1.0Introduction The basic function of a turbocharger in an automotive application is to utilise expensed exhaust gases (emanating from an Internal Combustion Engine (ICE)) that would otherwise have been vented to atmosphere to turn a turbine. This turbine uses a common shaft connected to a compressor wheel which produces compressed air which is fed into an engine in order to increase engine performance (Bell 1997). In other words, the turbocharger is basically a powerful air pump which is driven by exhaust gas energy. The specific operation of the turbo is of course far more complicated. However, for the purposes of this paper the basic function of the turbocharger is enough to substantiate the following information. As stringent emission standards for vehicle manufacturers are tightened particularly throughout Europe, engine manufacturers are looking towards ways in which to meet these standards (Bayomi& Abd El-Maksoud 2012). One such method has led to the rise in popularity of the turbocharger as vehicle manufacturers have opted to downsize the engine and still keep performance up with the use of a turbo (Chiong et al. 2012). This uprising of the turbo engine has resulted in a large amount of research into turbocharger technology (particularly in the aftermarket industry) to be undertaken with large amounts of investment from companies including Garret (Honeywell), Borg Warner and Mitsubishi Heavy Industries (MHI) all of which are at the forefront of state of the art turbo technology. This report will specifically focus on the Borg Warner EFR (Engineered for Racing) series of petrol engine turbochargers designed for use in high powered race cars as well as an aftermarket enhancement for street use. The EFR series of turbocharges uses stainless steel twin-scroll turbine housing capable of handling higher temperatures with better structural stability compared to its cast steel counterparts. The twin-scroll design also allows for more efficient use of the unstable environment of exhaust pulses with which it uses to generate compression. The “hot” side of the EFR turbo also features an adjustable high-flow internal waste-gate which is used to precisely control the amount of exhaust gases flowing into the turbine housing with the remainder of diverted exhaust gases efficiently expelled back into the exhaust system. The introduction of the Gamma-Ti alloy in the turbine wheel of the EFR turbo results in a 40% lighter turbine wheel (Tetsui 1999) which in turn results in more efficient operation. The shaft connecting the turbine and compressor wheels is supported by ceramic dual layer ball bearings which is push-fit into the water cooled bearing housing separating the turbine and compressor sides of the turbo. The use of the ceramic ball bearings and multi-approach cooling capabilities improves overall efficiency, responsiveness and reliability of the turbo, meaning the EFR can comfortably handle the demanding environments within which it is intended to be used. The EFR turbo by Borg Warner also harbours some electronic sensory equipment to enable the end user to accurately log information on the performance parameters of the turbo as well as the engine it is bolted to. Some of the electronic sensors include compressor wheel speed sensor and integrated boost control solenoid valve. The introduction of these electronics as integrated components is an innovation which reduces the complexity of installation. This means that the user does not have to provide external fabrications to mount these necessary pieces of equipment as well as not having to purchase aftermarket electronic sensory equipment which in-turn reduces the risk of purchasing incompatible components. Lastly, the EFR turbo features a “flexible” compressor housing outlet with an integrated bypass valve for relief of excess pressure. The “flexible” housing basically allows the user to use multiple methods of attaching the compressor to the necessary pipe work. The way in which this is achieved is by providing provisions for both V-band and conventional hose fittings. The result of these inclusions is not only ease of installation due to the end user not having to purchase aftermarket bypass valves but also saving space in the engine bay. There is not just one single type of EFR turbocharger. The end user has the ability to mix and match compressor and turbine geometries to suit their particular application. For example, if the intended use of the turbo is for a rally car which requires an estimated 400 horse power, the customer has the ability to choose the size of turbine and compressor housing required to comfortably perform its required duties. Though there are many variant of this turbo, the features listed above are the same for each and every instance. These features are summarised in Figure 1 below: 2.0 Customer Satisfaction 2.1 Stakeholder Analysis Table 1 below lists both internal and external stakeholders of the Borg Warner EFR turbo charger. As Borg Warner is such a large international company, only the most influential stakeholders are considered. Table 1 Stakeholder Analysis Stakeholder Expectations Internal Borg Warner Executive Team • High return on initial investment in research and development • Greater portfolio in the automotive aftermarket sector through increase in reputation • Increase in market share in the automotive aftermarket sector Engineering Department • Greater investment in development • Greater influence on innovation in automotive aftermarket parts • Increase in reputation as innovators Shareholders • High return on investment • Increased innovation on new ground breaking technologies Borg Warner Floor Employees • Stable work environment • Safe working conditions • Increase in consumer demand to reaffirm employment status. External End User • Reliable and robust product • Excellent after-sales technical support • Long warranty period • High functional support of other common race-car aftermarket products • Comparative pricing compared to smaller specialist competitors. • Good value for money Suppliers • Increase in demand prompting more business • Improved business relationship through regular orders • Increase in innovative cooperation during new product development • Increase in involvement with new product lines • Higher profits Distributers/Retailers • On-time and prompt delivery • Excellent after-sales technical support • Predictable and fair pricing • Robust and reliable products • Effective communication channels for customer enquiries Competitors • Lower market share to decrease aftermarket dominance • Stalled innovation to decrease competitive advantage Environmental Interest Groups • High innovation in turbo technology for implementation to automotive manufacturers resulting in greater efficiency. • Use of recyclable materials for easy re-use. Government regulators • Adherence to minimum product warranty periods. • Adherence to current consumer legislation. 2.2 Capturing Customer Needs When using the Voice of the Customer (VoC) tabulated methodology of collecting and categorising customer needs and requirements (requirements being a subset of customer needs) there are many different methods of data acquisition which can be used. Some of these methods of capturing customer needs are listed below and described in the following page (Akpolat 2004; Cooper &Dreher 2010): • Customer interviews • Survey questionnaires • Customer visits • Customer advisory boards • Focus groups • Customer complaints • Ethnographic methodologies Customer Interviews Customer interviews are a succinct and cost effective method of extracting customer needs. The way in which this method is performed is dependent on various factors including: which customers to interview (past, present, most trusted, new customers etc.), the number of customers to interview at one time (one-on-one vs group interview), where the interview is to be taken and how the interview is to take place (face to face, via internet or telephone) (Akpolat 2004; Griffin & Hauser 1993). Even though the way in which we structure the interview is dependent on the various parameters listed above, the basic framework behind holding the interview remains relatively the same. Basically, the interviewer pre-prepares specific questions targeted towards obtaining the desired information (in this case customer needs) and structured to suit the interview method. Once the desired information is obtained, it can then be analysed and the information categorised to determine the classification of customer attributes (customer basic, spoken and subconscious needs) which will then be used to identify the potential functional requirements of a specific product (Yang 2008). This method is generally capable of providing qualitative data, however, under the right circumstances (question style and structure) and sample size this data can also be quantifiable. It is also important to note that performing customer interviews can also provide the wrong information depending on the style of questions e.g. customers may provide a problem or a solution rather than a need or requirement. Survey Questionnaires Customer survey questionnaires provide a cost effective method in capturing customer needs and can engage a vast number of participants through multiple media including mail and via the Internet. Unlike customer interviews, customer surveys only provide a one directional form of contact where there is no way to communicate with and engage the customer other than what it written in the questionnaire. This means great emphasis is required on the quality of the questions (open ended, direct etc.) in order to obtain a reasonable level of accuracy. This can be achieved by providing the customers with a range level (for example 1-5) indicating various level of satisfaction, dissatisfaction, agreeableness or level of disagreeableness depending on the intended outcome, which the customer can respond to. Like the customer interview method, this methodology captures customer needs and categorises them according to the responses given in the questionnaire. This method can provide quantifiable data for use in analysis depending on the quality of the questions and inherent answers given by the participants. Customer Visits Customer visits can be effective ways of indirectly acquiring customer needs. The basic methodology of acquiring these needs are the same as the previous two methods described earlier. Basically, the employee of the organisation who is accompanying the customer (whether the customer is visiting the organisation premises or vice versa) prepares a structured conversational guide before the customer visit in which the customer, through conversation, will subconsciously provide insights into their needs which the employee can then use as qualitative data in an analysis (McQuarrie 2008). This method of acquiring customer needs can be very time consuming and the amount of customers visits (depending on the size of the organisation) required to achieve a decent amount of data may not be feasible. The choice of customer to invite for a visit also becomes an important factor and the results can be highly subjective depending on the style of questions asked during conversation and the response of the customer to these questions (McQuarrie 1991). Focus Groups Focus groups are typically formed by 5-8 individuals (preferably current and past customers) and relies on the communications of these participants in order to generate some form of qualitative data (Akpolat 2004; Kitzinger 1995). This method differs from the previously described interview based methods of acquiring customer needs in that participants are encouraged to discuss amongst themselves rather than answering direct questions given by an interviewer. A focus group leader may ask “open ended” questions to encourage participants into discussing various product or service requirements without the bounds of a formal interview. This in turn result in more accurate qualified data due to the fact that this method doesn’t require the participants to be able to read or write and encourages participants to engage in conversation who normally won’t in a one-on-one interview, thus also providing a broader range of participants and results (Kitzinger 1995). Ethnographic Methodologies Ethnographic research methodologies are one of the most comprehensive (albeit time consuming) methods of capturing customer needs (Cooper &Dreher 2010). These types of research methodologies commonly involve communicating with the customer in their own environment using contextual interviews and simultaneously performing systematic observation through video recording the customer whilst utilising a particular product (Goffin et al. 2012). The contextual interview method is less formal than the customer interview and survey questionnaires where the interviewer does not strictly conform to a set of predefined questions but rather asks particular questions as the interviewer sees fit whilst the customer is performing their regular duties in their own environment. This type of informal interview and through observation, the interviewer or organisation can determine if the customer is correctly or incorrectly using a particular product and at the same time capture both directly and indirectly the customer needs. 2.3 Customer Needs Analysis (Kano Analysis) The method of choice to capture customer needs is the customer survey questionnaire. This method is undertaken using a sequential non-iterative 5-step process illustrated in Figure 2 below: Figure 2 Survey Questionnaire Process (Akpolat 2004) Step 1. Define the goals and objectives of the survey The goals and objectives of this particular survey are listed below: • To provide Borg Warner with information regarding customer needs. • To provide evidence of customer dissatisfaction and areas of improvement • To give insight to the research and design department in order to potentially predict the future customer requirements and in turn breakthrough technologies. Step 2. Develop a survey plan In order to reach a higher number of participants, a decision has been made to conduct the survey questionnaire via the Internet through a link on Borg Warner EFR home page. The required sample size is 384 people and the questionnaire will not be timed to provide the participant enough time to answer questions as rationally as possible. Participants will remain anonymous, however a specific serial number of the turbocharger will need to be provided in order to ensure only true Borg Warner customers are participating and increase the accuracy of results. Step3. Design and test the questionnaire The questionnaire will require the participant to score the importance of each turbocharger attributes from 1 to 5 (1 being least important and 5 being most important). The questionnaire will appear as in Figure 3 below: A smaller scale of 1-5 is used to encourage participants to think rationally about their answer. A trial run is to be performed prior to the actual deployment of the questionnaire to ensure the validity of the responses and calibrate the style of questioning accordingly. Once results have been collected from all 384 respondents, each attribute’s score will be summed to provide a cumulative score which will be categorised into customer musts, wants and extras as per the criteria described in Table 2 below: Table 2 Category Criteria Musts Wants Extras Cumulative Score Greater than 1400 1000-1400 Less than 1000 Step 4. Carry out the survey The survey will be carried out according to the scale and style of questionnaire as described previously. Step 5. Analyse the results The results of the survey are shown in Table 3 on the following page:   Table 3 Survey Questionnaire Results Attribute Cumulative Score Average Score 1 Oil supply to bearings 1536 4.0 2 Turbine shaft supported by journal bearings 1440 3.8 3 Welded turbine flange 1584 4.1 4 Cast-iron turbine housing 1584 4.1 5 Adjustable internal waste-gate 1560 4.1 6 Adjustable angle of compressor and turbine housings 1632 4.3 7 Compressor failure containment 1656 4.3 8 Pressure control valve outlet barb fitting 1560 4.1 9 Integrated compressor entry mesh cover 1536 4.0 10 Easy maintenance operations 1584 4.1 11 Low rev-range turbine response 1176 3.1 12 Integrated boost control solenoid 1008 2.6 13 Integrated pressure relief valve 1080 2.8 14 Twin scroll turbine housing 1200 3.1 15 Adaptable turbine flange 1224 3.2 16 Turbine shaft supported by ball bearings 1032 2.7 17 Integrated compressor speed sensor 1128 2.9 18 Lightweight construction (below 10kg) 1176 3.1 19 Adjustable entry and outlet adapters for various pipes/hoses 1200 3.1 20 Water cooling capabilities 1128 2.9 21 Adjustable turbine geometry 912 2.4 22 Low housing expansion due to heat (decreased build tolerances) 768 2.0 23 Remote boost controller (in-car installation) 720 1.9 24 Integrated heat shield 672 1.8 25 Turbine energy recovery system 744 1.9 According to the results of the questionnaire illustrated previously and the category criteria described in the survey design phase, the customer needs can be categorised using the Voice of the Customer chart shown in Figure 4 on the following page. Figure 4 Voice of the Customer Chart (Akpolat 2004) 3.0 Design Requirements 3.1 Translating Customer Needs into Design Requirements (QFD) The customer requirements listed previously are correlated to 10 specific product design requirements using the Quality Function Deployment (QFD) method. The importance rating of each customer need is gathered from the average score of the customer survey questionnaire completed earlier. These values are multiplied by 2 to provide a value out of ten in order to sustain a high degree of accuracy and consistency throughout the analysis process. The relationship of the customer needs and quantifiable design requirements are recorded using the logarithmic 1-3-9 value system (1 representing a weak relationship, 3 representing a medium relationship and 9 for a strong relationship). Once the relationship values have been assigned, an absolute score is calculated by multiplying the relationship with the corresponding importance factor for each attribute associated to the design requirements and summed up to an absolute value. The resultant QFD form is shown in Table 4 on the following page.   3.2 Design Risk Analysis (Failure Mode and Effect Analysis) The following Failure Mode and Effect Analysis (FMEA) is conducted on the 5 highest scoring design requirements using the risk severity and likelihood measures outlined in Table 5 and 6. Part Potential failure Effect on the system Root cause of failure Risk before action is taken Action to reduce risk Risk after action is taken P S R P S R Turbine Housing Overheating Reduced efficiency and potential for complete system failure Incorrect application of turbo onto exhaust manifold 3 4 12 Ensure the size of the turbo is applicable to the engine in which it is bolted to 1 4 4 Turbine Wheel Complete dissipation Complete system failure as the turbo is no longer affective Excessive pressure through unregulated exhaust gases 2 5 10 Ensure boost control is programmed correctly and is in good working order 1 5 5 Oil feed jacket Blockage in feed ducts Potentially high oil pressures and no oil feed to the bearings potentially resulting in complete system failure Jacket machined not according to specifications Incorrect viscosity of oil and incorrect assembly 2 4 8 Ensure quality assurance methods are being upheld before assembly and the correct oil specifications used 1 4 4 Connecting Shaft Housing Crack through material No oil retention, thus bearings risk operating in temperatures they weren’t designed for Incorrect assembly of housing to turbine and compressor and potential misuse by operator 2 3 6 Ensure the shaft housing is assembled according to correct torque specifications and checked to ensure conformance 1 3 3 Compressor Housing Boost pressure leakage Drastically reduced efficiency of the turbo Crack or hole in the aluminium housing. Incorrect assembly procedures. 2 2 4 Use higher strength aluminium alloy and conform to proper assembly procedures 1 2 2 15 Table 5 Measures of Consequence Level Designation Technical 1 Marginal System remains operational – very little rectification required 2 Minor System remains operational – minor amount of rectification required 3 Moderate System does not remain operational – minor amount of rectification required 4 Significant System does not remain operational – significant amount of rectification required 5 Major System does not remain operational – rectification isn’t possible Table 6 Measures of Likelihood Level Descriptor Probability Description 1 Unlikely <0.01 Event is expected to occur in most cases 2 Possible 0.02-0.05 Event would probably occur at some time 3 Likely 0.05-0.49 Event should occur at some time 4 Probable 0.49-0.99 Event could occur at some time 5 Highly probable >0.99 Event could occur only under special circumstances 4.0 Supplier Selection and Evaluation 4.1 Identifying Components The list below shows 10 major components of the Borg Warner EFR turbocharger which can potentially be sourced from external suppliers: 1. Aluminium compressor housing 2. Stainless steel turbine housing 3. Gamma-Ti turbine wheel 4. Aluminium compressor wheel 5. Cast iron shaft housing 6. Integrated bypass valve 7. Integrated speed sensor 8. Adjustable waste-gate 9. Connecting shaft 10. Boost control solenoid 4.2 Developing Supplier Selection Criteria The component chosen is the stainless steel turbine housing. 10 different selection criteria have been listed below to enable effective supplier selection: 1. Lead time 2. Quality Assurance procedures 3. Reputation 4. Price to manufacture 5. Location 6. Environmental sustainability 7. Production capacity 8. Reliability 9. Customer service 10. Safety record 4.3 Constructing and Using a Supplier Selection System Due to the vast technical nature of the EFR turbocharger and the volatility of the automotive aftermarket sector, selection of a suitable supplier of key components becomes of upmost importance. The method in which Borg Warner uses to undertake this supplier evaluation is comprised of 3 steps: 1. Each supplier evaluation criteria is weighted according to what Borg Warner perceives to be important factors. This weighting carries a value from 1-10 where 1 is of least importance and 10 represents a very important criterion. 2. Once each criterion has been weighted accordingly, the suppliers under consideration are then given a score from 1-10 according to how well these suppliers meet the required criteria (1 being the lowest score and 10 being the highest). Of course this is purely subjective according to the overall perception of these suppliers by Borg Warner, however it is this perception, which ultimately guides the organisation to the most suitable supplier. 3. After each supplier has been assigned a score from 1-10 based on their performance on the specific criteria, each score is multiplied by the corresponding weighting and the resultant weighted score is added up for each criterion to provide a total weighted score. This total score provides an overall guide to Borg Warner as to which supplier to select. The higher the total weighted score, the more likely this supplier is to meet Borg Warner’s requirements. The process described above is undertaken using a tabulated format and the outcome is shown below in Table 7. Table 7 Stainless Steel Turbine Housing Supplier Evaluation Matrix Suppliers Selection and Evaluation Criteria Weighting Quality Castings Austral Engineering Turbine Co. Score Weighted Score Score Weighted Score Score Weighted Score Lead time 9 6 54 8 72 9 81 Quality Assurance procedures 8 9 72 7 56 8 64 Reputation 5 7 35 7 35 8 40 Price to manufacture 9 6 54 8 72 2 18 Location 6 6 36 6 36 5 30 Environmental sustainability 3 4 12 5 15 6 18 Production Capacity 7 8 56 6 42 7 49 Reliability 8 8 64 6 48 8 64 Customer Service 7 8 56 7 49 3 21 Safety Record 6 9 54 9 54 8 48 Totals 71 493 69 479 64 433 4.4 Supplier Evaluation Report The supplier evaluation matrix illustrated on the previous page has resulted in the supplier Quality Castings scoring the highest with 493 points, followed by Austral Engineering with 479 points and finally Turbine Co. with the lowest score of 433. The results of the total weighted scores is too close to be able to definitively determine the most suitable supplier for the stainless steel turbine housing and because each criteria is mutually exclusive, an analysis of the different scores in themselves can be used to help identify which of the three suppliers is the best. When considering all of the factors involved in the decision making process, the conclusion is that Quality Castings is the most suitable supplier. The following paragraphs outline the reasoning behind this decision with specific comparisons made with the other 2 competitors. Turbine Co. had the highest score for lead-time with a score of 9 compared to Quality Casting’s 6. However, the poor performances in the price and customer service criteria does not justify the excellent lead-time with Turbine Co. Borg Warner value customer service and price to a high degree as they would like to be certain if design specifications change midproduction, it would be imperative to have excellent customer service channels in order to minimise any further monetary losses as well as being able to accurately determine key delivery dates necessary for production. If Turbine Co. does not have great customer service; there is a possibility that these aforementioned dates are inaccurate and production will not cease when certain parameters of the turbine housing are changed mid-process. Reliability, safety record, QAP, location and production capacity remain relatively comparable with only minute differences in scoring and as such the decision to rule-out Turbine Co. was made purely due to the high cost of production and poor customer service. Austral Engineering had superior results in lead-time and price compared to Quality Castings, however lacked in production capacity and reliability. Due to the fact that the EFR turbocharger is part of the automotive aftermarket sector, it is extremely important that customer demands are met when it becomes necessary. If Austral Engineering can’t keep up with the necessary production rates, it will stall the remainder of the assembly procedures at Borg Warner and as such cost the company excessive amounts in lost revenue. In addition to this, Borg Warner hold reliability in relatively high regard as well because they would like to have a good relationship with suppliers who are consistent in both delivery and manufacturing accuracy and hold true what is agreed upon in the supply contract. If Austral Engineering struggle with the supply rate and are already stretched in resources to this extent, the combination of average reliability is a recipe for disaster as this can result in below the required amount of turbine housings delivered to Borg Warner with a chance of delayed delivery and sub-par manufacturing quality. 5.0 Statistical Process Control (SPC) 5.1 Identifying Processes The following 10 processes contribute to the production of the EFR turbocharger in no particular order: 1.Design process This process finalises the design phase of the turbo taking into consideration current customer needs, market trends and ground-breaking technology. This process develops final drawings for use during manufacture. 2.Supplier selection process The supplier selection process identifies components which are to be manufactured outside the organisation and which suppliers will be contracted. 3.Supplied goods inspection process This process checks to ensure all supplied goods are manufactured to specification and verifies the completeness of any Inspection and Test Plans (ITP’s), which may be applicable. 4.Single component internal manufacture process This is an internal manufacturing process that involves dispatch of drawings to appropriate departments along with manufacture process guidelines (routing forms) and actual manufacturing processes in order to produce single components. For example, the internal manufacturing process to produce the compressor housing can include: • Dispatch of drawings to casting and NC milling department, • Casting of the housing including any machining allowances • 3 axis NC machining of internal bore and mating surfaces • Dimensional check of critical dimensions • Hard anodising • Final over-all quality inspection 5.Final assembly process The final assembly process involves consolidating all of the internally manufactured, and supplied components and systematically assembling the turbo charger together. 6.Packing process This process involves collecting all of the assembled turbochargers and packaging them with the necessary quality documentation and auxiliary components necessary for correct installation. 7.Distribution process The distribution process includes loading packaged products into traceable batches onto delivery trucks and distributing to their respective recipients. 8.Invoicing and consolidation process Process to ensure timely collection of funds and appropriate accounting activities. 9.Internal Quality Assurance conformity process This process involves periodic checking of conformance to all internal quality assurance measures including process forms, material traceability reports, routing forms, ITP’s and non-conformance reports (NCR’s). This process ensures the credibility of all quality measures as well as making sure that each product has been manufactured to its corresponding specifications. 10.Marketing and advertising process This process develops marketing and advertising strategies to ensure that the EFR turbocharger has good market exposure and a better chance of succeeding.   5.2 Monitoring Processes using SPC Charts Statistical process control charts are an effective tool used to visualise and pin-point statistical anomalies in specific process performance by an organisation undergoing SPC. These charts provide vital information on process effectiveness by plotting data related to a measure of effectiveness of a particular process and allow the user to visualise patterns in process performance. As a result, this chart gives an organisation the ability to identify variations in process performance and analyse the accuracy of any given process. The identification of certain variations (both as a result of the inevitable natural occurrence and possible special causes) allows the user to effectively optimise any given process. The typical SPC chart is comprised of the following features: • Upper Control Limit (UCL) and Lower Control Limit (LCL), which indicate the acceptable range of process variation. If any data point goes beyond these limits, the process is deemed out of control. • Centre line, which indicates the mean value of a particular process characteristic. This line represents the value in which a process is deemed in control. SPC charts can be categorised according to the data set in which they use, namely: attribute (or discrete) data, and variable (or continuous) data. Attribute data refer to quality characteristics that cannot be numerically represented e.g. amount of defects or non-conformance items. Three of the most common SPC control charts which use attribute data are described below: • P- Chart The P-Chart is used to plot the proportion (or fraction) of defective (or nonconforming) items in a varying or constant sample size. It is important to note that the P-chart only registers the number of defective items as a proportion of the whole sample size, not the number of defects per item. As such, there are only two distinct possibilities i.e. conformance or non-conformance. This chart can be used to monitor a process by tracking the proportional amount of defective items per unit of time and comparing this value with the mean value designated by the centreline of this chart in relation to the UCL and LCL. The variations can be visually assessed and statistical patterns analysed to ensure a particular process is in control or not. For example, this chart can track the process of impeller machining by recording the proportion of defective impellers per day. If the sequence of values plotted resembles an uncontrolled process, measures can be taken to restore it back to peak performance. • Np-Chart The np-chart, in comparison to the p-chart, plots the number of defective items in a constant sample size. The UCL and LCL are constant in this chart as the centreline value remains constant for every subgroup. It is important to note that this chart does not plot the proportion of defective items in a varying sample size, but rather counts the number of defective items evident in a constant sample size. The np-chart can be used for process monitoring by tracking the amount of defective items per day and (similarly to the p-chart) analyse the statistical pattern of defective items with reference to the UCL and LCL in order to determine process stability. If the pattern produced by these values on the chart resembles an unstable process (for example if one of the values falls well beyond either control limit) then this alerts the operator or organisation that something needs to be done to regain control of the process. • C-Chart The c-chart plots the number of defects detected in a constant sample size. The centreline is found by calculating the average number of defects at a finite amount of time within a constant sample size. In comparison to the np-chart and p-chart, the c-chart counts the number of defects not defective items within a specific sample size. For example, if there were 5 defects found in 1 turbine housing within a constant sample size of 10 housings in one day, the np-chart will register only 1 defective item in one day whereas the c-chart will register 4 defects in one day. This particular chart can be used for monitoring processes in the same way as the previously mentioned charts. As an example, if while recording the number of defective turbine housings, no number of defects fall above the standard deviation (1𝜎) within 15 units of time, this value is too good to be true and the organisation can re-evaluate the process parameters to ensure true determination of defects is being recorded. If not, then, as previously mentioned, measures can be taken to ensure accurate data logging (for example, the process used to detect defects can be fault tested then calibrated or refined to provide the correct data). Variable data is gathered through measuring specific quality characteristics during a process including the width, temperature, thickness etc. of a particular product. This data is expressed numerically and can provide for a much more in-depth analysis of process procedure variations than the non-numerical attribute data. Two SPC charts that use variable data are described below: • X-Chart Unlike the charts described previously, the x-chart uses measured data in a sample size of only 1 at a given time frame. For example, if the inside diameter of the turbine housing is measured as a quality characteristic, only 1 measurement will be made per day (if the period of measurement is in days). The way in which process monitoring is achieved with the x-chart is by analysing the patterns formed of the measurement values made per unit of time and comparing these patterns with known patterns of uncontrolled processes. There are some limitations when it comes to using this chart to monitor a particular process as only a sample size of 1 per subgroup is utilised. This makes effectively monitoring the process of large sample sizes tedious. • 𝑋̅ and R-Chart The 𝑋̅ and R-Chart actually depict two different charts that are batched to provide a more thorough analysis. The 𝑋̅ chart plots the average of measurements obtained within a specific subgroup over time. The centreline is depicted by the average of all the average measurements made within each subgroup. The R chart, on the other hand, plots the range of measurements identified in each subgroup and the centreline is found using the average of all of these ranges. It is important to note that this chart is only valid from sample sizes of between 2-9. How these charts enable effective process monitoring is by combining two areas of interest which in turn allows for a more in-depth analysis of process effectiveness. The range chart allows the user to track sample range compared to the average (stable range) in order to detect any anomalies which may suggest an out of control process whilst the 𝑋̅ chart provides the user with information regarding average measurements of a subgroup within the sample size of 2-9. Again, it is up to the user to interpret the chart in order to gain an understanding of how the process is performing. 5.3 Constructing and Using a SPC Chart As the number of defective turbine housings would like to be scrutinised within a variable sample size (20% of all housings in a batch will be inspected and because the batch size varies depending on customer demand, the sample size will vary accordingly), Borg Warner have decided to make use of the p-chart to help carry out its SPC activities. The results from the first 3 weeks of production (Monday to Friday) have been tabulated below in Table 8 and then graphically represented in a P-Chart in Figure 5: Table 8 P-Chart Metrics Day 𝑥 𝑝 𝑝̅ 𝑈𝐶𝐿 𝐿𝐶𝐿 1 18 0.200 0.195 0.320 0.070 2 17 0.179 0.195 0.317 0.073 3 17 0.179 0.195 0.317 0.073 4 17 0.183 0.195 0.318 0.072 5 19 0.209 0.195 0.320 0.070 6 18 0.196 0.195 0.319 0.071 7 22 0.244 0.195 0.320 0.070 8 17 0.185 0.195 0.319 0.071 9 19 0.211 0.195 0.320 0.070 10 17 0.187 0.195 0.320 0.070 11 17 0.183 0.195 0.318 0.072 12 17 0.183 0.195 0.318 0.072 13 18 0.198 0.195 0.320 0.070 14 20 0.211 0.195 0.317 0.073 15 19 0.200 0.195 0.317 0.073 2.926 The proportion (𝑝) of defective items is given by Equation 5.1 below: (5.1) Where x represents the number of defective items and n is the sample size. The centreline (𝑝̅) is found using Equation 5.2: (5.2) Where k is the number of days (number of subgroups). Finally, The UCL and LCL were found using the equations (5.3) and (5.4) respectively: (5.3) (5.4) The P-chart in Figure 9 shows the process of manufacturing the turbine housing is technically in control. The variance in the proportion of defective items per day is minimal with the highest proportion of defective items occurring on day 7. This variance is coincidently the reason why this process is deemed in control because if that particular value fell under the standard deviation line (indicated by the green and red dotted lines labelled “sigma” on the p-chart) this process would technically be regarded as out of control according to Nelson Rule No. 7. The majority of the proportion of defective item values fell well within the upper and lower control limit showing excellent accuracy and consistency. However, the draw back of using this particular chart is that it does not show how many defects where evident in the particular turbine housings deemed defective. Other more in-depth variable data charts may provide a more concise analysis which pin-points minor deviations of a particular quality characteristic, though the p-chart is a computationally inexpensive method providing a broad picture of the effectiveness of the turbine housing manufacturing process. 6.0 Problem Solving 6.1 Identifying Problems and Problem Solving Tools During the operation of the EFR turbocharger, some problems may occur including: 1. Bearing failure (cooked bearings) 2. Overheating 3. Internal waste-gate jammed open or shut 4. Faulty boost pressure regulator (irregular electronic signals to internal solenoid) 5. Complete compressor wheel dissipation 6. Debris stuck in the compressor housing 7. Leaking boost pressure 8. Pressure relief valve failure (internal compression spring stuck in a single position) 9. Over-boosting Some problem solving techniques which can be used to treat these problems include (Kanji & Asher 1996; Smith 1998) 1. Cause and effect diagram 2. Brainstorming 3. Why-why diagram 4. Pareto diagram 5. Check sheet 6. Process flow chart 7. Gantt chart 6.2 Constructing and Using Problem Solving Tools The potential problem chosen to be rectified is bearing failure. The chosen problem solving methods include the following: 1. Brainstorming 2. Why-why diagram 3. Check sheet 4. Pareto diagram These problem-solving methods are applied below: Brainstorming The brainstorming method is performed by putting forward the problem to a group of people (preferably team members involved in the functional design of the turbocharger) and recording all responses which may indicate possible causes. Some such responses to the brain storming session are illustrated below in Figure 6. This initial brainstorming session provides a broad range of possible causes, some of which may not even be very applicable to the problem at hand. However, this provides an appropriate platform required to undertake the proceeding problem solving tools which will expand on this initial assessment. Why-Why Diagram The why-why diagram in Figure 7 below is an effective tool to help discover the underlying root causes of particular problems. In this case, each line connecting two causes represents the question “why?” This question can be asked repeatedly in order to draw out any possible root causes of bearing failure in the EFR turbocharger. The why-why diagram below is representative of only some of the possible causes and can be expanded much further, however for the purposes of demonstrating how this method is to be used as a problem solving tool, this is sufficient. Figure 7 Why-Why Diagrams Check Sheet Once potential causes of the problem have been identified (using the brainstorming session and the why-why diagram), it is now possible to determine which of these causes is the predominant reason behind bearing failures in turbochargers. Once such method to be able to accomplish this is the check sheet. The check sheet in Table 9 is populated using documented cases of root causes collected within one week. These root causes have been documented by an official Borg Warner EFR turbocharger maintenance sub-division by every technician that has found these causes to be responsible for bearing failure. This table shows the most pronounced cause of bearing failure to be the lack of oil flow with overheating being the least pronounced. Table 9 Check Sheet Root Cause Occurrence Absolute % Accumulative % Lack of oil flow 25 30.9 30.9 Cracked oil jacket 17 21.0 51.9 Distorted connecting shaft 15 18.5 70.4 Over boosting 10 12.3 82.7 Cracked bearing housing 7 8.6 91.4 Clogged water cooling channels 5 6.2 97.5 Overheating 2 2.5 100.0 Totals 81 100 Pareto Analysis The Pareto chart in Figure 8 is populated using the occurrence and accumulative percentages of identified root causes. The outcome of the analysis has identified 3 areas of concern (namely: lack of oil flow, csracked oil jacket and distorted connecting shaft). These 3 possible root causes represent approximately 80% of all causes and should be given priority in terms of treatment then the remainder of root causes. Through these 4 method of problem solving, Borg Warner was able to determine a broad range of possible causes through a brainstorming session, go deeper into the causes and find underlying root-causes, identify which of these causes is occurring the most and finally figure out which causes deserve priority treatment and which can be deemed negligible. Once the causes which need immediate attention are identified measures can be taken to reduce the possibility of these causes. In this case possible remedies may include better technician education in regards to installation and regular maintenance checks and maintenance log-book for better traceability. These processes can also be conducted using a problem solving form however this form shouldn’t be a standalone document filled out by one person. Exhaustive identification methods as per the methods used above should be used to find possible remedies to rectify the problem. References Akpolat, H. 2004, Six Sigma in Transactional and Service Environments, Gower. Bayomi, N.N. & Abd El-Maksoud, R.M. 2012, 'Two operating modes for turbocharger system', Energy Conversion and Management, vol. 58, no. 0,pp. 59-65. Bell, C. 1997, Maximum Boost: Designing, Testing, and Installing Turbocharger Systems, Robert Bentley Automotive Publishers. Chiong, M.S., Rajoo, S., Martinez-Botas, R.F. &Costall, A.W. 2012, 'Engine turbocharger performance prediction: One-dimensional modeling of a twin entry turbine', Energy Conversion and Management, vol. 57, no. 0,pp. 68-78. Cooper, R.G. &Dreher, A. 2010, 'Voice-of-Customer Methods What is the best source of newproduct ideas?', Marketing Management, vol. 19, no. 4,p. 38. Goffin, K., Varnes, C.J., van der Hoven, C. &Koners, U. 2012, 'Beyond the Voice of the Customer', Research Technology Management, vol. 55, pp. 45-53. Griffin, A. & Hauser, J.R. 1993, 'The Voice of the Customer', Marketing Science, vol. 12, no. 1,pp. 127. Kanji, G.K. & Asher, M. 1996, 100 Methods for Total Quality Management, SAGE Publications. Kitzinger, J. 1995, 'Qualitative research. Introducing focus groups', BMJ: British medical journal, vol. 311, no. 7000,p. 299. McQuarrie, E.F. 1991, 'The customer visit: qualitative research for business-to-business marketers', Marketing Research, vol. 3, no. 1,pp. 15-28. McQuarrie, E.F. 2008, Customer visits: Building a better market focus, ME Sharpe. Parts, E.J. 2013, Borg Warner, viewed 1 May 2014 2014, . Smith, G.F. 1998, Quality Problem Solving, ASQ Quality Press. Tetsui, T. 1999, 'Gamma Ti aluminides for non-aerospace applications', Current Opinion in Solid State and Materials Science, vol. 4, no. 3,pp. 243-8. Yang, K. 2008, Voice of the Customer : Capture and Analysis, Book, McGraw-Hill, New York.