1
7.5 Pitfalls of Engineering Change
7.5
Pitfalls of Engineering ChangeChange Practice duringComplex Product Design
Timothy Jarratt, Claudia Eckert and P. John Clarkson[1]
Abstract
The majority of design projects involve adapting a known solution to meet new requirements. Therefore, understanding the issue of engineering change is of vital importance if companies are to deliver product development projects on time and to budget. When a change is made to part of a product, the change is likely to propagate to affect other components or systems. This paper examines the engineering change process within a UK engineering firm and focuses on the issue of change propagation. The findings are compared with an earlier study in the aerospace industry. Four reasons why propagation occurs are proposed and discussed.
Keywords:Design, Management, Engineering change
7.5.1.Introduction
From a business perspective, changes to a design are “a fact of life” in taking a product from concept, through design, manufacture and out into the field [1]; they are the rule and not the exception in product development processes in all companies and in all countries [2]. As an example of the importance of engineering change, a survey of German engineering businesses found that approximately 30% of all work effort was due to engineering changes [3]; this included rework as well as the adding of functionality to a product. It has been reported that engineering changes in the automotive industry consume between a third and a half of the engineering capacity and account for 20-50% of tool costs [4].
Engineering Change Research
Relatively little work has been published on engineering change and engineering change management. Wright [5] conducted a survey of engineering change management literature published between 1980 and 1995 and found only 15 “core” papers. Other researchers have also commented upon this scarcity (e.g. [6]).
Historically change has been seen by the design community as the responsibility of manufacturing research groups [5] and any alterations made to a product during design were just regarded as normal iterations of the design process. However, the rise of concepts such as concurrent engineering and simultaneous design, plus the influence of business disciplines such as configuration management, has seen production and organizational issues becoming an integral part of the design process and associated research. Thus, engineering change is starting to be featured much more prominently in academic work.
Background and Motivation
The work described in this paper is part of an ongoing research project into the field of engineering change and the engineering change process. An earlier, general study into change processes at an aerospace company (reported in [7]) indicated that potential change propagation was a major pitfall of engineering change management. Propagation occurs when an alteration to a component or system spreads to affect other parts of the product. A few authors (e.g. [3] and [4]) have mentioned propagation as a possible effect of implementing a change. There have been studies into change practice in UK [6], Hong Kong [8] and Swedish [9] companies. However, there has never been a study that specifically investigates the issue of change propagation within a company.
The study described in this paper was carried out at a large UK engineering company and was complemented by many informal interviews with engineers from a wide range of other companies. Underpinning this research is the hypothesis that changes potentially can propagate between the elements of a product when an alteration is made to a particular component or system. The purpose of this work was to prove this main hypothesis and to discover how engineers assess the possibility of change propagating, which tools or methods (if any) they use to support their evaluation and what sort of support is required.
7.5.2.The Engineering Change Process
The formal engineering change process is a critical business process that affects all aspects of product design and development. An engineering change process can be triggered at any point in the product life-cycle once the concept has been selected, because, once the concept has been chosen, information about design decisions starts to be formally released to design teams, suppliers, potential customers, etc. Any changes to this information as the product evolves must be regarded as an engineering change process.
Perhaps the clearest description of the engineering change process is provided by Leech and Turner [10], who state that the process is a mini, highly constrained design process or project and “like any project, is only worth undertaking if its value is greater than its cost”.
Figure 7.5.1. A generic engineering change process
Figure 7.5.1 shows a generic high-level engineering change process based upon various processes outlined in literature (e.g. [11] [12] [13]). The process is initiated by a change trigger. Eckert et. al. [7] describes changes as emerging from the product (i.e. errors) or being initiated from outside (i.e. customer requests, legislation, etc.).
The highest risk of the six phases is the third: assessing the impact or risk of implementing each solution. Various factors must be considered: for example the impact upon design and production schedules; how relationships with suppliers will be affected; and will a budget overrun occur. The further through the design process a change is implemented, the more disruption is caused. Several authors refer to a “rule of 10” (e.g. [2]), which states that the cost of making an engineering change rises by a factor of 10 between each phase. Thus, a change made during manufacture would be 1000 times more expensive than making the same change during the detail design phase.
There are possible iterations within the process, two of which are marked by arrows in Figure 7.5.1. For example a particular solution may be too risky for the company to implement and so the process will return to phase 2, so that other possible solutions can be identified. At the approval stage, the Engineering Change Board may feel that more risk analysis is required (maybe in the form of more testing) and so the process will return to phase 3. There are other possible iterative loops, but they are not marked for sake of clarity. The most extreme loop would be when, if during the review phase, it was realized that the implemented solution had been ineffectual or made matters worse. In that instance the process would return to the start with a new change request being raised.
7.5.3.Details of the Study
This section briefly describes the company involved and outlines the elements of the study.
The Company
The company designs and manufactures diesel and gas engines for two main markets: generator sets and off highway. There is a wide spectrum of customers ranging from famous global businesses to tiny privately owned companies. Large concerns take delivery of many thousands of engines per year, whilst small manufacturers of specialist vehicles may require less than 50. One basic engine can go into a wide variety of applications and be used in a range of environments. Thus, adapting engines to fit customers’ needs is a standard activity for the company.
A major issue facing the business is complying with ever tightening environmental legislation, which has caused a significant decrease in product life-cycles and led to huge technological changes. Various aspects of engine performance are being regulated; the most important area is that of exhaust emissions. There is not room for a thorough discussion of this issue; further details are provided by Jarratt et. al. [14]. New legislation requires the development of a new generation of engines approximately every five years.
Methodology of the Study
The study consisted of four parts: (1) interviews with a range of employees across the company; (2) observations of meetings; (3) shadowing an engineer involved in managing the engineering change process; and (4) structured sessions with engineers filling in a connectivity matrix (see [15] for more details of this last aspect).
Twenty engineers and managers were interviewed for between 45 minutes and two hours in early 2002. The interviewees came from a wide range of roles and functions, and had a range of experience within the company and industry. Just under half of the interviewees had been employed by the company for over 20 years. Each session was recorded on audiotape and was later transcribed. The interviews were semi-structured. After initial questions about the background of each person the interview progressed to focus more on design processes and in particular the engineering change process. During the discussion of engineering change, the interviewees were asked to:
- describe the company's engineering change process and their relationship to it;
- give examples of changes that had been successfully implemented and examples where there had been unexpected difficulties;
- identify those tools and techniques (if any) that were used to support engineering change;
- discuss where the engineering change process could be improved and what support methods / tools they felt were needed to assist them.
Three Design Change Meetings were observed during the interview period to gain an understanding of the issues and trade-offs that are considered when assessing and authorizing engineering changes. Attendance at these meetings also enabled the authors to appreciate the tools and methods used during the risk analysis assessment of each change. The Design Change Meetings are held twice a week and the purpose is to review and authorize engineering changes to the company's product range. A wide variety of people from across the business (manufacturing, after sales support, etc.) attended in conjunction with designers and engineers.
The first author spent a week shadowing the Technical Design Manager. This employee had a number of roles, one of which was being the “owner” of the Design Change Process. Shadowing this person enabled the author to gain a clear understanding of how the engineering change process worked in reality and it helped highlight which areas of the process required support.
7.5.4.Findings
The company is very successful and on the whole is well organized. A key issue that came out of the study was that all of the company's design and development activities are dominated by engineering change in some form or other. This includes the New Product Introduction (NPI) process, which is used to develop engines to meet new tiers of legislation. When a new engine is required and an NPI process is launched, one of the first acts is to take the current engine and decide which pieces of the architecture and technology can be carried forward into the next generation and which must be altered to meet the new requirements. Diesel engines are evolutionary products with a well-established architecture. However, this is not to say that there is no room for innovation. Over the past decade rapid advances have been made in materials technology and manufacturing processes. Perhaps the biggest area of innovation has been the addition of electronics to engines.
Reasons for Triggering the Change Process
At a high level there are three sources of engineering change at the company: (1) suppliers, (2) customers and (3) internal departments.
Suppliers: either a current supplier is changing a process (e.g. manufacturing) or a new supplier is being brought in. These relationships are vitally important as a large amount of the engine is bought in (70-80%).
Customers: the reasons are either to make a change to an existing sales option or to request a new sales option. For example, a customer that designed vineyard equipment requested the development of a new filter head option. Vineyard tractors need to be very narrow and the standard filter head impeded the turning circle of the vehicle.
Internal departments: these changes can be for a wide variety of reasons such as quality and reliability improvements, cost savings, changes to manufacturing processes, improve servicing or a company-wide initiative.
Effects / Impacts of Change
During the interviews it became clear that most of the changes carried out are quite mundane. For example changes are made to paperwork when there is a switch in supplier and the new source's manufacturing process is slightly different to that of the previous supplier's. However, every so often an engineering change can propagate dramatically, as the product engineering manager noted: “On each design project there will be 4-5 things that are major ... things that we did not predict at all”. Less dramatic propagation will also occur: “in a lot of cases [propagation] occurs, but only to affect one or two components”
The interviewees gave many examples of past changes to the engines that they had been party to, most of which described situations where the implementation of a change had not been completely smooth. Three of these examples are given below with quotes from the engineers who described them. All show that engineering changes can propagate to affect other parts of the product or other business processes such as after sales support.
Example1:outlet pipe change – a project was undertaken to replace the metal water outlet pipes in an engine series with plastic ones to reduce overall engine weight and lower manufacturing costs. The water outlet pipes contain a temperature sensor and this was transferred from the old design of pipe to the new one. It was only once the redesigned engine was in production that it was noticed that the sensor no longer functioned: it had been designed to earth through the pipe, which was no longer possible due to the change of material. The solution was to redesign the sensor with a return wire. “Nobody thought about it when they introduced plastic pipes. It was very embarrassing and very expensive.”
Example2:change to bush – a new bush design was suggested for the gears for the power-take-off to reduce cost. As a result, the oil circuit system was altered without any analysis work. The first engine with this change included quickly broke during testing. After much analysis, it was revealed that the change to the oil circuit was the critical factor, which had initially been overlooked. The eventual solution meant changes to four or five other components within the engine. “It was a real systems approach”.
Example3:gear train change – another key requirement (both from customers and legislation) for new engine design is to reduce engine noise. One way to achieve this focused on the backlash between the gears. The gear train was redesigned and some of the gear ratios were altered. Engineers designed a gear train that was perfectly durable, but unfortunately, to get the gears to fit within the same space that had been available before and be located at the same centres, meant that the helix angle had to be altered. These introduced extra loads in the gears and, as the bearing systems were being carried over from common parts, those bearing systems were not adequate. So in order to stop the change from spreading the gears were redesigned again: “otherwise we would have fixed the next problem down the chain and that would have had a knock on effect somewhere else”. “You can’t just […] work on your own component – how it interacts with other parts of the engine must be considered”.
Tools to support the engineering change process
During the interviews, when the subject of tools or methods was raised, the interviewees were first asked to describe how they analyzed each engineering change before describing the tools and methods they used.
Analysis of engineering changes
When an engineering change is initially evaluated, a group of engineers will meet and use their experience to determine what level, if any, of testing and analysis is required. If the form, fit or function of a component is affected, a Failure Mode and Effects Analysis (FMEA) will be carried out to identify the critical characteristics of the situation: for example whether emissions legislation compliance will be affected. This will be done at three different levels: engine, system and part. People will also consider wider risk issues such as program risk i.e. whether the change will affect the whole design program.
Several engineers talked about having a “mental checklist” that they went through with simple components. “For example with flywheels: we have almost standard quotes – the scheme work will be the same, etc. The time to design and detail the new flywheels is quite standard”. With customer requests for new options, things are not always so easy: “it comes down to experience. There is no hard and fast way of doing it, there is no checklist as such”.
There was universal agreement that a lot of problems with analyzing engineering changes come from oversights and mistakes rather than the unknown: “a lot of the problems come from stupid mistakes – not from horrible ones – the big ones people think about and apply their formidable brains to it, but the little details are overlooked”; “[paradoxically] big changes are less likely to propagate [unexpectedly] than little ones...”
Tools and methods
In terms of computer tools, the company has an Enterprise Resource Planning (ERP) system, which has a Product Change Control module that supports the workflow aspect of the change process i.e. applying a change once it has been authorized. The company also has a Product Data Management (PDM) system, which is integrated with the CAD system and used to manage all the drawings, etc. for each product. This PDM system is also used to populate the company's configurator software, which is used by sales engineers to accurately configure engines for each customer.