Theory and Practice of Quality and Reliability Engineering in Asia Industry

Statistical process control charts are important tools for monitoring process stability in manufacturing and service industries. Using multivariate control charts to monitor two or more related quality characteristic has become increasingly popular in today’s environment. Out-of-control signals in multivariate control charts may be caused by one or more variables or a set of variables. In the practice of quality control, in addition to the quick detection of process change, it is also critical to diagnose the change and identify which variable(s) has changed after an out-of-control signal occurred. In this paper, we propose a decision tree (DT)-based ensemble classifier to approach the problem of diagnosing out-of-control signal. The focus of this paper is on the variance shifts of a multivariate process. Simulation studies indicate that the proposed ensemble classification model is a successful method in identifying the source of variance change. A comparative study also reveals that DT using extracted features as input vector has slightly better classification performance than using raw data as input. The proposed method could be a useful diagnostic aid in monitoring shifts in process dispersion.

Signal processing of vibration signals from rotating machinery has been an active research area for recent years. Especially, discrete wavelet transform (DWT) is considered as a powerful tool for feature extraction in detecting fault on rotating machinery. However, the number of retained DWT features can be still too large to be used for standard multivariate statistical process control (SPC) techniques although DWT significantly reduces the dimensionality of the data. Even though many feature-based SPC methods have been introduced to tackle this deficiency, most of methods require a parametric distributional assumption that restricts their feasibility to specific problem of process control and thus limits their applications. This study introduced new feature extraction technique to alleviate the high dimensionality problem of implementing multivariate SPC when the quality characteristic is a vibration signal from bearing system. A set of multiscale wavelet scalogram features was generated to reduce the dimensionality of data, and is combined with the bootstrapping technique as nonparametric density estimation to set up an upper control limit of control chart. Our example and numerical simulation of a bearing system demonstrated that the proposed method has satisfactory fault-discriminating ability without any distributional assumption.

The Anderson leadership team “discovered” a new formal approach to continuous improvement in 2004—Lean Six Sigma (LSS). Senior management integrated the two approaches into Lean Six Sigma—attacking both the product and process variation and waste. Our processes are supported with over 100 engineers, mechanics, and technicians. A significant training effort allowed all of them to be Lean Certified. They have saved over $10 M annually in improving equipment and setups alone. Anderson has developed 18 of 200 Green Belts (our best) into Black Belts and six of them into Master Black Belts. They saved almost over $7 M annually. The Steering Committee has committed everyone in every department to participate. To that end our projects are focused in every area of the company: Production 44%, Quality Control/Assurance 26%, Maintenance 8%, Warehouse 8%, Sales/Administrative/Finance 8%, Engineering 4%, and Human Resources 2%. The results are cost reduction, improved customer service and measurable improvement in quality, employee and customer satisfaction resulting in over $15 M annual savings plus cost avoidance, and increased capacity. In addition, over 25% of our projects improve our regulation and compliance. As a result of these efforts in 2011 Anderson won the IQPC’s “Best Manufacturing Project Award.” In 2012 Anderson won the IQPC’s “Best Manufacturing Continuous Improvement Program Award.” In 2013 they won the “State of Illinois Malcolm Baldrige Award.”

Internal audit planning is a critical step in achieving effective and efficient internal audits. This study looks at how risk management approach is used to plan internal audit. Purpose is to reduce the subjectivity and time spent in audit planning. A two-step risk management approach is used. In Step 1, the audit frequency is determined based on the department’s Workplace Environment, Health and Safety (EHS) risk. In Step 2, the level of readiness of a department is determined by looking at the department’s level of documentation and level of measuring and monitoring of processes. From the risk assessment, a 3-year cycle plan is developed to ensure all departments are audited at least once in every 3 years. Higher risk departments would be audited more frequently. A review of risk factors is conducted every 3 years or as and when necessary. The subjectivity in selecting a department for internal audit is reduced. Another significant benefit achieved is the reduction in time spent to do audit planning. With the risk-based audit planning, new staff who is not familiar with audit planning can complete the audit planning process efficiently and effectively. Through the risk management approach, the internal audit planning is more structured and effective.

Over the years, there is an increasing trend of using quality tools and techniques in service sectors to address customer feedback. Deployment of such quality tools in the education sector is relatively new. Coleman College in Singapore faced feedback from the student community that the Advanced Diploma in Accounting course is too focused on academic learning. This issue was brought up to the Academic Board for further discussion. There were a couple of suggestions and the team embarked on a root cause analysis to analyse the feedback. One root cause identified is the lack of skill-based learning in the current course. The team then brainstorm to incorporate MYOB, the accounting software, into the course programme as a new module. In addition to adding more practical aspects to the course, this implementation also benefits the students by improving their employability. A post-implementation feedback was gathered again to ascertain the effectiveness of the learning outcome. The feedback was positive and the implementation is successful.

The creation of the AUN-QA was initiated in 1998 as a network of Chief Quality Officers (CQOs) appointed by the AUN member universities as the focal point for coordinating activities to realise the mission of harmonising educational standards and to seek continuous improvement of academic quality of universities in ASEAN. Since then, AUN-QA has been actively promoting, developing, and implementing quality assurance practices based on an empirical approach where quality assurance practices are shared, tested, evaluated and improved. AUN-QA assessment was inaugurated in December 2007. From 2007 to 2012, 16 AUN-QA assessments covering four countries, 7 universities and 37 programmes were carried out. These 37 programmes were assessed by AUN-QA assessors based on the 15 criteria of the AUN-QA Model at the programme level [1]. To assess the impact of AUN-QA assessments on the quality of educational programmes, an analysis of the results was made using medians and categories of criterion rated based on the AUN-QA Rating Scale. Based on the medians of the AUN-QA criterion rated at the programme level, the number of criteria rated “Better than Adequate” has increased from 1 in 2007 to 13 in 2012. This demonstrates that the quality of the assessed educational programmes at the participating universities has improved over the last 6 years. Recommendations to further improve the quality of the educational programme in the 15 criteria are made.

Due to the government’s education policy significantly change in higher education, number of new universities has established dramatically in the past decades in Taiwan. As a result, the universities are facing the strongly competitive pressures. Most of the universities adjust their strategy in improving service quality to enhance their competition among the universities. Usually they used the customer satisfaction survey to ensure the actual service quality from student’s perspectives, and then find out the service items or quality attributes with low satisfaction levels as the items to be improved. Then they will adjust and response to those items which need to be enhanced or improved furthermore to improve their overall service quality and attributes. Several studies assert that only using the satisfaction survey and the related improvement actions may not be obtained the significant results in improving service quality. The main cause is that the improvement items may not concerned by student. Usually the student will concern the service context and quality attributes with high importance. Therefore, we consider the famous I-S model and the related questionnaire to evaluate the service quality for university. We have conducted this model and the related questionnaire survey in a private university in Taiwan. According to I-S model analysis, the results of this research survey show the main pole to be improved (high importance and low satisfaction), including equipment, facilities need to be replaced or upgraded. The survey results will be improved subsequently as further reference for improvement.

Data management is one of the necessary processes in quality management. Chulalongkorn University has developed the tool for this process called Quality Assurance Information System (QAIS). This system was originally developed based on “Lotus Notes” software. Currently, its system consists of the processes of data input, reference input, data checking and correction, data evaluation, and report preparation. By using this system, the data collection process changes from “Offline” to “Online via Intranet.” It has been used to support internal and external quality assessments according to national quality standards of Thailand since this system can connect and transfer data to the national database called CHE QA Online (Commission on Higher Education Quality Assessment online system). Chulalongkorn University was nationally awarded by the Office of the Public Sector Development Commission in the category of Servicing Innovation for implementing this system. Therefore, Chulalongkorn University would like to share and exchange the concept and the operation of this system for further improvement.

The concept of Total Quality Management (TQM) has been discussed in the present work stressing its application in the field of education. Side by side definition, concept, principles of TQM, and philosophy are discussed. In this present work also discussed synergistic relationship, continuous improvement, self-evaluation, a system of ongoing process leadership, etc. TQM process in educational institutes generally includes (a) Mission and Passion, (b) Administrator as a role model, (c) Environment factors, (d) Accountability, (e) Humans relations and (f) feedback. Implementation principles and processes of the concept in education have been narrated after identifying the barriers for implementing total quality in education. TQM will help achieve excellence, which only can guarantee the survival of institutions now in a highly competitive world, with ever decreasing subsidy in the education sector.

Preventing falls is one of the clinical problem and causes patients’ significant problem, since it may cause serious injuries such as an abrasion and fractures. An assessment score sheet for inpatients is one of the falls prevention system. In this paper, we focus on sheet for inpatients’ fall risk. In this paper, we validate the appropriateness of the falls estimate in patients hospitalising. We reassessed the score of 84 cases which consist of 42 sets of the actual patient data for patients who have fallen and 42 sets of the data for those who have not from seven wards from October 2013. The results showed four items as greatly influence the falls. With these items, we calculated fall probabilities of each patients, and compared them of patients who have fallen and those who have not. Then, we could not recognise the difference between two groups since the number of people whose probability of fall is more than 60% are 34 cases in patients who fallen and 24 cases who did not fall. From the result of the analysis, we discuss it is not enough that nurses estimate the patient’s fall risk only using the assessment in patients at the time of hospitalising. Instead, the importance of daily assessment is shown.

Medical technology has been consistently improving in recent years. With that, people are more concerned with medial safety and its quality. In medical service provided by hospitals, quality is greatly influenced by human knowledge and skills. Therefore, it is fundamental to improve these factors. Actually, at many hospitals, medical safety education is provided to employees. To conduct more efficient education, the effectiveness of the education must be evaluated, and the specific contents must be revised after its administration. It is said that effects of an education can be evaluated in terms of learning (knowledge and skill) and the degree of behavior modification and results. In evaluation, we should begin by evaluating learning, and then move to behavior and results. However, the methodology for measuring these has not been established. In this study, we analyzed the long-term findings in a hospital and implemented an evaluation model for medical safety education. Further, we designed the effectiveness measurement tool based on that model. We then propose a method for measuring the effects of health-care safety education. Using this method, medical workers can evaluate students’ learning free of bias and extensive work after providing education. In addition, they can review the contents of education and provide sufficient feedback on the test results to students.

Home healthcare (HHC) service systems have been attracting researchers’ attentions in recent decades. Many resources are involved in delivering HHC service, consisting of different categories of human resources (nurses, physicians, physiotherapists, social assistants, and psychologists), support staff, and material resources. One of the main issues encountered when managing HHC services is the patient assignment problem, i.e., assigning operators to admitted patients given some sets of constraints (e.g., the continuity of care). This paper addresses the patient assignment and grouping problem for HHC system, Alice Ho Miu Ling Nethersole Hospital (AHNH) in Hong Kong, which focuses more on assigning new-admitted patients in a period to operators with achieving the aims of workload balance and minimum delay visits, and meanwhile grouping the assigned patients for each operator based on the patients’ geographical location and next visit due time to reduce the travel time and improve the efficiency of operators. This hospital suffers from ineffective manual planning methods to manage their human resource when providing home health care. In order to improve the quality of service and efficiency of operators, we model the patient assignment and grouping problem and will use heuristic method such as discrete particle swarm optimization, Tabu Search, genetic algorithm to obtain the optimal solution.

In studying the service quality of a service provider, customers are often asked to rate the services they have received through a questionnaire. Based on the responses received, various models/techniques that analyze the scores of different components have been proposed to evaluate overall service quality. In particular, the average score of a service provider often serves as a basis for comparison between two or more service providers. In practice, the scores associated with each response level (e.g., agree, strongly agree) are often based on intuition or some simple rule, which may not correctly reflect the differences between different response levels. Therefore, it is important to examine the robustness of a conclusion (e.g., service provider A is better than service provider B) with respect to changes in the scores assigned to each response level when studying the service quality of service providers through customer surveys. It is commonly assumed that the difference between the average scores of two service providers provides a good indication of the robustness of the conclusion. In this paper, we illustrate that this may not be true and highlight that a separate analysis on the robustness of a conclusion is necessary. We describe how the robustness of a conclusion can be represented and show how it can be computed in practice.

Environmental quality for the general population is dominated by air quality. Thus modeling of air quality is the first step toward any program for quality improvement. This paper describes the use of the ARIMA (Autoregressive Integrated Moving Average) time series modeling approach, illustrated by the tracking of the daily mean PM2.5 concentration in the north region of Singapore. A framework for ARIMA forecasting revised from the general Box-Jenkins procedure is first outlined; T-test and three information criteria, namely, AIC (Akaike Information Criterion), BIC (Bayesian Information Criterion), HIC (Hannon-Quinn Information Criterion) are employed in addition to analyses on the ACF (autocorrelation function) and PACF (partial autocorrelation function). With forecasting as the primary objective, the emphasis is on out-of-sample forecasting more than in-sample fitting. It is shown that for 30 such forecasts, one-step ahead MAPE (mean absolute percentage error) has been found to be as low as 8.0%. The satisfactory result shows the classical time series modeling approach to be a promising tool to model compound air pollutants such as PM2.5; it enables short-term forecasting of this air pollutant concentration for public information on air quality.

It is widely acknowledged that environmental quality and protection should not be ignored in the development of a society. Environmental quality actually depends not only on government programs and policies but to a large extent awareness and actions from the public. However, the latter would then relate to human behavior which could be subjective, irrational, and impressionable as well. In recent years, the fast development of Internet technology has propelled the boom of social media and instant communication tools which enable the public to express their opinions on current events more conveniently and quickly. Nevertheless, the majority of swift comments tend to be emotional or one-off knee-jerk reaction which may fail to appreciate the actual impacts behind. With this background, the importance of critical and statistical thinking in the Web era has been illustrated in this paper. One case study, which relates a recent actual incident (discarding of rubbish) on Tiananmen Square in the capital of China, is used to underscore the need for rational opinions and the circumspect in spreading the message for enhancing environmental quality and handling related behavioral issues. It has been found that typical comments on the Internet in terms of this particular environmental issue reveal the usual emotional reactions and lack of statistical knowledge among the public. It is concluded that for advances in the importance quality of the environment, guidance of critical and statistical thinking is indispensable.

During charging of Lithium-ion battery (LiB), the charging cutoff voltage (COV) may exceed the manufacturers’ specification because of incorrect monitoring of the charging control circuit, either due to the aging of the control circuit or the design/manufacturing errors of the control circuit. In fact, it is found that overcharging LiB cell is a common abuse. This work shows the effect of excessive COV on cell’s discharging ability, and the use of a novel nondestructive method to evaluate if the damage made in the cell by the excessive COV is rendering the cell from further safe usage or it is still acceptable with minor degradation in reliability and safety, thus providing a basis for quality consideration of the cell. The method also enables battery manufacturers to identify the internal components for their cells that are most vulnerable to the excessive COV so that quality improvement of their batteries can be designed and produced. This method also alerts electric vehicles user on the hidden safety issues of their battery pack, and enables battery management system to perform reliability balancing, a new patented technique to ensure the safe and reliable operation of battery pack.

The field failure data is important in the reliability field. Component failure data, instead of system failure data, are usually collected by the system owners. However, in many cases the individual component time-to-failure data are not recorded. Instead, cumulative operation time and the number of failure times are recorded. This can be explained by the fact that exponential distribution is often used to fit the data. Unfortunately, the assumption of exponential distribution is not valid in many cases. In our paper, Gamma distribution and Inverse Gaussian (IG) distribution are used to fit the data. The point estimate for parameters of these two distributions can be obtained easily by general computer software. We then proposed methods to obtain the confidence interval (CI) for parameters of gamma distribution and IG distribution when the data are merged. In particular, for the rate parameter and mean of a gamma distribution, no efficient methods have been found to get the interval estimation. We borrow the idea of General Pivotal Quantity to obtain the interval estimation for the rate parameter and mean of a gamma distribution. The simulation study shows our method outperforms Wald method in terms of coverage probability. At last, an illustrative example is given to demonstrate the applicability of proposed models.

An electronic system contains millions of components on a single chip today, and the failure of any of these components may lead to the system failure. Conventionally, the system reliability is ensured by having a high reliable subsystem or component, and thus efforts are placed to study the reliability of individual components in electronic system. However, the failure mechanisms of a system is lot more complex, and the correlation between the system failure and component failure is not straight forward, depending on the system configuration, as demonstrated in this work. Therefore, electronic system reliability may not be derived from its components’ reliability.

In order to set an adequate lifetime target for each market, quantitative evaluation of variation of lifetime characteristics is required. In particular, the lifetime of vehicle unit depends heavily on customer’s usage (e.g., gross vehicle weight, road gradient, and acceleration operation). We thus have developed an online monitoring system that continually collects some information such as usage and environmental conditions. A method has been developed for predicting vehicle component lifetimes using data from an online monitoring system that collects an extensive amount of data during vehicle operation. The linear model used for prediction takes into account variations in usage conditions and models’ data as covariates. The prediction procedure was generalized to enable it to make predictions using a new data sample. The large amount of information on usage and environmental conditions obtained with the online monitoring system enabled the usage of each sample to be quantified and treated as a stratification factor. A stratified analysis produced fairly accurate results, meaning that using online monitoring data should be useful for lifetime prediction.

Online monitoring becomes a common tool for keeping high reliability of products and systems in these days. The information to maintain the target product includes usage history, system conditions, and environmental conditions. Nowadays they are monitored and reported in real time and stored as big data. On modeling the failure mechanisms statistically, these variables are primary candidates for covariates which affect the failure mechanism. There is literature on modeling the lifetime of products with the accelerated lifetime model on an accumulation of the covariate effects, which is a nonlinear transformation of the observed lifetime. The existing literature requires a parametric form of the transformed lifetime distribution in advance. However, such knowledge may be difficult to acquire in advance in some cases. When we assume an incorrect distribution, it is called misspecification. This paper proposes a strategy to use the log-normal likelihood for the estimation of the parameters of the covariate effects when the underlying distribution is either the Weibull or log-normal distribution. It is derived that the score function of the log-normal likelihood is identified as an approximation for the Weibull cases. The simulation study shows the relationship among the sample distribution of parameter estimates and underlying distributions. For the Weibull cases, if the shape parameter is large, the bias of the resulting estimates is small.

When tailor-made systems are introduced into fields, markets, or factories, the replacement parts or subcomponents for maintenance are produced and purchased simultaneously so as to keep the quality and reliability of those parts homogeneous against the parts installed into systems at the beginning. Those parts are then kept in stock at yards and will be kept waiting for their turns. Furthermore, the maintenance records of such systems are often window censored, under which the records are both left and right censored. In this talk, we model such situations as a renewal process with incomplete repairs, in which a failed subcomponent is replaced with an unused-but-not-new subcomponent. The statistical inference of the failure time distribution and the effect of in-stock term are implemented using stochastic EM algorithm, which is needed to solve the maximization problem of the incomplete data likelihood under window censoring. A real case is analyzed using the proposed model and the effect of in-stock-at-yards term on the total lifetime is assessed by assuming Weibull distribution. We also examine the sampling properties of the estimators for Weibull distributions.