Augmenting the Taguchi loss function with AHP to perform economic evaluation of reducing variations in processes and products

Purpose – This paper deals with the identification and diagnosis of operational variability in chemical processes, which is a common problem in mills but little explored in literature. The Cross-Industry Standard Process for Data Mining (CRISP-DM) is a widely used approach in problem solving. The purpose of this paper is to: first, contribute to the body of knowledge on applying CRISP-DM in a pulp mill production process and the special issues that need to be considered in this context. Exact amounts of a cost increase due to variation in pulp production have not been reported previously. Second, to quantify the cost of variation. Design/methodology/approach – In the case studied, the variation in a pulp mill batch cooking process had increased. In order to identify the causes of variation, CRISP-DM was applied. Findings – The cycle of variation was identified and found to be related to the batch cooking process cycle time. By using information from this analysis it was possible to detect otherwise unobserved defective steam nozzles. The defective equipment was repaired and improved. Further improvement was achieved when the fouling of a heat exchanger was found by analysis to be the root cause of long-term variability parameters. By applying CRISP-DM, equipment defects and fouling were identified as the root causes of the higher manufacturing costs due to increased variation were detected and estimated. The Taguchi loss function is a possible tool for estimating the cost of variation in pulp manufacturing. Originality/value – This paper provides new knowledge in the context of implementing CRISP-DM and the Taguchi loss function in the pulp and paper manufacturing process.

Augmenting the Taguchi loss function with AHP to perform economic evaluation of reducing variations in processes and products

Production yield is usually defined as a function of the variability and tolerance interval of processes, and off-centering (bias) is often neglected. When working in precision assembly, where tolerance intervals are specified in the range of some micrometers, off-centering effects can no longer be neglected. Indeed, the smaller the interval, the more sensitive the yield becomes, inertial tolerancing is a methodology derived from Taguchi loss function, to specify part dimensions and assemblies so that off-centerings are taken into account. We discuss different inertia allocation methods proposed by Pillet and propose a new method where inertia are simply allocated according to possible process performances. Concerning production indicators, due to this methodology, we suggest to abandon traditional capability indices and use sigmadelta-tolerancing with square error. We benefit from sigmadelta-graphs to monitor processes and to help to visually quantify the lack of production and assembly processes. Finally, we describe an entire design procedure, that leads from product functional requirements to production specification, and apply it to an industrial case study. A key element that we provide in this procedure is a novel formulation of assembly yield in the presence of off-centering. The procedure is not single-shot but rather iterative, starting from both opposite ends : customer needs and break-even production costs

The conventional perception about the quality of the product was concentrated only around the specification limits. If the product characteristic was within the specifications, it was found to be good and accepted for use. If it did not meet the specifications, the product was rejected, or reworked and used. This is the concept of step function. Taguchi’s quality loss function, which is quadratic in nature, evaluates the quality even when the product characteristic is within the specification. This paper discusses the application of Taguchi’s quality loss function to evaluate the quality loss during spline hobbing operation of a shaft manufacturing process. The quality characteristic considered here is fraction defective type. The fraction defective data has been collected from five industries manufacturing the same kind of shaft; Taguchi’s loss function has been applied and average quality loss has been calculated for each of the industries. The loss function graphs are constructed for these industries for the purpose of comparison. Using the loss function, the industries having a comparatively larger quality loss are identified and necessary action has been suggested to improve their product quality.

In this paper, Taguchi's quadratic loss function is incorporated in the economic design of the control chart. This is done by redefining the in-control and out-of-control costs using Taguchi's loss function in the general model for the economic design of x macr -control charts developed by Banerjee and Rahim [1]. Both cases of increasing hazard rate (Weibull failure rate) and constant hazard rate (exponential distribution) will be presented and followed by numerical examples. Finally, sensitivity analysis is conducted to study the effect of important parameters on the cost. Suggestions for further research are also included.

Taguchi defined product quality in a universal way by adopting a quadratic quality loss function. In 1987, Kapur and Wang addressed the problem of setting economic specification limits for processes being not capable to meet given tolerances and, therefore, product quality has to be improved by inspection. Kapur and Wang selected for their investigations Taguchi's quadratic loss function. In this paper, we assume a asymmetrical linear quality loss function and show that the obtained optimal values differ greatly from those obtained by Kapur and Wang.

The functional performance of an assembled product depends upon its component tolerance and assembly clearance. For producing components with tight tolerance, advanced machining process may be required but it will increase the product cost. In any manufacturing process, final step is assembly and plays a vital role in determining the product quality. The selective assembly is a method of assembly that provides an effective way for producing high precision product from low precision components. The conventional Taguchi's loss function targets on a single value but in real situations an interval of values will be essential. To overcome this issue, a symmetrical interval-based Taguchi (SIT) loss function is proposed in this paper and it is applied into the selective assembly method to evaluate the assembly loss. Also, an improved sheep flock heredity (ISFH) algorithm is proposed to obtain the best combination of selective group with minimum clearance variation and least assembly loss value.

In this paper, Duncan's cost model combined Taguchi's quadratic loss function is applied to develop the economic-statistical design of the sum of squares exponentially weighted moving average (SS-EWMA) chart. The genetic algorithm is applied to search for the optimal decision variables of SS-EWMA chart such that the expected cost is minimized. Sensitivity analysis reveals that the optimal sample size and sampling interval decrease; optimal smoothing constant and control limit increase as the mean and/or variance increases. Moreover, the combination of optimal parameter levels in orthogonal array experiment plays an important guideline for monitoring the process mean and/or variance.

In this paper we incorporate Taguchi's loss function in the economic design of the control chart. This is done by redefining the in‐control and out‐of‐control costs using Taguchi's loss function. The new model brings Taguchi's off‐line concepts to the classical SPC approach and continues to have the advantages of the economic design model by taking into consideration the cost consequences of the design. Sensitivity analysis with respect to important model parameters is discussed.

This paper takes over the common dilemma facing a consumer receiving a lot from a supplier with unavailable information about the supplier's process level, or the information being available but untrustworthy or uncertain. This paper aims to model and optimize an economical single sampling plan that is independent of the supplier's process level, where the loss caused by accepting low quality lots is treated as a Taguchi's loss function; the model also considers inspection cost, and replacement cost. The Taguchi's loss function in this paper is a function of the expected percent defect in the accepted lots, which later through standardizing the Operating Characteristic (OC) curve becomes a function of the sample size n, and the defectives rejection limit c achieving independence from the supplier's process level. The standardization is attained through mathematical estimation and use of the beta function properties; the reliability associated with using the expectation is assessed later through the variance. The optimization technique used to find the value of n and c that minimizes the total cost associated with this sampling plan is direct search since both variables are discrete and bounded by the lot size.

ABSTRACTAlthough the normality is a usual assumption for quality measurements, it may be erroneous and misleading in many industrial applications of quality control online models. In addition, incorporating Taguchi's loss function to these inspection activities leads the management to better decisions. In this paper, the integrated model of Taguchi's loss function and the economic statistical design of -control charts are investigated under both normal and non normal quality data. This purpose has been achieved by modifying the cost model of Banerjee and Rahim (1988) in which a non uniform sampling scheme is utilized for systems with increasing failure rate and Weibull shock model. For the same values of time and cost quantities and Weibull distribution parameters, the effects of three most popular quality characteristic distributions (Normal, Burr, and Johnson) that are fitted on a case study data in Rahim (1993) are studied and compared. The optimal parameters of both economic and economic statistical designs reveal that there is an insignificant difference between Normal and Burr distributions, whereas a relative difference can be observed in the case of Johnson distribution.

The accurate determination of production duration remains challenging in real‐time batch operations, making it difficult to optimize the termination time with minimal time and material consumption. To enhance batch profitability, this work develops a machine learning‐based approach for online batch endpoint identification. In this framework, quality‐related variables are used to represent the batch duration, and a data‐driven soft sensor is established to estimate these variables. Given the complex nonlinear dynamics and significant inter‐batch variability in batch processes, a dedicated recurrent neural network is designed with a time‐weighted loss function to improve prediction accuracy in the later stages of production. Furthermore, a decision‐making strategy combined with a Savitzky–Golay (S–G) filter is introduced to robustly identify the optimal endpoint for each batch. The proposed method is validated through an industrial‐scale penicillin fermentation process. Results demonstrate that the batch operation can be terminated approximately 58.5 h earlier on average, leading to a 27.68% increase in productivity and saving about 10,157 RMB in material costs per batch.

Recognition for SAR deformation military target from a new MiniSAR dataset using multi-view joint transformer approach

New American College of Chest Physicians Lung Cancer Guidelines*: An Important Addition to the Lung Cancer Guidelines Armamentarium

Purpose of this study is to define a novel process capability analysis regarding to loss function. Loss functions such as Taguchi loss functions, assess the process losses, but this function do not distinguish real harmful loss. So in this study an improved Taguchi loss function method is constructed. Then regarding to this new function a novel process capability is developed. The logical idea is to use improved Taguchi quality loss function in order to measure the process capability. Main findings are improved Taguchi loss function that is more sensitive to losses and new process capability index base on improved Taguchi loss function, for qualitative response data, to evaluate the process capability for the qualitative response data.