CN103632054A - Spare part demand predicting method based on state monitoring and reliability of equipment part - Google Patents

Abstract

The invention relates to a spare part demand predicting method based on state monitoring and reliability of an equipment part and belongs to the technical field of mechanical manufacturing. The method includes the following steps that firstly, according to a maintenance record list of equipment, the reliability theory is used for processing burn-out life data of parts in the maintenance record with the reliability theory, and the probability cumulative distribution function of the burn-out life of the parts is obtained; the probability cumulative distribution function of the burn-out life of the parts, the total working time of the equipment, the history record of the real spare part demand quantity, the year-to-year value record and the month-to-month value record of the spare part demand quantity, the experience predicting values of planners and the like are compared with errors, used for comparison predicting, of the real spare part demand quantity so as to obtain the predicting value of the demand quantity of equipment spare parts. According to the method, the prediction result of the method has the practical bases, the inactive stock of an enterprise can be effectively reduced, resource waste is reduced, cost is reduced and the method has good adaptability to different spare parts which are prone to damage.

Description

A kind of spare parts demand Forecasting Methodology based on status monitoring and part of appliance reliability

Technical field

The present invention relates to a kind of spare parts demand Forecasting Methodology based on status monitoring and part of appliance reliability, belong to machinery manufacturing technology field.

Background technology

The maintenance process of large-scale and complicated device is complicated, for shortening the idle hours of repairing, guarantees maintenance efficiency, and the Parts Inventory of some is necessary.Parts Inventory shortage may cause equipment not come into operation by on-call maintenance, causes huge economic loss; Spares reserve too much produces a large amount of dead stocks, and causing warehouse site cost, retention fee, spare part to damage extra consumptions such as loss of causing increases, and takies a large amount of working capitals of enterprises, affects enterprise operation.Therefore, rational prediction spare parts demand amount reduces Parts Inventory amount under the prerequisite of the normal maintenance requirements of assurance equipment, for reducing business capital, is detained, increases economic efficiency, and has great significance.

Traditional standby demand forecast be the record of chain rate on year-on-year basis by spare parts demand amount, next month the factor such as sales policy carry out experience and estimate to predict spare parts demand amount, formulate procurement plan, this is also the method that current most of enterprise adopts.Owing to lacking digitized Forecasting Methodology, the estimation of spare parts demand mainly relies on planner's experience, and be subject to going into operation in the following a period of time impact of the many factors such as number of devices, on-stream time, seasonal factor, sales policy of spare parts demand amount, rule of thumb is difficult to accomplish demand forecast accurately.

The method of the quantitative forecast of mentioning in some scientific research documents is as computing method such as single exponent smoothing, Function Fitting recurrence, neural networks, conventionally need a large amount of data to carry out Fitting Analysis etc., but in actual production owing to being subject to the quantitative limitation of quantity matter, be difficult to matching accurate, the over-fitting often occurring and cross study phenomenon and cause predictablity rate lower.

Along with the development of wireless sensor technology and Condition Monitoring Technology, the maintenance based on state (Condition-Based Maintenance, CBM) obtains promotion and application widely just gradually.Maintenance based on state is the monitoring analysis of equipment running status data to be determined to the preventive maintenance pattern of maintenance of equipment demand.Monitored status data conventionally by sensor, is gathered or device intelligence terminal is uploaded, formation Condition Monitoring Data stream, and upload to continuously equipment condition monitoring analysis platform by data sensor networks such as wireless sense network, GPRS, Internet and satellites, resolve, a series of analytic processes such as storage and eigenwert extraction, denoising, data fusion, determine maintenance of equipment demand.The information such as the service time of the equipment that these Condition Monitoring Datas reflect, extent of deterioration, possible trouble location and reason, the to a certain extent health status in following a period of time and type and the quantity that may need to change spare part for predict device.

Prior art has proposed to utilize status monitoring information to carry out spare part Forecasting Methodology, but is all conventionally based upon in a basic condition: need the parts of prediction with sensing device, can obtain its status data.And in actual conditions, in view of problems such as sensor cost, volumes, be not that all parts on equipment can be monitored to.The emphasis of status monitoring is more crucial or valuable parts on equipment often, and the easily worn part large but less expensive to some demands, sensor installation is predicted the spare parts demand amount of its following a period of time, high cost separately.

Summary of the invention

The object of the invention is to propose a kind of spare parts demand Forecasting Methodology based on status monitoring and equipment dependability, consumable accessory for demand in enormous quantities in large-scale and complicated device carries out spare parts demand prediction, improve the accuracy of prediction, easy to implement, can under the prerequisite of the normal maintenance requirements of assurance equipment, reduce Parts Inventory amount, reduce business capital and be detained, save operation cost of enterprises.

The equipment and spare part needing forecasting method based on monitoring running state and equipment dependability that the present invention proposes, comprises the following steps:

(1) from equipment maintenance record list, obtain the burn-out life of part of appliance, according to reliability theory, the normal distribution model matching of use Weibull model, Weibull competing risk model or brachymemma obtains the probability cumulative distribution function F (t) in component failure life-span;

(2) N platform equipment is carried out to status monitoring, obtain every equipment j in N platform equipment at current time T _ithe total duration t of work _ij, 1≤j≤N wherein, the total duration of work of N platform equipment is

(3) according to the total duration of work of the equipment obtaining in step (2) and the probability cumulative distribution function F (t) in the component failure life-span obtaining in step (1), utilize following formula, calculate N platform its upper side part at current time T _ithe real work age

${\stackrel{\→}{a}}_i={\stackrel{\→}{a}}_{i-1}+\mathrm{\Δ}{\stackrel{\→}{t}}_i-\stackrel{\→}{g}\left(S_i\right),$

The real work age

for N dimensional vector,

a wherein _ijbe expressed as parts on j platform equipment at the real work age of current time Ti,

for parts are at T _i-1the real work age constantly,

mod represents remainder number, B _yfor the corresponding component failure life-span duration of the Y quantile of the cumulative distribution function F with step (1) (t);

for time interval [T _i-1, T _i] increment of the total duration of work of interior equipment,

for life-span recharge function, refer to that parts on equipment are at time interval [T _i-1, T _i] internal cause is replaced the value that real work age of causing reduces, wherein S _ifor time interval [T _i-1, T _i] interior actual Inventory of Spare Parts, life-span recharge function

the computing method of value are: respectively to parts at T _i-1real work age a constantly _{i-1, j}with time interval [T _i-1, T _i] the increment Delta t of the total duration of work of interior equipment _i,jsum up, obtain one and add and be worth, according to adding and be worth descending order, equipment is sorted, set sequence at front S _iparts on the equipment of name are at time interval [T _i-1, T _i] in be replaced, before S _ig (the S of name equipment _i) value is the burn-out life duration B of these parts _y, the g (S of all the other equipment _i) value is 0;

(4) utilize formula obtain T _i+1the parts real work age constantly, wherein

to be time interval [T _i, T _i+1] estimated value of increment of the total duration of work of interior equipment, according to historical same period the equipment situation that goes into operation estimate to obtain;

(5) step (4) is obtained in the real work age a of every its upper side part _{i+1, j}respectively with component failure life-span duration B _ycompare, if a _{i+1, j}>B _y, the parts on this equipment may lose efficacy, the component count that may lose efficacy add and, obtain time interval [T _i, T _i+1] the interior number of components that may lose efficacy

(6) to the above-mentioned number of components that may lose efficacy

the historical demand of equipment and spare part

with the experience estimated value of equipment management personnel to component demand amount

be weighted summation, obtain time interval [T _i, T _i+1] predictor formula of interior equipment and spare part demand:

α wherein ₀, α ₁, α ₂be respectively

with

weight;

(7) predicated error between the spare part actual demand amount S in the predicted value S ' of spare parts demand and equipment history run record is expressed as to ε, ε=| S '-S|, according to the spare part actual demand amount in equipment history run record, by minimizing m the predicated error sum in the time period

calculate component failure life-span duration B _yin Y value and above-mentioned weight α ₀, α ₁, α ₂value;

(8) Y value and the α that step (7) are obtained ₀, α ₁, α ₂value substitution formula

in, calculate [T _i, T _i+1] the equipment and spare part Demand Forecast value of period

The spare parts demand Forecasting Methodology based on status monitoring and equipment dependability that the present invention proposes, its advantage is:

(1) the inventive method is utilized current actual the use age of Condition Monitoring Data estimation section, bonding apparatus Condition Monitoring Data and part reliability model carry out spare part prediction, actual foundation has predicted the outcome, more accurate compared with Classical forecast method, can effectively reduce enterprise's dead stock, reduce the wasting of resources, reduce costs.

(2) the inventive method is for the large easy loss parts of demand, in conjunction with enterprise practical, consider, the total duration of work by monitoring component place equipment but not the state of each parts oneself predict, saves parts deploy sensor and gather the cost of automatically controlled signal.

(3) the inventive method is adjusted Prediction Parameters by minimizing predicated error, and different damageable spare parts is had to good universality.

Embodiment

The equipment and spare part needing forecasting method based on monitoring running state and equipment dependability that the present invention proposes, comprises the following steps:

(1) from equipment maintenance record list, obtain the burn-out life of part of appliance, according to reliability theory, the normal distribution model matching of use Weibull model, Weibull competing risk model or brachymemma obtains the probability cumulative distribution function F (t) in component failure life-span; With Weibull, be distributed as example, cumulative distribution function F (t)=1-exp[-(t/ α) ^β], wherein α, β are the parameter that needs matching.According to the idiographic flow of reliability theory matching cumulative distribution function, refer to the Jiang Ren of mechanical engineering publishing house speech, the left bright strong < < reliability model of showing and application > >.

(2) N platform equipment is carried out to status monitoring, obtain every equipment j in N platform equipment at current time T _ithe total duration t of work _ij, 1≤j≤N wherein, the total duration of work of N platform equipment is

(3) according to the total duration of work of the equipment obtaining in step (2)

and the probability cumulative distribution function F (t) in the component failure life-span obtaining in step (1), utilize following formula, calculate N platform its upper side part at current time T _ithe real work age

${\stackrel{\&RightArrow;}{a}}_i={\stackrel{\&RightArrow;}{a}}_{i-1}+\mathrm{\&Delta;}{\stackrel{\&RightArrow;}{t}}_i-\stackrel{\&RightArrow;}{g}\left(S_i\right),$

The real work age for N dimensional vector,

a wherein _ijbe expressed as parts on j platform equipment at the real work age of current time Ti,

for parts are at T _i-1the real work age constantly,

mod represents remainder number, B _yfor example, for the corresponding component failure life-span duration of the Y quantile of the cumulative distribution function F with step (1) (t), B ₉₀, represent parts cumulative failure probability corresponding burn-out life duration while being 90%.

for time interval [T _i-1, T _i] increment of the total duration of work of interior equipment,

for life-span recharge function, refer to that parts on equipment are at time interval [T _i-1, T _i] internal cause is replaced the value that real work age of causing reduces, wherein S _ifor time interval [T _i-1, T _i] interior actual Inventory of Spare Parts, life-span recharge function

the computing method of value are: respectively to parts at T _i-1real work age a constantly _{i-1, j}with time interval [T _i-1, T _i] the increment Delta t of the total duration of work of interior equipment _i,jsum up, obtain one and add and be worth, according to adding and be worth descending order, equipment is sorted, set sequence at front S _iparts on the equipment of name are at time interval [T _i-1, T _i] in be replaced, before S _ig (the S of name equipment _i) value is the burn-out life duration B of these parts _y, the g (S of all the other equipment _i) value is 0;

(4) utilize formula

obtain T _i+1the parts real work age constantly, wherein

to be time interval [T _i, T _i+1] estimated value of increment of the total duration of work of interior equipment, according to historical same period the equipment situation that goes into operation estimate to obtain;

(5) step (4) is obtained

in the real work age a of every its upper side part _{i+1, j}respectively with component failure life-span duration B _ycompare, if a _{i+1, j}>B _y, the parts on this equipment may lose efficacy, the component count that may lose efficacy add and, obtain time interval [T _i, T _i+1] the interior number of components that may lose efficacy

(6) to the above-mentioned number of components that may lose efficacy

the historical demand of equipment and spare part with the experience estimated value of equipment management personnel to component demand amount

be weighted summation, obtain time interval [T _i, T _i+1] predictor formula of interior equipment and spare part demand:

α wherein ₀, α ₁, α ₂be respectively with

weight;

(7) predicated error between the spare part actual demand amount S in the predicted value S ' of spare parts demand and equipment history run record is expressed as to ε, ε=| S '-S|, according to the spare part actual demand amount in equipment history run record, by minimizing m the predicated error sum in the time period

calculate component failure life-span duration B _yin Y value and above-mentioned weight α ₀, α ₁, α ₂value;

(8) Y value and the α that step (7) are obtained ₀, α ₁, α ₂value substitution formula

in, calculate [T _i, T _i+1] the equipment and spare part Demand Forecast value S ' of period _i+1.

Claims (1)

1. the equipment and spare part needing forecasting method based on monitoring running state and equipment dependability, is characterized in that the method comprises the following steps:

(1) from equipment maintenance record list, obtain the burn-out life of part of appliance, according to reliability theory, the normal distribution model matching of use Weibull model, Weibull competing risk model or brachymemma obtains the probability cumulative distribution function F (t) in component failure life-span;

(2) N platform equipment is carried out to status monitoring, obtain every equipment j in N platform equipment at current time T _ithe total duration t of work _ij, 1≤j≤N wherein, the total duration of work of N platform equipment is

(3) according to the total duration of work of the equipment obtaining in step (2)

and the probability cumulative distribution function F (t) in the component failure life-span obtaining in step (1), utilize following formula, calculate N platform its upper side part at current time T _ithe real work age

${\stackrel{\&RightArrow;}{a}}_i={\stackrel{\&RightArrow;}{a}}_{i-1}+\mathrm{\&Delta;}{\stackrel{\&RightArrow;}{t}}_i-\stackrel{\&RightArrow;}{g}\left(S_i\right),$

The real work age

for N dimensional vector,

a wherein _ijbe expressed as parts on j platform equipment at the real work age of current time Ti,

for parts are at T _i-1the real work age constantly, mod represents remainder number, B _yfor the corresponding component failure life-span duration of the Y quantile of the cumulative distribution function F with step (1) (t);

for time interval [T _i-1, T _i] increment of the total duration of work of interior equipment,

for life-span recharge function, refer to that parts on equipment are at time interval [T _i-1, T _i] internal cause is replaced the value that real work age of causing reduces, wherein S _ifor time interval [T _i-1, T _i] interior actual Inventory of Spare Parts, life-span recharge function

the computing method of value are: respectively to parts at T _i-1real work age a constantly _{i-1, j}with time interval [T _i-1, T _i] the increment Delta t of the total duration of work of interior equipment _i,jsum up, obtain one and add and be worth, according to adding and be worth descending order, equipment is sorted, set sequence at front S _iparts on the equipment of name are at time interval [T _i-1, T _i] in be replaced, before S _ig (the S of name equipment _i) value is the burn-out life duration B of these parts _y, the g (S of all the other equipment _i) value is 0;

(4) utilize formula

obtain T _i+1the parts real work age constantly, wherein to be time interval [T _i, T _i+1] estimated value of increment of the total duration of work of interior equipment, according to historical same period the equipment situation that goes into operation estimate to obtain;

(5) step (4) is obtained

in the real work age a of every its upper side part _{i+1, j}respectively with component failure life-span duration B _ycompare, if a _{i+1, j}>B _y, the parts on this equipment may lose efficacy, the component count that may lose efficacy add and, obtain time interval [T _i, T _i+1] the interior number of components that may lose efficacy

(6) to the above-mentioned number of components that may lose efficacy

the historical demand of equipment and spare part

with the experience estimated value of equipment management personnel to component demand amount

be weighted summation, obtain time interval [T _i, T _i+1] predictor formula of interior equipment and spare part demand:

α wherein ₀, α ₁, α ₂be respectively

with

weight;

(7) predicated error between the spare part actual demand amount S in the predicted value S ' of spare parts demand and equipment history run record is expressed as to ε, ε=| S '-S|, according to the spare part actual demand amount in equipment history run record, by minimizing m the predicated error sum in the time period calculate component failure life-span duration B _yin Y value and above-mentioned weight α ₀, α ₁, α ₂value;

(8) Y value and the α that step (7) are obtained ₀, α ₁, α ₂value substitution formula

in, calculate [T _i, T _i+1] the equipment and spare part Demand Forecast value S ' of period _i+1.