Manufacturing Technology Today

Susanta Kumar Gauri ¹, Shankar Chakraborty ²

Affiliations
1 SQC & OR Unit, Indian Statistical Institute, 203, B. T. Road, Kolkata, IN
2 Dept. of Prod. Engg., Jadavpur University, Kolkata, IN

Abstract

Researchers have attempted several approaches for determination of the process settings that can optimize the multiple performance measures (responses) of wire electrical discharge machining (WEDM) operations. The VIKOR method, applied so far for multi-response optimization of chemical processes, can overcome the limitations of the multi-response signal-to-noise (MRSN) ratio based approach. In this paper, the VIKOR method is modified to make it more generalized and a set of experimental data on multiple responses of WEDM process is analyzed using the modified VIKOR method. The results demonstrate that the optimal factor-level combination determined using the VIKOR method can lead to significantly better overall quality level than the MRSN ratio based approach.

Full Text

     

Susanta Kumar Gauri ¹, Shankar Chakraborty ²

Affiliations
1 SQC and OR Unit, Indian Statistical Institute, Kolkata, IN
2 Dept. of Production Engg., Jadavpur University, Kolkata, IN

Abstract

With the introduction and increased use of newer and harder materials like titanium, stainless steel, high strength temperature resistant (HSTR) alloys, fiber-reinforced composites and ceramics in aerospace, nuclear, missile, turbine, automobile, tool and die making industries, a different class of machining processes has been emerged. Instead of employing the conventional tools, these non-traditional machining (NTM) processes use energy in its direct form to remove materials from the workpiece. To achieve the best performance of these NTM processes, it is necessary to select the machining parameters at their optimal levels. Taguchi method of robust design has been extensively used to choose the optimal parametric levels in various machining processes. This paper exclusively reviews the applications of the Taguchi method adopted to select the optimal factor level combinations in different NTM processes.

Full Text

     

Susanta Kumar Gauri ¹, Shankar Chakraborty ²

Affiliations
1 SQC & OR Unit, Indian Statistical Institute, 203, B. T. Road, Kolkata-700108, IN
2 Dept. of Production Engineering, Jadavpur University, Kolkata-700032, IN

Abstract

Control charts usually exhibit one of the eight types of patterns. These patterns can be classified as normal and abnormal. Recognition of abnormal patterns in control charts can provide clues to reveal potential quality problems in the manufacturing processes. Neural network approaches (with features extracted from the pattern data as input vector representation) have been successfully applied by the researchers in recent years for recognition of control chart patterns. Usage of features leads to smaller network size and results in faster training and generally more effective and efficient recognition of control chan patterns. The reported feature-based approaches can only recognize six principal control chart patterns (CCPs). In this paper a new set of features is proposed and a multilayered perceptron (MLP) neural network trained by back-propagation algorithm is presented that can recognize stratification and systematic patterns in addition to the other six patterns as mentioned above. Extensive performance evaluation of the developed pattern recognizer is carried out using simulated data. Numerical results indicate that the artificial neurai network based pattern recognizer developed using the proposed set of features can perform well in real time process control applications.

Full Text

     

Susanta Kumar Gauri ¹, Shankar Chakraborty ²

Affiliations
1 SQC & OR Unit, Indian Statistical Institute, 203, Kolkata-700 108, IN
2 Department of Production Engineering, Jadavpur University, Kolkata-700 032, IN

Abstract

The time consumed from loading of ingredients into a furnace to the completion of pouring of melt in the molds is known as the heat length and a batch of production is called as heat. In a single heat, different types of castings are produced when the required raw material composition for all these castings are similar. If sufficient numbers of molds for different types of castings that can consume the entire melt quantity are not ready after packing, heat length increases resulting in lesser number of heats per day. However, deployment of too many resources for packing incurs unnecessary cost. In this context, determination of the optimal quantities of different types of castings to be produced in a heat (known as product-mix) and allocation of their molds for packing at different locations is very important. The optimal product-mix will ensure that the leftover quantity of melt in the furnace will be minimum and shortages in the delivered quantities as well as overproduction of different types of castings can be avoided. In this paper, a real time problem of optimal product-mix determination for an Indian foundry is discussed and Its solution Is presented. The expected annual tangible gain from the increased productivity achieved through the deployment of the optimal product-mix of molds for packing Is estimated to be around rupees 11.14 lakhs/annum.

Full Text

     

Susanta Kumar Gauri ¹, Shankar Chakraborty ²

Affiliations
1 SQC & OR Unit, Indian Statistical Institute, 203, B. T. Road, Kolkata-700108, IN
2 Dept. of Production Engineering, Jadavpur University, Kolkata-700032, IN

Abstract

Accurate monitoring and control of the manufacturing processes has become very important In today's dynamic world due to rapid Increase In demand for highly precise end products. For every process, there is a target and the goal of statistical process control (SPC) is to ensure that the process mean lies nearer to the target value without inflation o f process variability. Control charts are widely used to identify the situations when control actions will be needed for the manufacturing processes. Control charts usually exhibit one of the eight basic types of patterns. Identification of these patterns leads to more focused diagnosis and significantly minimizes the effort towards troubleshooting. Pham and Wani presented a feature-based heuristic approach for control chart pattern (CCP) recognition, which has been very appealing to the shop-floor people, because In this approach, the practitioners can clearly understand how a particular pattern has been identified by the use o f relevant features. The heuristics proposed by Pham and Wani can only differentiate six types of CCPs based on extraction of nine features. In this paper, a new set of nine features is proposed and the heuristics for CCP recognition based on these features is also presented that can efficiently differentiate all the eight basic types of CCPs. The features are chosen such that the need for human intervention for their extraction is eliminated and thus the CCP recognition is totally automatic. The proposed feature-based CCP recognition approach can be applicable to any process operating on the known target value and will be considerably robust against the variation of true process standard deviation from its estimated value. Thus it has enough potential for use in real-time process control applications.

Full Text