7. 資料收集與失效分類

hlperng

7 資料蒐集與失效分類 (Data Collection and Failure Classification)

Depending on the type of reliability testing being performed, either the relevant test item failures should be observed throughout the test or the number of relevant failures observed at the termination of the test.  The monitoring of test item performance and the definitions of failures and relevant test time should therefore be clearly stated in the detailed reliability test plan.  IEC 60300-3-2 and the following subclauses give rules and guidance on this.  
根據所執行的可靠度試驗種類不同，決定應該要全程觀察試驗物品相關失效或是在試驗結束時觀察相關失效數量，所以在可靠度試驗計畫中應該詳細說明所需監視的試驗物品性能、失效的定義與相關試驗時間，IEC 60300-3-2及以下各節細目對此一議題提供了一些規則與指引。

The data to be collected should include installation details, where these might affect the repaired item reliability.  The reports should also describe in detail the operating and environmental conditions during the test period.  
蒐集的資料應該包括可能會影響到已修復物品可靠度的安裝細節，報告也應該詳細描述試驗過程中的操作和環境條件。

7.1 試驗物品性能監視 (Monitoring of Test Item Performance)

The following steps for monitoring test item performance should be clearly stated in the detailed reliability test plan.  
詳細可靠度試驗計畫中應該清楚地說明下列試件性能監視步驟。

7.1.1 功能參數 (Functional Parameters)

The functional test item parameters to be monitored during the test should be specified.  Limits for such parameters are often a criterion of failure (see 7.2).  
應該規定試驗過程要監視的試驗物品功能參數，這些參數的界限值往往就是失效的判定準則（參閱第7.2節）。

It may be required that all or only some of the specified parameters need to be monitored additionally.  In some cases, it may be useful to monitor additional parameters since trends can indicate imminent failures.  
有時候可能需要額外監視所有或只有部分規定參數，在某些情況下，監視額外參數可能有助於從趨勢識別即將發生的失效。

Where redundancy is employed within the item to be tested, the monitoring of functional parameters of the redundant sub-assemblies should be considered.  
當受試物品中採用複聯設計時，應該考慮監視複聯子組件的功能參數。

7.1.2 量測方法 (Measuring Method)

The measuring method and the measuring accuracy required should be specified for each monitored parameter.  Procedures for estimating the total measuring error should be given, if relevant.  
對於每個被監視的參數都應該詳細規定其量測方法及必要的量測精度，並提供相關的程序推定所有的量測誤差。

7.1.3 監視間隔 (Monitoring Interval)
Key functional parameters should, if possible, be continuously monitored in order to record intermittent failures.  If monitoring cannot be made continuously, the time interval between monitoring and the points in the test cycle at which monitoring should be done, should be specified.  
如果可以的話，應該連續監視關鍵的功能參數，以便記錄間歇性失效。如果無法連續監視，應該詳細規定監視的時間間隔，以及試驗週期中應該執行監視的時間點。

The time interval between monitoring should be short enough to avoid substantially biasing the test results, preferably not longer than the test cycle time period.  
監視的時間間隔應該要足夠短，以避免造成大幅的試驗結果偏差，最好不要長於試驗週期的時間區間。

7.2 失效類型 (Failure Types)

For each parameter to be monitored, the limits for acceptance performance should be specified.  A failure should be considered to have occurred when any one of these limits including output data is exceeded permanently or intermittently under the specified environment.  All failures should be analysed in accordance with the following subclauses.  
針對每一項被監視的參數，都應該詳細規定其可接受的界限。當在規定的環境下，任何一項，包含輸出資料，永久性的或間歇性的超出這些界限時，應該視為已經失效，所有的失效應該根據下列各節細目說明來分析。

Failure indications caused by erroneous measurement or failures in external measuring equipment are not considered test item failures.  
因錯誤的量測或是外部量測儀器失效造成的失效現象不能視為試驗物品失效。

If more than one parameter has deviated beyond it specified limits, each of these deviations should be considered as one failure of the tested items if it has not been proved that they are due to the same failure cause (secondary failure), in which case they are combined and considered as one test item failure.  
如果有一項以上的參數偏異超過其規定界限時，若這些偏異並未證實是來自於相同的失效原因（二次失效）時，每一個偏異都應該視為是試驗物品的一次失效；若為二次失效時，則合併視為試驗物品的一個失效。

If two or more independent failure causes are present, each of the resultant failures should be considered together as one test item failure.  
若是存在兩項或是更多的獨立失效原因，每一項的失效結果應該合併視為一個試驗物品失效。

Each test item failure should be classified as a relevant or a non-relevant failure.  All test item failures that cannot be clearly classified as non-relevant failures according to 7.2.1 or to any additional rule given in the detailed reliability test plan should be considered as relevant test item failures.  This includes no fault found unless specified in the test plan.  
每一項試驗物品失效都應該歸類為相關或是非相關失效，任何試件失效無法依據第7.2.1節或其他任何訂在詳細可靠度試驗計畫中的規則，清楚地歸類為非相關失效時，則應該被視為相關試驗物品失效。除非在試驗計畫中另有規定，未發現故障視為相關失效。

All relevant test item failures observed during or at the termination of the test should be taken into account in the statistical analysis and in decision-making.  
在統計分析及決策時，所有試驗過程中或試驗結束時所觀察到的相關試驗物品失效都應該列入計算。

7.2.1 非相關失效類別 (Classes of Non-relevant Failures)

A test item failure may be regarded as a non-relevant failure only if the circumstances at the occurrence provide clear evidence for the failure to be classified according to one of the classes defined in 7.2.1.1 to 7.2.1.3  below.  The evidence should be documented and included in the test report.  
只有在失效事件發生時有明顯證據能依據下列第7.2.1.1節至第7.2.1.3節所定義的一項分類種類的試驗物品失效，才能被視為非相關性失效，這些證據應該被記錄並納入試驗報告中。

Additional classes of non-relevant failure applicable in a particular case may be defined in the detailed reliability test plan.  
其他適用於特殊情況下的非相關性失效種類可以定義於詳細可靠度試驗計畫中。

7.2.1.1 二次失效 (Secondary Failures)

Secondary failures are considered non-relevant.  The corresponding primary failure is classified relevant or non-relevant according to the test specification.  Note that a secondary failure may occur after a time delay following the occurrence of the primary failure.  The duration of the time delay should be approved by the customer and the test agency.  
二次失效被視為是非相關的，其相對應主要失效則根據試驗規格歸類為相關或非相關。要注意的是二次失效可能會在主要失效發生後延遲一段時間才發生，延遲時間應該由顧客及試驗機構核可。

7.2.1.2 誤用失效 (Misuse Failures)
Misuse failures during testing may be due to unintentional test conditions, for example severity exceeding those specified for the item under test, rough handling by test or repair personnel, etc.  Misuse failures are considered to be non-relevant unless agreed upon by the parties.  
試驗過程中的誤用失效可能是由於某些非故意的試驗條件所造成的，例如試驗條件嚴厲度超過受試物品規定的數值、試驗或維修人員的粗魯搬運等。除非經過雙方同意，否則誤用失效被視為非相關的。

NOTE.  If equipment is required to be robust against operation and handling errors, this should be considered by the parties.  
註解：如果對於設備有要求抵抗操作與搬運的錯誤時，雙方應該將此列入考慮。

7.2.1.3 藉由設計改正消除的失效 (Failure Elimination by Design Correction)

A type of failure observed early in the test may result in a design change or other remedy, implemented after the test on all items in the population.  If such a corrective action is proven to be effective, the failures of this type may, upon agreement, be reclassified as non-relevant for the test already performed.  
在試驗早期所發現的失效，可能可以引發設計變更或其他改善，在試驗後應用於群體的每一個物品。如果這樣的改正行動經證明有效，這一類型的失效可能可以經過同意後在試驗完成後重新歸類為非相關。

NOTE.  The criteria for classifying a failure as non-relevant should be specified.  
註解 : 應該規定非相關性失效的分類準則。

7.2.2  失效之特殊類別 (Special Classes of Failure)

The detailed reliability test plan may define additional special classes of test item failures, based for example on the effect upon the performance of the tested item, on the location of the failure with respect to redundant units or non-essential units in the item, or on repair costs and time.  The types of failure defined in IEC 60050(191) may be useful for this purpose.  
詳細可靠度試驗計畫可能根據一些議題定義其他特殊的失效類別，例如根據對受試物品性能的影響、試件中與複聯單機或非主要單機有關的失效位置、或維護成本及維護時間等。在IEC 60050(191)中所定義的失效類型可能有助於此項目的。

7.2.2.1 需要立即決定拒收決策之失效 (Failure Demanding Immediate Reject Decision)

In particular cases for reliability compliance tests, it may be appropriate to define those test item failures the occurrence of which, regardless of number, should result in  an immediate reject decision that overrides the normal accept/reject criteria.  In this situation, the definition of such failures should be included in the detailed reliability test plan.  
在特殊情況下的可靠度符合試驗，可適當定義試驗物品發生該失效時應該立即採取拒收決策，不管數量多少，而忽略正常的允收/拒收準則。在這種情況下，此類失效的定義應該包含在詳細可靠度試驗計畫中。

For example, test item failures which are likely to result in hazardous or unsafe conditions for persons using, maintaining or relying on the tested item, or failures which are likely to cause extensive material damage should be included in this category.  
例如，試件失效可能導致使用、維護或信賴通過試驗的試件人員危險或不安全的狀況，或是失效可能引起大規模的材料損害都應該歸為此類別中。

7.2.2.2 重複性失效 (Recurrent Failures)

Recurring failures may imply a weakness in the design of the item ro indicae the use of items of inferior quality.  Recurring failures are an important indication of possible wear-out or other degradation that results in an increasing failure rate.  
重複性失效可能意味著物品設計上有缺陷或是使用較差品質的物品，對於可能磨耗或是退化所造成失效率增加，重複性失效是一項重要的指標。

When failure recur, special investigations should be initiated in order to establish the cause, and assumptions regarding the distribution of times to failure should be re-examined (see 8.4).  
當失效重複發生時，應該啟動特別調查以確定原因，並重新檢討有關失效發生時間分布的假設（參閱第8.4節）。

7.2.2.3 複聯物品的失效 (Failures in Redundant Items)

In a redundant item the faults causing system system failure may, upon agreement between customer and manufacturer, be classified separately from those causing only a reduced level of redundancy.  The type of failure and its consequences for the item should be considered in the test plan.  The test specification should state to what extent item failures and sub-assembly failures (reduced level or redundancy) are classified as relevant failures.  
在客戶跟製造商協議後，複聯物品誤差所造成系統的失效與只影響下層的失效可以分別歸類，在試驗計畫中應該考慮到此種失效的類型及其結果。試驗規範中應該要說明什麼範圍的物品失效跟從屬組件失效（低層的複聯層級）歸類為相關性失效。

NOTE.  The same type of failure may cause system failure or reduced redundancy depending on the condition of the item.  
註解：同樣類型的失效可能會造成系統失效或下層複聯物品失效取決於試件的狀況。

7.3 失效分析 (Failure Analysis)

Each failure should be analysed at least to the extent that the failure can be classified clearly as relevant or non-relevant.  
每個失效應該至少要清楚的分析到可以歸類於相關或是非相關的範圍。

The failure analysis should, as far as possible, identify the primary failure cause and also the other likely contributing factors.
失效分析應該盡可能地定義出主要失效原因以及其他可能的要素。

The purpose of this more thorough failure analysis is to form a basis for corrective action, for example, design changes and changes in production processes.  
這樣完善的失效分析目的在於建立一個改善行動的基準，例如設計變更與生產過程變更。

In order to facilitate failure analysis, all relevant information about each failure should be recorde in the test results or file, and all exchanged parts or modules retained for physical examination.  
為了有效地進行失效分析，每個失效的所有相關資訊都應該記錄在試驗結果或是檔案上，而且要保存所有更換零件或模組的以備實體檢驗。

The failure analysis should confirm that the failure modes observed in the test are relevant for the intended and expected use of the tested item in the field.  This is especially important in comparison tests and accelerated tests; for example step-stress tests.
失效分析應該確認試驗過程中發生的失效模式與試件在實際所使用時所預期的環境是相關的，尤其在比較試驗和加速試驗時特別重要，例如步進應力試驗。

Failure modes that are not relevant for the intended and expected use of the tested items in the field may, upon agreement, be reclassified as non-relevant for the test concerned.  Where no fault had been found during the failure analysis, this event should be counted as a relevant failure, unless agreed otherwise between the parties.  Software failures in a combined software / hardware item may require special failure analysis techniques.  
失效模式與試件在實際所使用時所預期的環境無關時，經過協議可以從試驗的觀點重新歸類為非相關失效。失效分析找不出任何錯誤時，應該將此失效事件視為相關性失效，除非另有協議。包含軟硬體一起的物品發生軟體失效時可能需要特別的失效分析技術。

7.4 相關試驗時間 (Relevant Test Time)

The relevant test time is the time used in connection with the number of relevant test item failures to show compliance with reliability requirements or to calculate the value of the reliability performance characteristic.  
相關試驗時間是結合相關試件失效數量的相對應時間，用來證明符合可靠度需求或是計算可靠度性能特性的數值。

The relevant test time to be recorded during testing may be the individual relevant time for each test item or the relevant time accumulated by all test items as specified by the detailed reliability test plan.  Any reliability stress screening, maintenance time and down time of the test items should be excluded.  When monitoring at intervals, any failure should be considered to have occurred at the middle of the time interval concerned.  
按照詳細的試驗計畫中的規定，在試驗中所要記錄的相關試驗時間可以是每個試件個別的相關時間或是所有試件累積的相關時間。所有試件的可靠度應力篩選、維護和停機時間都應該排除。當區間監視時，任一失效應該視為在監視時區中點發生。

The detailed reliability test plan may prescribe a minimum and/or a maximum relevant test time on each individual test item.  
詳細可靠度試驗計畫中可以規定個別試件的最小和/或最大的相關試驗時間。

For test items consisting of two or more separate sub-items, the relevant test time recorded for the test item should taken into account the relevant test time on any of its sub-assemblies.  
對於包括兩個或是更多單獨子項的試件，其相關試驗時間紀錄應該考慮每一個從屬組件的相關試驗時間。

The relevant test time should be defined in the detailed reliability test plan.  If the test is based on test item operating time, an elapsed time indicator or cycle counter installed in every repairable item on test is recommended.  Clock readings should be recorded at each failure observation and maintenance action.  Where this is impractical, another time basis, such as vehicle usage time (for example flight time for airborne equipment) may be used with or without the application of an appropriate corrective factor.  
相關試驗時間應該定義於詳細可靠度試驗計畫中，如果試驗是以試件的操作時間為主，試驗時建議在每個可修復的物品上安裝計時器或是周期計數器，每次的失效觀察或是維護行為都應該記錄其時間。當時間紀錄不適用時，可以使用其它的時間基礎，像是載具使用時間（例如飛行裝備的飛行時間）加上適當的修正因子或不需修正因子。