WO2010017745A1

WO2010017745A1 - Reliability predicting method of communication device

Info

Publication number: WO2010017745A1
Application number: PCT/CN2009/073069
Authority: WO
Inventors: 尤荣贤; 张学渊
Original assignee: 中兴通讯股份有限公司
Priority date: 2008-08-14
Filing date: 2009-08-04
Publication date: 2010-02-18
Also published as: CN101340326A; CN101340326B

Abstract

A reliability predicting method of a communication device is provided, and the method comprises: collecting the fail data or the reworking data of the communication device on which failure appears; determining the failure rate of the communication device according to the fail data or the reworking data. The factors influencing the failure rate of the communication device on which failure appears comprise the production factor, the hardware failure rate and/or the software failure rate.

Description

Reliability prediction method for communication equipment

The present invention relates to a reliability prediction method, and in particular to a reliability estimation method for a communication device.

Background technique

Reliability design is an important part of communication product design, and reliability is expected to be an essential foundation in reliability design. The commonly used prediction methods are mainly failure rate accumulation method, similar equipment method, similar circuit method and so on.

The failure rate accumulation method is based on the reliability data of components to predict the reliability index of the system. Further, it can be further divided into a counting method and a stress method. The failure rate data of components is mainly based on GJB/Z299C (National Military Standard: Electronic Equipment Reliability Prediction Manual) and MIL-HDBK-217 (US Military Standard: Electronic Equipment Reliability Prediction Manual).

The similar device method uses the reliability level of similar devices to predict the failure rate and mean time between failures (MTBF) of the device. The similar circuit method uses the reliability data of similar circuits to infer the reliability level of a similar functional unit/single board design. This method requires field failure rate data or reliability test data according to the type of circuit.

In recent years, with the development of science and technology, new products and new components have emerged one after another. The acquisition time of the component failure rate data lags behind the design requirement time, making it impossible to predict the reliability with the failure rate accumulation method. The software failure caused by the existence of software devices such as microprocessors and DSPs (digital signal processors) causes device failures, which causes the board or device failure rate of the circuit to be substantially the same.

Summary of the invention

The technical problem to be solved by the present invention is to provide a method for predicting the reliability of a communication device, which can accurately and objectively evaluate the market expectation of the product.

In order to solve the above problems, the present invention provides a method for predicting reliability of a communication device, including Includes:

Collecting failure data or repair data of a failed communication device;

The failure rate of the designed communication device is predicted based on the failure data or the repair data.

Further, factors affecting the failure rate include: production factors, hardware failure rates, and/or software failure rates.

KS

Further, the hardware failure rate I = F ₀ + ∑ ^ N, . , wherein the constant term Fo is related to the production factor, M is the number of components of the i-th component Γ ¹ is the failure factor, and S is the total failure factor Number.

Further, when a component failure occurs due to software reasons, the software failure rate is determined according to a software failure factor, a software complexity factor, and a software maturity factor.

Further, the software failure rate is a product of a software failure factor, a software complexity factor, and a software maturity factor.

Further, the software maturity factor is related to demand stability, software reusability, and software commerciality; the software complexity factor is determined according to a code size.

Further, the communication device is a unit/board, or any component in which components are mounted. Further, the failure rate of the communication device is the sum of the failure rates of the units/boards constituting it. Compared with the existing similar failure rate accumulation reliability prediction method, the prediction method provided by the present invention has the following technical effects:

First, the present invention can directly calculate the failure rate, MTBF and repair rate of the veneer.

Second, the present invention takes into account the software, hardware, and accidental factors of the communication device, and can accurately and objectively evaluate the market expectations of the product.

Third, the present invention collects a large amount of software failure data by induction, and finds that the law is applied, and provides a feasible software reliability evaluation scheme, and can also directly predict the reliability of the software.

Fourth, the present invention can be made into software, and automatically read into the BOM of the corresponding unit/board, and input the corresponding parameters to conveniently complete the expected work.

Fifth, the present invention provides a practical method for collecting field component failure rate through rework data. Law. BRIEF abstract

1 is a flow chart of processing data for failure rate in the present invention;

2 is a flowchart of determining a software complexity factor C _s , a software maturity factor M _{s ,} and a software failure factor F _s in the present invention;

Figure 3 is a flow chart of the estimated unit/board failure rate in the present invention.

Preferred embodiment of the invention

The invention utilizes the failure data and the repair data of the communication equipment in the network operation to predict the reliability of the reworkable unit/board of the communication product, and utilizes the BOM of the unit/board (the Bill of Material, which is a product structure recognizable by the computer) The data file) form quickly predicts the corresponding reliability indicator MTBF. It is characterized by a combination of the components used in the reworkable unit/board and the software effects attached to the unit/board, as well as other human and non-controllable factors.

First, the method of collecting the invalid data will be described in detail.

The management of components generally uses code management, and the major categories, subclasses, and subclasses of component categories can be separated from the code.

1.1 For the reworkable unit/board, determine the failure rate of the rework unit/board according to the rework data corresponding to its code.

1.2 For the components identified by the component code classification, the repair data of the corresponding components is determined according to the component code, and the failure rate of the corresponding components is determined according to the repair data. First, according to the large class calculation, if the number of components used is enough to calculate the failure rate in a small class according to the given accuracy, the failure rate is calculated in a small class, and even the sub-class and the code level (specific model) can be determined. Failure Rate.

According to the coding rule, the material code can distinguish whether it is a single board/unit or a component. In the present invention, in order to distinguish the reworkable unit/board and components, the code is mentioned when referring to the reworkable unit/board. The code is used for components.

1.3 Failures caused by uncontrollable factors such as non-components and non-software reasons (accidental factors) are classified as "other" categories. 1.4 For the reworkable unit/board with software, the failure rate caused by the software problem is counted. Next, some relevant parameters affecting the failure rate are described in detail.

The failure rate of the unit/single proposed by this method /1 consists of the following three factors:

The first factor is the production factor, which is recorded as a constant term F. (or /l o ) , F. It is only related to factors such as process, operation level, management, etc., and is the same for all units/boards.

The second factor is hardware failure rate, which is related to the device used. The hardware failure rate is equal to !; where M is the number of components used in the i-th class, the failure factor is actually the field failure rate of the 13th component, and S is the failure. The total number of factors. The sum of the first and second factors is called hardware failure rate:

KS

/1 1=F ₀ + .

The third factor is software failure rate, related to the loaded software, software failure rate /1 2 =

F _S ' C _S ' M _S , where F _s is the software failure factor, C _s is the software complexity factor, which can be determined according to the code size, is the software maturity factor, can be based on demand stability, software reusability and software commercial The degree is determined. If the unit/board does not have software, the software failure rate factor does not exist.

The unit/board MTBF indicator can be predicted by the above three factors, for example, the unit/single board failure rate = hardware failure rate 1 + software failure rate 2, and the reciprocal of the failure rate is the unit/single board MTBF indicator value.

The present invention will be further described in detail below by way of specific examples in conjunction with the accompanying drawings.

As shown in FIG. 1 , which is a flowchart of the failure rate data processing in the present invention, the process of extracting component failure rate of the present invention is as follows:

Step 101: Determine the product type, delivery period, and running time. If the GSM product is determined, the delivery period is from January 2005 to December 2006, and the operation period is from January 2007 to December 2007.

Step 102: Extract the GSM product from the repair database. In 2007, all the repair data of the equipment shipped in the previous two years, and classify the failed components.

Step 103: Extract the shipment data of the GSM product from January 2005 to December 2006 from the delivery database, and extract the online operation data of the component from the shipment data in combination with the BOM.

Step 104: Calculate the failure rate of various components:

Where ι3⁄4 is the total number of online operations of the i-th device in the field application during the time period T examined, 13⁄4 is the class examined The number of failures of components running online during the T period.

For example, for class j devices, the number of failures is 16 in a year, and the corresponding number of online runs is 690,584, then the field failure rate is ^; 16/( 690584x8760 ) =2.6448 l0" ⁹ /h, where T = 8760 It is the number of hours of the year.

As illustrated, the flowchart of C _{_s,} M _s F _s and 2 of the present invention is determined, the following detailed description of the invention calculated C _s, the workflow of F _s and M _s.

Step 201: Extract the failure data of the repairable unit/board failure due to software reasons from the repair data. For example, 60 different boards were analyzed by software for failures caused by software.

Step 202: Count the software scales of the various reworkable units/boards processed and classify them by statistical methods.

For example, the software code size of the 60 boards is counted, and the distribution map is drawn according to the code quantity, and the quarter and third quarters are taken as simple, general and complex demarcation points, and the integer is determined. Less than 200,000 lines of code is simple, 200,000 lines and 600,000 lines of code are general, and more than 600,000 lines of code are complicated.

Step 203: Calculate the average value of the failure rate caused by the software cause of each type of reworkable unit/board according to the classification determined by 202. For example, the software failure rates of the above 60 types of boards are classified according to code complexity (simple, general and complex), and the average values of failure rates caused by software causes are calculated.

Step 204: Compare the average value of the failure rate caused by the software for each statistical classification, and calculate the software complexity factor Cs.

Example: If the average of the failure rates of the simple, general and complex three types is calculated to be 1:3:5, then we will take the complexity factor as 1, 3, and 5.

Step 205: For the same complexity, classify the failure rate by maturity, and calculate the maturity factor.

Ms.

For example, the maturity factor Ms can be considered from the three aspects of demand stability Ml, software reusability M2 and software commerciality M3. From the actual situation, the demarcation point is determined and discretized, and the combination of three factors of Ml, M2 and M3 is used to invalidate the software. Perform classification statistics and compare the average values to obtain the maturity factor Ms. For example, the demand stability M1 can be divided according to the rate of change of demand. The Ml with a demand change rate of less than 10% is taken as 1, the demand change rate is greater than or equal to 10%, and the Ml of less than 30% is taken as 2, and the demand change rate is greater than 30%. Ml takes 3.

Software reusability M2, can be based on the software reuse rate of 30% or more to 70% or less, Ml takes 2; reuse rate is greater than 70%, Ml takes 1; reuse rate is less than 30%, Ml takes 3.

The commercial level of software M3 can be less than one year, one year to two years, and M3 is equal to 3, 2, 1 for more than two years.

Ml, M2, and M3 are divided into M1, M2, and M3 according to the values, and the average values of each combination are calculated separately, and the values corresponding to Ms are obtained by comparison and rounding. Example: GSM product Ms <U port under ^ J; mouth:

Step 206: Combine the failure rate, the comprehensive complexity and the maturity, and obtain the software failure factor Fs, that is, the software failure rate when Cs=l and Ms=l.

As shown in FIG. 3, which is a flowchart of the estimated unit/board failure rate in the present invention, the steps of predicting the failure rate are as follows:

Step 301: User input unit/board BOM name, software mark If the unit/board with software, enter = 1 otherwise enter = 0.

Step 302: Calculate the number of various components according to the BOM.

KS

Step 303: The hardware failure rate is expected to be ⁷ = F. +∑ K _t N _t .

i=l

Step 304: Determine whether the fault of the component is related to the software; if no software fault is involved (the case of S _T =0), then go to step 308, otherwise go to step 305.

Step 305: Input software code complexity level C, demand stability level M1, software reutilization level M2, and software commercial level M3, C, Ml, M2, and M3 may have corresponding codes. When inputting, input corresponding code. .

Step 306: The complexity level corresponding to the input code, the requirement stability level, and the software reuse The degree of grading and the degree of commercialization of the software are converted into a software complexity factor Cs and a maturity factor Ms. Step 307: The software failure rate Fs*Cs*Ms is estimated, and the hardware failure rate calculated in step 303 is added to obtain the failure rate of the unit/board.

Step 308: Output the calculation result (the failure rate of the unit/board).

The above is a detailed description of the present invention in connection with a specific actual product, and the specific implementation of the present invention is not limited to the description. It is to be understood that those skilled in the art to which the invention pertains may make a number of simple derivations or substitutions without departing from the spirit and scope of the invention.

Industrial applicability

The invention utilizes the failure data and the repair data of the communication equipment in the network operation to predict the reliability of the reworkable unit/board of the communication product, and uses the BOM form of the unit/board to quickly predict the corresponding reliability index MTBF, comprehensively Consideration is given to the components used in the reworkable unit/board and the software effects attached to the unit/board, as well as other factors such as human and non-controllable.

Claims

Claim

A method for predicting reliability of a communication device, the method comprising:

Collecting failure data or repair data of a failed communication device;

Determining the failure of the failed communication device according to the failure data or the repair data

2. The method for predicting reliability of a communication device according to claim 1, wherein

- production factors, hardware failure rates and/or software failure rates _t

3. The method for predicting reliability of a communication device according to claim 2, wherein

KS

The hardware failure rate /1 = F ₀ + ∑K _i N _i , wherein the constant term F _{0 is} related to the production factor, and M is the number of use of the faulty communication device ¥ i component The failure factor of the i-th component, S is the total number of failure factors.

4. The method for predicting reliability of a communication device according to claim 2, wherein

The software failure rate is determined based on the software failure factor, the software complexity factor, and the software maturity factor.

The method of predicting reliability of a communication device according to claim 4, wherein the software failure rate is a product of the software failure factor, a software complexity factor, and a software maturity factor.

The method for predicting reliability of a communication device according to claim 4 or 5, wherein the software maturity factor is related to demand stability, software reutilization, and software commercialization degree; and the software complexity factor is based on code size. determine.

7. The method for predicting reliability of a communication device according to claim 2, wherein said occurrence occurs

8. The method of predicting reliability of a communication device according to claim 2, wherein said occurrence occurs The communication device of the barrier is a unit/single board, or any component that is manufactured according to a determined production process, has components installed, and has certain functions.

9. The method of predicting reliability of a communication device according to claim 8, wherein the failure rate of the communication device in which the failure occurs is the sum of the failure rates of the unit/board in which the failure occurs.