Loss Function: an Analytical Tool of Taguchi Techniques For Quality Engineering |
The Taguchi philosophy is a new point of view in thestatistical decision instead of classical dichotomic type of decision good/bad,with the principal goal of improving quality of products in the industrialpractice. In the present paper are introduced some new adaptive models ofTaguchi’s loss functions using statistical software. We propose quality losslaws, which are symmetrical and asymmetrical distributions, and therefore moreadequate models, which give a better approximation in the real world. This paper presents a new approach to determiningeconomical values for manufacturer lower and upper specification limits using anewly derived hybrid function for expected cost. The identification and use ofspecification limits are essential in protecting the producer from shippingdefective products that pass unnoticed due to measurement error. This costfunction is composed of four parts: a generic Taguchi loss function, a functionfor rework cost, a function for scrap cost, and a function that describes avariance to cost tradeoff. The expected cost function does not assume theprocess mean to be equal to the target value of the process nor is itrestricted to being symmetrical. The cost function is implemented amongdifferent production systems and optimal values of manufacturer lower and upperspecification limits are determined for each system. In this study, the Taguchi method is utilized to discoverthe optimal cutting parameters in turning operations. The orthogonal array, thesignal-to-noise proportion, and investigation of fluctuation are utilized tostudy the execution qualities in turning operations of AISI 1030 steel barsutilizing Tin covered tools. The model was created at first for unidiametercase and afterward acclimates to other workpiece geometries. An Adaptive NeuroFuzzy Inference System (ANFIS) is proposed in this paper to control aconsistent cutting power turning process under different cutting conditions.The ANFIS comprises of two parts: indicator and the fuzzy rationale controller.The step size of the indicator and the scaling components of the fuzzycontroller are balanced for guaranteeing soundness and acquiring optimalcontrol exhibitions. The Taguchi-genetic method is connected in this paper tohunt down the optimal control parameters of both the indicator and the fuzzycontroller such that the ANFIS controller is an optimal controller. PCreenactments are performed to check the viability of the above optimal fuzzycontrol plan composed by the Taguchi-genetic method.