International Journal of Quality Engineering and Technology (5 papers in press)

Regular Issues

• DESIGN OF EWMA CONTROL CHART FOR MONITORING TRANSFORMED RAYLEIGH DISTRIBUTED DATA  
  by Olatunde Adeoti 
  Abstract: Monitoring statistical process for the detection of assignable causes of variation is based on the assumption that the process characteristic follow the normal distribution. But, in practice, this is often not the case as process characteristic seldom follow the non-normal distribution. This paper design a new control chart to monitor quality characteristic that follow the non-normal distribution. The proposed control chart based on the EWMA statistic is constructed after transforming the Rayleigh distributed data to approximate normal using the power transformation method. The ARL and SDRL values of the proposed control chart are evaluated for different shift sizes. The performance of the proposed chart is compared with the recent CUSUM chart for transformed Rayleigh distributed data. The study shows that the proposed chart outperforms the recent CUSUM control chart for transformed Rayleigh data. Real-life and simulated dataset to illustrate the design and applications of the proposed control chart is given.
  Keywords: Control chart; Transformed Rayleigh data; EWMA; ARL; Power transformation.
  
  
• Joint optimization of production run length and maintenance policy for an imperfect process with multiple correlated quality characteristics  
  by Ali Salmasnia, Maryam Kaveie 
  Abstract: The earliest economic production quantity models assumed that the manufacturing process and the quality of produced items are perfect. While in a real situation, non-conforming products are fabricated and machine failure happens. Hence, the production systems are increasingly engaged in the improvement of machines availability and products quality. In this regard, this paper presents an integrated production and maintenance planning model under monitoring multiple quality characteristics. To adapt to the real production conditions, it is considered that quality characteristics are correlated. Furthermore, to improve the power of process monitoring, a shewhart control chart is designed by considering both economic and statistical criteria. Due to the complexity of the problem, the particle swarm optimization algorithm is employed to optimize the expected total cost per time unit, subject to statistical quality constraints. Here, an industrial example is given to show applicability of the presented mathematical programming. Furthermore, to demonstrate the validation and effectiveness of the suggested approach, a comparative study is presented. It confirms that the integration of production planning, maintenance policy, and statistical process monitoring leads to a significant increase in the cost savings.
  Keywords: Production run length; maintenance policy; statistical process monitoring; multiple-quality characteristics.
  
  
• Design and Implementation of ARL-unbiased CCCr-chart for Monitoring High-yield Processes  
  by Nirpeksh Kumar, Ranjeet Kumar Singh 
  Abstract: In order to overcome the shortcomings of the conventional charts such as p-, c-, u-chart in monitoring the high-quality processes with low fraction nonconforming, the cumulative counts of conforming (CCC) charts are recommended in Statistical Process Control (SPC) literature. To improve further their ability of detection early shifts in the fraction nonconforming, the CCC_r-charts are proposed considering the cumulative count of conforming items up to the r-th nonconforming one. But the CCC_r-charts perform poorly in detection of small downward shifts in the fraction nonconforming because of their undesirable ARL-biased property. This results in the larger out-of-control (OOC) average run length (ARL) values than the in-control ARL value for some values of the fraction nonconforming. In this paper, we eliminate the ARL-biasedness property and propose the ARL-unbiased CCC_r-charts using the notion of uniformly most powerful unbiased (UMPU) test to ensure that a user will get an OOC signal more quickly than a false alarm for the shifts in both upward and downward directions. The performance of the proposed chart is also compared with the existing ARL-unbiased CCC chart and it is found that the former has an improved ability of detecting shifts in the fraction nonconforming over the latter. An illustrative example is given and a summary and conclusions are offered.
  Keywords: average run length; ARL-unbiased; control chart; fraction nonconforming; geometric distribution; high-yield processes; in-control and out-of-control performance; uniformly most powerful unbiased test.
  
  
• A fuzzy Kano-QFD approach for prioritizing NBA quality parameters for service quality enhancement: a case of Indian PEIs  
  by Pramod Kinker, Vikas Swarnakar, A.R. Singh, Rajeev Jain 
  Abstract: The aim of this work is to identify and prioritise the National Board of Accreditation (NBA) of India quality parameters as design characteristics (DCs) to direct and develop educational services by the influence of service quality factors in students and teachers perspective in polytechnic education institutes (PEIs) of Madhya Pradesh, India. A framework for identification and prioritisation of the service quality improvement factors for students and teachers perspective have been presented in the present study. A fuzzy Kano approach has been used to prioritise service quality attributes into different Kano categories. Then, QFD is used to rank the DCs. It facilitates polytechnic education planners, decision makers, and administrative peoples of institutions to design and develop a strategy to enhance the service quality of PEIs.
  Keywords: education; polytechnic education institutes; PEIs; service quality factors; fuzzy Kano; quality function deployment; QFD.
  
  
• Statistical Analysis of Factors Associated with Recent Traffic Accidents Dataset: A Practical Study  
  by Emad Imreizeeq, Jamal N. Al-Karaki, Amjad Gawanmeh 
  Abstract: In this paper, we propose a logistic model to fit accidents dataset of 10,000 road crash incidents for the Emirate of Abu Dhabi published in 2020. After cleaning up the dataset, we use descriptive and inferential statistical tools to study the attributes of each variable. Then, we identify the main independent variables that can be incorporated in a general logistic regression model which also includes the interactions between them. Our analysis using the significance level of (alpha = 0.05) found that there is a reduced logistic regression model that can fit the data in which the location of accident can be represented using type of accident and the age of people involved in the accidents. Moreover, the results show that the interaction terms are not significant to be included in the model. Furthermore, the study shows that the odds for accidents by young age group (less than 40 years old) in external streets is 27% higher than the odds for internal streets, and that the odds for sequential type accidents in external streets is 13% higher than the odds for internal streets.
  Keywords: logistic regression; traffic accidents; accident data analysis; chi-square test.