Reliability Engineering: Trends, Strategies and Best Practices

WHITE PAPER September 2007

Predictive HCL’s Predictive Engineering encompasses the complete product life-cycle process, from concept to design to prototyping/testing, all the way to manufacturing. This helps in making decisions early Engineering in the design process, perfecting the product – thereby cutting down cycle time and costs, and Think. Design. Perfect! meeting set reliability and quality standards. Reliability Engineering: Trends, Strategies and Best Practices | September 2007

TABLE OF CONTENTS Abstract 3 Importance of reliability engineering in product industry 3

Current trends in reliability engineering 4

Reliability planning – an important task 5

Strength of reliability analysis 6

Is your reliability test plan optimized? 6

Challenges to overcome 7

Outsourcing of a reliability task 7

About HCL 10

© 2007, HCL Technologies. Reproduction Prohibited. This document is protected under Copyright by the Author, all rights reserved. Reliability Engineering: Trends, Strategies and Best Practices | September 2007

Abstract

In reliability engineering, product industries now follow a conscious and planned approach to effectively address multiple issues related to product certification, failure returns, customer satisfaction, market competition and product lifecycle cost. Today, reliability professionals face new challenges posed by advanced and complex technology. Expertise and experience are necessary to provide an optimized solution meeting time and cost constraints associated with analysis and testing. In the changed scenario, the reliability approach has also become more objective and result oriented. This is well supported by analysis software. This paper discusses all associated issues including outsourcing of reliability tasks to a professional service provider as an alternate cost-effective option.

Importance of reliability engineering in product industry

The degree of importance given to the reliability of products varies depending on their criticality. For instance, failure of a critical unit in a flight control system may cause loss of human life and property, whereas an unreliable ignition system in a car may lead to customer dissatisfaction. Reliability has a direct or indirect impact on some aspects like profitability due to poor market feedback and reduced sales. The respective product manufacturers have a common concern about reliability but for different reasons.

Regulatory requirements warrant compliance to applicable Reliability and Safety (R&S) standards. In the aerospace and life sciences domain, inadequate product reliability may lead to loss of safety critical functions of an aircraft or a life-supporting system, respectively. Reliability engineering practice during the product development program is mandatory. Manufacturers strictly prioritize associated tasks during the development program in accordance with R&S requirements.

On the contrary, reliability practice in automotive, semiconductor and hi-tech domains is driven by product functional requirements, business profitability, market competition and customer satisfaction.

The need for products to conform to higher performance limits has increased the complexity and poses a challenge to reliability practitioners. One recent unfortunate example is of space shuttles Challenger and Columbia with the state-of-the-art technology that met the required performance limits. The fatal accidents of these space shuttles after 9 and 27 respective successful missions raise a serious question on R&S of these products.

Current trends in reliability engineering

Does the accomplishment of reliability tasks assure definite improvement in product reliability? No; in case the selected tasks are not effective and efficient, the effort may not yield the expected result. We need to learn from our past experience and have a focused approach in line with current trends. The world reliability community believes in sharing best practice experience so as to provide reliable products to all customers. The following are a brief on the current trends:

Is your reliability assurance dependent only on numeric values?

The reliability prediction of a medical device may show compliance to 99% target reliability. However, current reliability practice additionally recommends applicable qualitative failure analyses such as hazard analysis and Failure Modes Effect Analysis (FMEA) to analyze all associated failure conditions and modes. The objective is to improve reliability addressing design weaknesses. Reliability prediction, allocation and FTA are the common tools used to address specific quantitative requirements. The current reliability practice does emphasize qualitative analysis without ignoring the importance of quantitative analysis.

Are you tracking preferred practice and regulatory changes?

In the aerospace domain, FAR/ JAR 25.1309 and ARP 4761 have brought clarity and minimized subjectiveness in the approach. Risk analysis of medical devices is one of the most important tasks for conformance to regulatory requirements. Design control and validation requirements emphasize need for risk analysis vide 21 CFR section 820.30, and relevant guidelines are provided in ISO 14971. The structured approach with design for reliability (DFR) is now being practiced during product development in other non-aerospace domains. Hazard analysis, FMEA, Fault Tree Analysis (FTA) and reliability prediction are invariably part of the DFR activity.

Looking for an alternate reliability life test?

The current product lifecycle is short, and hence conventional long-duration reliability tests have been phased out. Short-duration test approaches such as Accelerated Life Test (ALT) and Highly Accelerated Life Test (HALT) are now preferred options for reliability validation.

Can analysis be a substitute for a reliability test?

This can be explained by drawing an analogy, as follows. A sales representative may provide details on technological features of a mobile handset or a car, but customers always prefer to have a performance demonstration test in spite of their prior knowledge and confidence on the product technology. Similarly, a reliability validation test cannot be replaced by an analytical assessment; however, it can certainly minimize the dependence on testing. Late availability of information on product reliability with a requirement of a sizeable budget is the common concern against reliability tests. A rational approach with analysis and/ or testing may be judiciously planned. A reliability test on a product, which is not safety critical, may be waived if supportive design study, component reliability data, engineering stress analysis, similarity analysis and product design qualification test data are available.

How to minimize the uncertainty of a quantitative prediction?

A great deal of effort has been put into the preparation of a common failure database that significantly supports the reliability prediction task. NPRD95, MIL-HDBK-217FN2, FMD97 and NPRD3 are different failure databases in the public domain. Manufacturers need to create and maintain failure databases of their own products. Accuracy of analysis will be better with the use of product-specific failure data. Product repair and defect investigation data need to be preserved and maintained for creating a product-specific failure database over a period of time.

Do you accept failure as a random occurrence?

Experience has taught manufacturers not to treat failures as random occurrences. Now they try to understand failure mechanism with a physics- of-failure approach which consequently may contribute in reduction of constant failure rate and extension of useful life. Advanced technology for failure investigation is available and helps identify the root cause and provides options to fix it.

Reliability planning – an important task

Reliability analysis findings are normally reviewed during an important phase of product design and development. It becomes an uphill task to complete the analysis in a short time unless continuous effort is made throughout the development program as per the schedule detailed in the reliability plan. Inadequate planning may result in an incoherent approach. This may lead to either omission of or delay in carrying out important analyses in the development program, thus leaving limited options for design improvement. DFR underlines the importance of a reliability plan.

Do not use a common template for a reliability plan

Every product design and its development program are unique. Meticulous planning is required to map the reliability tasks to the identified program objectives and constraints.

What should a reliability plan essentially contain?

A reliability plan is a framework for achieving program objectives for product reliability which include identification and scheduling of reliability tasks traceable to requirements, approach for applicable tasks (analysis and testing), responsibility matrix, resource optimization and planning and guidelines on preferred practice during design and development.

Reliability tasks are carried out along with the design development activity and completed at an appropriate phase. The schedule of the identified tasks should be included in the plan.

The plan should capture the reliability priorities, program objectives and constraints. This input should be used for planning and optimization of effort. The strategy and the framework of the identified tasks related to reliability analysis and testing should be detailed unambiguously.

Strength of reliability analysis

You may rely on reliability analysis, select an appropriate one!!

Total dependence on reliability test is an approach associated with risk. There are powerful analytical tools capable of providing early assurance on design reliability, viz. reliability apportionment, prediction, part derating analysis, FMEA/ Criticality Analysis (CA), Functional Hazard Analysis (FHA), Software Hazard Analysis, FTA, Common Mode Analysis (CMA), Zonal Safety Analysis (ZSA), Cascading Failure Analysis (CFA), warranty analysis and reliability estimation from field failure data of similar products.

A product with complex multidisciplinary integrated functions requires specialized analytical tools. These tools are required to be used to assess built-in safety and reliability features of the product under an applicable environment. Some of these analytical tools are qualitative and very effective in weeding out design weaknesses. Each analysis has got its strength and weaknesses. For instance, FTA can be used for analyzing critical failures of a complex system; however, when complexity is not high, FTA may not be required. Some regulatory guidelines provide insight into the selection of applicable analysis and help in the appropriate selection of analytical tools that enhance the effectiveness of reliability analysis.

Software for quantitative analysis

The use of dedicated reliability software minimizes the possibility of error in quantitative analysis. Different software such as Relex, Reliasoft, Cafta, Prism, Item and Isograph has independent modules which can be used to carry out particular analysis. Selection of the software should be done with due care after assessing its features, cost and feedback from other users.

Is your reliability test plan optimized?

Our experience on project support to different customers in the life sciences domain is that a reliability test plan invariably needs optimization to attain

target reliability with a specified confidence limit against constraints such as schedule, budget and sample size. Mathematical approach objectively helps in the DFR test planning.

There are tests of different types such as HALT, ALT and reliability growth test that need to be appropriately selected for a specific product requirement.

Challenges to overcome

The challenges to be overcome by the reliability team are:

Technological challenge

A technologically advanced and highly integrated system with hardware and software performing complex functions certainly poses a stiff challenge to reliability engineers.

We find products or systems in life sciences, semiconductor, automotive and hi-tech performing important functions with real-time data analysis, monitoring and control. Administration of an excess dose of a drug during surgery due to failure of some unit of an automated sedation delivery system may prove to be fatal. Similarly, poor engine response of a racing car may cause annoyance due to a fault in the controller-based fuel injection system. Analysis of all probable causes needs sound understanding of the system.

In aerospace, flight control and actuation, avionics, landing gear, propulsion, wheel and brake, and nose wheel steering systems have complex functionality and offer a stiff challenge during FTA and FMEA which could be overcome with experience and proper application of concepts.

Program schedule

An analysis task is a time-bound activity which is further constrained by the limited or late availability of information in a typical development scenario. All development programs have stringent schedules. There is immense pressure when there is a backlog or delayed start of reliability analysis. In our experience more than 70% of the projects are highly constrained by the stringent delivery timeline.

Resource planning

A major effort is required to prepare analysis reports in compliance to regulatory requirements at an appropriate phase of projects and reviews. Deployment of skilled and experienced resources ensuring availability of software needs prior planning. Ramp-up of the project team size and optimized use of software are the typical challenges related to the resource.

Outsourcing of a reliability task

Is outsourcing cost effective?

(i) Availability of skilled resource and management

In the product industry, forming a dedicated team of engineers having the specialized skill set and experience in reliability is difficult. Also the requirement of ramp-up of team size to meet the program and certification deadlines makes it more difficult to manage with the available resource due to design and development delays or concurrent planned activities.

(ii) Investment on software

Investment on reliability analysis software is a costly affair. Sometimes use of particular software is recommended by an aircraft manufacturer/ customer for data compatibility. This is to enable the integration of analysis into a common database. Non-availability of the necessary software may lead to additional cost.

To summarize, outsourcing of reliability tasks to professional groups having domain experience is a cost-effective measure for product manufacturers.

On the other hand, service providers are required to address the abovementioned issues related to skilled resource and software investment. They are also required to have good infrastructure including secure and efficient communication links for analysis data transfer. The success of time- bound projects depends on availability of infrastructural support.

Who is the preferred reliability service provider?

One who meets the following minimum expectations of a product manufacturer qualifies as the preferred service provider:

■ Provides best optimized solution meeting development program schedules. The optimized solution includes analysis and reliability test tasks

■ Possesses an adequate and up-to-date knowledge base on changing trends and regulatory requirements to recommend and to follow the best practice methods

■ Capable of carrying out analysis with proper understanding of complex system functionality with multiple interfaces

■ Prepares a reliability plan, recommends and carries out appropriate reliability analysis and tests during the design and development phase

■ Generates analysis reports for the product certification and meets the expected quality level

■ Meets the delivery timeline

■ Has the required resource base to support reliability tasks

The following are the four important criteria for the selection of a qualified service provider:

(i) Reliability domain expertise

It is important that the service provider has a perfect blend of qualification, experience and skill set as their resource base. The skill set includes understanding of fundamental concepts of reliability, analytical tools and use of dedicated software.

(ii) Relevant experience

Domain-specific experience is desirable particularly for aerospace and life sciences to address R&S issues during analysis aligned with current practice in respective domains. In general, the service provider is expected to have an in-depth understanding and application of reliability concepts in varied domains, which is possible only if they have worked on a variety of projects of higher complexity and challenges.

(iii) Understanding of regulatory requirements

Again in aerospace and life sciences domains, relevant R&S requirements need to be addressed in compliance to regulatory requirements. Therefore, understanding of these requirements is expected from the service provider.

(iv) Resource base and infrastructure

It is very important that the service provider has a sound resource base and infrastructure. The resource base includes experienced and skilled manpower with analysis software. The ability to ramp up the team according to a project need is one of the important considerations. The required infrastructural support for the success of a project includes availability of systems, efficient communication links, data confidentiality, backup and data recovery provisions.

PACKAGE IMPLEMENTATION

INFRASTRUCTURE MANAGEMENT

BPO

R&D

TECHNOLOGY ABOUT HCL

About HCL Technologies

HCL Technologies is one of India’s leading global IT Services companies, providing software-led IT solutions, remote infrastructure management services and BPO. Having made a foray into the global IT landscape in 1999 after its IPO, HCL Technologies focuses on Transformational Outsourcing, working with clients in areas that impact and re-define the core of their business. The company leverages an its extensive global offshore infrastructure and global network of offices in 17 countries to deliver solutions across select verticals including Financial Services, Retail & Consumer, Life Sciences & Healthcare, Hi-Tech & Manufacturing, Telecom, and Media & Entertainment (M&E). For the quarter ended 31st March 2007, HCL Technologies, along with its subsidiaries, had last twelve months (LTM) revenue of US$ 1.27 billion and employed 40,149 professionals. For more information, please visit www.hcltech.com.

About HCL Enterprise

HCL Enterprise is a leading Global Technology and IT enterprise that comprises two companies listed in India – HCL Technologies and HCL Infosystems. The three-decade-old enterprise, founded in 1976, is one of India’s original IT garage start-ups. Its range of offerings spans Product Engineering, Custom and Package Applications, BPO, IT Infrastructure Services, IT Hardware, Systems Integration, and distribution of ICT products. The HCL team comprises approximately 45,000 professionals of diverse nationalities, who operate from 17 countries including 360 points of presence in India. HCL has global partnerships with several leading Fortune 1000 firms, including leading IT and Technology firms. For more information, please visit www.hcl.in.