks bright True relation information, quota difficulty is big.The electronic product rack assembly work computational methods based on gray theory are proposed for this.

The content of the invention

During the present invention is for electronic product rack assembly work quota, quota personnel authorization difficulty is big, and efficiency is low, fixed A kind of the problems such as volume result is inaccurate, it is proposed that rack assembly work computational methods based on gray theory.According to history by norm Grey Systems Modelling information is obtained in database, rack assembly work computation model is set up, acquisition treats that the rack of hour norm sets Meter information, calculates rack assembly work, and quota data of going forward side by side are stored in history norm database, supplement history norm database.

The technical scheme is that：

A kind of electronic product rack assembly work based on gray theory determines method, it is characterised in that：Including following step Suddenly：

Step 1：Obtain the factor to affect of influence rack assembly work：According to the requirement of rack assembly technology and assembly technology, Analysis obtains influenceing the main factor to affect of rack assembly work.

Step 2：Set up the rack assembly work computation model based on gray theory：Choose rack assembly work quota history Data, set up the rack assembly work computation model based on GM (1, N) model.

Step 3：The cabinet design information of quota is treated in acquisition, is calculated using the rack assembly work based on GM (1, N) model Model calculates the rack man-hour for treating quota.

Step 4：Modeling data is updated, modeling essential information is screened, it is ensured that set up the correctness of computation model.

In the step 1, the factor to affect of rack assembly work is obtained, rack assembler is mainly passed through according to technique expert Skill draws, predominantly the quantity of the plug-in unit of rack, panel, screw, pad.

The rack assembly work computation model for setting up gray theory described in step 2, it is specific as follows:

The form of GM (1, N) model is exactly single order, and wherein N refers to the gray model for prediction of N number of characteristic factor. Gray theory has achieved the effect of highly significant by prolonged practice in many fields.GM (1, N) model is specific The process of foundation is that original series are carried out into one-accumulate generation, possesses stronger regularity by the new ordered series of numbers for generating, then use allusion quotation Type curve goes to fit correspondence trend, finally produces model with closest curve, is predicted finally by the system set up.

OrderIt is system features data sequence

OrderIt is correlative factor sequence

Initial dataBy 1-AGO Accumulating generations, single order Accumulating generation sequence is obtainedI.e.：

Wherein：

TakeSeries of mean

ThenWherein(as system features number According to) it is grey derivative,It is background value, then claims

It is GM (1, N) model.

In GM (1, N) model, a is referred to as System Development coefficient, b_ix_i ⁽¹⁾K () is referred to as driving item, b_iReferred to as drive factor,Referred to as parameter is arranged.

Construction data matrix B, Y

Wherein parameter ordered series of numbersFollowing relationship is met by application least squares estimate：

And the model of GM (1, N) isAlbefaction equation (the shadow side of this model Journey) it is as follows：

Known parameters a, solves albefaction equation, obtains：

WhenAmplitude of variation it is very small in the case of, in albefaction equationRegarded It is grey constant, then the approximate response type of GM (1, N) model can be reduced to：

Accumulation reduction-type is：

Numerical prediction is carried out using gray model GM (1, N), the numerical value for coming is predicted and is sometimes fluctuated meeting than larger, logarithm Influence is produced according to the development trend of row, ideal predicting the outcome can not be predicted.So must be to used herein GM (1, N) model is modified.When needing to be modified when rack assembly work is predicted, using the side of residual GM Method：

The residual error ordered series of numbers for making generation is ε⁰=(ε⁰(1),ε⁰(2),...,ε⁰(n)), obtain GM (1, N) model predication values with Difference between actual value, whereinOrdered series of numbers is X⁽¹⁾Residual sequence.If there is k₀, under satisfaction Row condition：

(1)Symbol it is consistent；

(2)n-k₀>=4, then claim (| ε⁽⁰⁾(k₀)|,|ε⁽⁰⁾(k₀+1)|,...,|ε⁽⁰⁾(n) |) for residual error rear can be modeled, still It is designated as ε⁽⁰⁾=(| ε⁽⁰⁾(k₀)|,|ε⁽⁰⁾(k₀+1)|,...,|ε⁽⁰⁾(n)|)。

To the modeled residual error rear sequence ε for obtaining⁽⁰⁾GM (1,1) model is set up, coefficient i.e. P=[a are calculated and obtain_ε, b_ε]^T, obtained by following computing formulaThe analogue value：

Then formula (12) is modified to

WhereinWith residual error rear ε⁽⁰⁾Symbol be consistent.

The design information of quota rack is treated in acquisition, uses the rack assembly work computation model based on gray theory set up The assembly work of quota rack is treated in calculating.

The beneficial effects of the invention are as follows：

The present invention has no the relation of determination, and not obvious rule due to quota historical data in the hour norm, This method is not merely simply to find probability distribution in quota historical data, obtains Rule Summary, but use generation Method weakens its randomness, obtains the ordered series of numbers of regular reinforcement.The computation model of rack assembly work is finally obtained, is not only dug Potential rule in historical data has been dug, the accuracy of hour norm has been improve, and also improve the efficiency of quota.

Specific embodiment

Below by example, technical method of the invention is described in further detail.

A kind of electronic product rack assembly work based on gray theory determines method, and it is comprised the following steps：

The first step：Obtain influence rack assembly work factor

Understood as Research foundation analysis using the structure composition and assembly technology of rack：Because rack uses structural module Design, is constituted, generally using same or analogous assembling technology procedure comprising same or similar assembling.Therefore these structures The assembly work of the parts being similar in composition also is more or less the same.For efficient, accurate completion rack Automatic manual transmission man-hour Intelligent estimation, by the research to cabinet structure feature and assembly technology：Rack Automatic manual transmission mainly by feature card, Panel, pinboard are fitted into rack, and mounting means is filled for spiral shell, it can be seen from Assembly part and mounting means, influence rack machinery The principal element of assembly work is as follows：

(1) the quantity N1 of plug-in unit.Plug number is more, and the man-hour for being installed into extension set is more, and adjusts the work of card address When it is more long.

(2) the quantity N2 of panel.The quantity of panel, directly determines the man-hour length for installing panel.

(3) the quantity N3 of screw.Main on-link mode (OLM) is spiral shell dress in rack Automatic manual transmission, and the quantity of screw directly affects dress The man-hour matched somebody with somebody, screw number is more, and the man-hour of consumption is more long.

(4) the quantity N4 of pad.Pad is the part that must be used in assembling process.The number of pad quantity is determined Assembly work needed for assembling process.

Information above can be obtained from design document, and influence assembly work also has some other factors, according to influence journey The size of degree, four factors of the above are that influence is larger, and other factors can be ignored.

The initial data of table 1

Second step：Set up the rack assembly work computation model based on gray theory：

Understand that grey systems GM (Isosorbide-5-Nitrae) model chooses plug-in unit, panel, screw, pad by studying rack assembly technology As influence factor and the prediction of rack assembly work is carried out, finally predicted value is compared with actual value, verify the model Accuracy and practicality.Model data is gathered by certain enterprise's erecting yard, and 12 groups of data are gathered altogether.Wherein 10 groups used as reality Data are tested, 2 groups used as verification data.According to the initial data in table 1, traditional GM (1, N) model (wherein N=4) is set up, counted Calculation obtains each coefficient in model, then obtains assembly work Grey models GM (1, N).

Make x₁,x₂,x₃,x₄,x₅Respectively assembly work and plug-in unit, panel, screw, the quantity of pad；Wherein assembly work Index X₁ ⁽⁰⁾It is the main behavior references object of system, and each variable has 10 groups of data as correlative factor sequence, i.e.,It is as follows：

The data in formula 17,18,19,20,21 are carried out into one-accumulate respectively respectively, is obtainedAnd Corresponding grey differential equation is set up, because this equation can not obtain accurate solution, equation is obtained using least square method herein Least squqre approximation solution.

Calculating parameter row u=[a, b₂,b₃,b₄,b₅] estimate beWherein B, Y Value is as follows respectively：

The value of B, Y is substituted intoU=[a, b can be calculated₂,b₃,b₄,b₅]^TLeast square Estimate isAsh is determined micro- Divide the coefficient of equation, be so as to can obtain assembly work forecast model：

According to 10 groups of experimental datas and the model of foundation, the assembly work of prediction and the assembly work such as following table of reality are drawn Shown in 2：

The assembly work of the prediction of table 2 compares with practical set man-hour

3rd step：Residual GM and model checking

The man-hour prediction ordered series of numbers of system features sequence is carried out into one-accumulate, obtaining new ordered series of numbers is

Remember that model prediction sequence is

The raw residual sequence of generation is

ε⁽⁰⁾=(0.2014,0.219,0.2332,0.2545,0.2966,0.3207,0.288,0.4803) (25)

According to data above, GM (1,1) model is set up, by calculating a_ε=-0.1223, b_ε=0.1631.Then

Obtain based on Residual GM (Isosorbide-5-Nitrae) grey forecasting model, being calculated rack Automatic manual transmission man-hour forecast model is：

Amendment principle according to Remanent Model, so this model is only to the 3rd (including the 3rd numerical value) predicted value later It is modified.In order to reach the requirement of actual quota, respectively using common grey GM (1, N) models and based on Residual GM (1, N) correction model carries out the prediction of rack Automatic manual transmission man-hour.I.e. with serial number 1-10, totally 10 sample trainings are modeled, and predict sequence number I Assembly work with II.Result is as shown in table 3.

The assembly work of the prediction of table 3 compares with practical set man-hour

Result is readily available from upper table, based on Residual GM (1, N) repairing model than common grey GM (1, N) model Accuracy increases, and Practical Performance is more preferable, it is adaptable to the prediction in Automatic manual transmission man-hour.

According to the standard of posteriority difference method of calibration, the science that C carrys out overall merit forecast model with P the two parameters is calculated Property.If system features sequence is x⁽⁰⁾, corresponding forecasting sequence isResidual sequence is obtained for ε, x is finally calculated⁽⁰⁾Average It is with variance：

And calculate the average and variance of residual epsilon and be：

Minimal error probability P=1 is can be calculated, mean square deviation ratio is C=0.12.Example proves the rack Automatic manual transmission work When forecast model possess good prediction effect, and error is controlled in less than 5%.

The present invention is summarized and analyzes to the influence factor of rack assembling first, and machine is set up using emerging gray theory Cabinet assembly work computation model, by the present invention so that when quota personnel carry out the hour norm, can quickly carry out rack assembling The quota in man-hour, effectively shortens the time of the fixed quotas of work, and the difficulty for solving rack assembly work quota is big, and accuracy difference etc. is asked Topic, and algorithm is feasible, high precision.

The preferred embodiments of the present invention are the foregoing is only, the present invention is not restricted to, for the technology of this area For personnel, the present invention can have various modifications and variations.It is all within the spirit and principles in the present invention, made any repair Change, equivalent, improvement etc., should be included within scope of the presently claimed invention.

Claims (3)

1. a kind of electronic product rack assembly work based on gray theory determines method, it is characterised in that：Comprise the following steps：

Step 1：Obtain the factor to affect of influence rack assembly work：According to the requirement of rack assembly technology and assembly technology, analysis Influence the main factor to affect of rack assembly work；

Step 2：Set up the rack assembly work computation model based on gray theory：Choose rack assembly work quota history number According to rack assembly work computation model of the foundation based on GM (1, N) model；

Step 3：The cabinet design information of quota is treated in acquisition, uses the rack assembly work computation model based on GM (1, N) model Calculate the rack man-hour for treating quota；

Step 4：Modeling data is updated, modeling essential information is screened, it is ensured that set up the correctness of computation model.

2. method according to claim 1, it is characterised in that：In the step 1, electronic product rack assembly work is influenceed Factor include：

Rack assembling is main based on manual operations, influence assembly work factor to affect include Assembly part quantity number and Assembly difficulty；Summing up main factor to affect according to rack assembly technology and flow is：Plug-in unit, panel, screw, the number of pad Amount.

3. method according to claim 1, it is characterised in that：Foundation described in the step 2 is based on GM (1, N) model Rack assembly work computation model GM (1, N) form for single order, wherein N refers to the gray system of N number of characteristic factor Forecast model；The process of specifically setting up of GM (1, N) model is that original series are carried out into one-accumulate generation, by the new number for generating Row possess stronger regularity, then are gone to fit correspondence trend with typical curve, finally produce model with closest curve, finally It is predicted by the system set up, is comprised the following steps that：

OrderIt is system features data sequence

$X_1^{\left(0\right)}=\left(x_1^{\left(0\right)}\right(1),x_1^{\left(0\right)}(2),...,x_1^{\left(0\right)}(n\left)\right)---\left(1\right)$

OrderIt is correlative factor sequence

$X_2^{\left(0\right)}=\left(x_2^{\left(0\right)}\right(1),x_2^{\left(0\right)}(2),...,x_2^{\left(0\right)}(n\left)\right)---\left(2\right)$

$X_3^{\left(0\right)}=\left(x_3^{\left(0\right)}\right(1),x_3^{\left(0\right)}(2),...,x_3^{\left(0\right)}(n\left)\right)---\left(3\right)$

$X_n^{\left(0\right)}=\left(x_n^{\left(0\right)}\right(1),x_n^{\left(0\right)}(2),...,x_n^{\left(0\right)}(n\left)\right)---\left(4\right)$

Initial dataBy 1-AGO Accumulating generations, single order Accumulating generation sequence is obtainedI.e.：

$X_i^{\left(1\right)}=\left(x_1^{\left(1\right)}\right(1),x_2^{\left(1\right)}(2),...,x_N^{\left(1\right)}(N\left)\right)---\left(5\right)$

Wherein：

$X_i^{\left(1\right)}\left(k\right)=\underset{j=1}{\overset{k}{\Σ}}{x}_i^{\left(0\right)}\left(j\right);(k=1,2,...,n)---\left(6\right)$

TakeSeries of mean

$z_1^{\left(1\right)}\left(k\right)=\frac{1}{2}\left(x_1^{\left(1\right)}\right(k)+x_1^{\left(1\right)}(k-1\left)\right),(k=2,3,...,n)---\left(7\right)$

ThenWherein(as system features data) are Grey derivative,It is background value, then claims

$x_1^{\left(0\right)}\left(k\right)+{\mathrm{az}}_1^{\left(1\right)}\left(k\right)=\underset{i=2}{\overset{N}{\Σ}}{b}_i{x}_i^{\left(1\right)}\left(k\right)---\left(8\right)$

It is GM (1, N) model.

In GM (1, N) model, a is referred to as System Development coefficient, b_ix_i ⁽¹⁾K () is referred to as driving item, b_iReferred to as drive factor,Referred to as parameter is arranged.

Construction data matrix B, Y

$B=\left[\begin{array}{cccc}-z_1^{\left(1\right)}\left(2\right)& x_2^{\left(1\right)}\left(2\right)& ...& x_N^{\left(1\right)}\left(2\right)\\ -z_1^{\left(1\right)}\left(3\right)& x_2^{\left(1\right)}\left(3\right)& ...& x_N^{\left(1\right)}\left(2\right)\\ ...& ...& ...& ...\\ -z_1^{\left(1\right)}\left(n\right)& x_2^{\left(1\right)}\left(n\right)& ...& x_N^{\left(1\right)}\left(n\right)\end{array}\right]$

$Y=\left[\begin{array}{c}{x}_1^{\left(0\right)}\left(2\right)\\ x_1^{\left(0\right)}\left(3\right)\\ ...\\ x_1^{\left(0\right)}\left(n\right)\end{array}\right]$

Wherein parameter ordered series of numbersFollowing relationship is met by application least squares estimate：

$\hat{a}={\left(B^{T}B\right)}^{-1}{B}^{T}Y---\left(9\right)$

And the model of GM (1, N) isThe albefaction equation (shadow equation) of this model is such as Under：

$\frac{{\mathrm{dx}}^{\left(1\right)}}{dt}+{\mathrm{ax}}_1^{\left(1\right)}=b_2{x}_2^{\left(1\right)}\left(k\right)+b_3{x}_3^{\left(1\right)}\left(k\right)+...+b_n{x}_n^{\left(1\right)}\left(k\right)---\left(10\right)$

Known parameters a, solves albefaction equation, obtains：

$x^{\left(1\right)}\left(t\right)=e^{-at}\[\underset{i=2}{\overset{n}{\Σ}}\∫b_i{x}_i^{\left(1\right)}\left(t\right)e^{at}dt+x^{\left(1\right)}\left(0\right)-\underset{i=2}{\overset{n}{\Σ}}\∫b_i{x}_i^{\left(1\right)}\left(0\right)dt\]---\left(11\right)$

WhenAmplitude of variation it is very small in the case of, in albefaction equationIt is considered as ash Constant, then the approximate response type of GM (1, N) model can be reduced to：

${\hat{x}}_1^{\left(1\right)}(k+1)=\left(x_1^{\left(1\right)}\right(0)-\frac{1}{a}\underset{i=2}{\overset{n}{\Σ}}{b}_i{x}_i^{\left(1\right)}(k+1\left)\right)e^{-at}+\frac{1}{a}\underset{i=2}{\overset{n}{\Σ}}{b}_i{x}_i^{\left(1\right)}(k+1)---\left(12\right)$

Accumulation reduction-type is：

${\hat{x}}_1^{\left(0\right)}(k+1)=a^{\left(1\right)}{\hat{x}}_1^{\left(1\right)}={\hat{x}}_1^{\left(1\right)}(k+1)-{\hat{x}}_1^{\left(1\right)}\left(k\right)---\left(13\right)$

Numerical prediction is carried out using gray model GM (1, N), the numerical value for coming is predicted and is sometimes fluctuated and data can be arranged than larger Development trend produce influence, ideal predicting the outcome can not be predicted；So must to GM used herein (1, N) model is modified；When needing to be modified when rack assembly work is predicted, using the method for residual GM：

The residual error ordered series of numbers for making generation is ε⁰=(ε⁰(1),ε⁰(2),...,ε⁰(n)), obtain GM (1, N) model predication values and actual value Between difference, whereinOrdered series of numbers is X⁽¹⁾Residual sequence；If there is k₀, meet following condition：

(1)Symbol it is consistent；

(2)n-k₀>=4, then claim (| ε⁽⁰⁾(k₀)|,|ε⁽⁰⁾(k₀+1)|,...,|ε⁽⁰⁾(n) |) for residual error rear can be modeled, still it is designated as ε⁽⁰⁾=(| ε⁽⁰⁾(k₀)|,|ε⁽⁰⁾(k₀+1)|,...,|ε⁽⁰⁾(n)|)。

To the modeled residual error rear sequence ε for obtaining⁽⁰⁾GM (1,1) model is set up, coefficient i.e. P=[a are calculated and obtain_ε,b_ε]^T, By arriving for following computing formulaThe analogue value：

${\widehat{\mathrm{\ϵ}}}^0(k+1)=(-a_{\mathrm{\ϵ}})\left({\mathrm{\ϵ}}^0\right(k_0)-\frac{b_{\mathrm{\ϵ}}}{a_{\mathrm{\ϵ}}})e^{-a_{\mathrm{\ϵ}}(k-k_0)}---\left(14\right)$

Then formula 12 is modified to

${\hat{x}}_1^{\left(1\right)}(k+1)=\left\{\begin{array}{c}\left(x_1^{\left(0\right)}\right(0)-\frac{1}{a}\underset{i=2}{\overset{n}{\&Sigma;}}{b}_i{x}_i^{\left(1\right)}(k+1\left)\right)e^{-at}+\frac{1}{a}\underset{i=2}{\overset{n}{\&Sigma;}}{b}_i{x}_i^{\left(1\right)}(k+1),k<k_0\\ \left(x_1^{\left(1\right)}\right(0)-\frac{1}{a}\underset{i=2}{\overset{n}{\&Sigma;}}{b}_i{x}_i^{\left(1\right)}(k+1\left)\right)e^{-at}+\frac{1}{a}\underset{i=2}{\overset{n}{\&Sigma;}}{b}_i{x}_i^{\left(1\right)}(k+1)\&PlusMinus;{\widehat{\mathrm{\&epsiv;}}}^{\left(0\right)}(k+1),k\&GreaterEqual;k_0\end{array}\right.---\left(15\right)$

WhereinWith residual error rear ε⁽⁰⁾Symbol be consistent；

Treat that quota cabinet design information performs that above-mentioned algorithm sets up based on gray theory rack assembly work computation model, i.e., The assembly work for treating quota rack can be calculated.