Optimization of Manufacturing Processes

The traditional trial-and error method applied to derive empirical relation and optimize the process is time consuming and results in reduced productivity, high rejection and cost. Hence, current research in foundries focussed towards development of statistical modelling and optimization tools. The present research work is focused on modelling and optimization of Alpha-set moulding sand system. The variables such as percent of resin and hardener, and curing time will influence the sand mould properties, namely, compression strength, permeability, mould hardness, gas evolution and collapsibility. Experimental data is collected as per CCD design matrix and non-linear models have been developed for all responses. The behaviour of all responses is studied by utilizing surface plots. The statistical adequacy of all models is tested with help of ANOVA. All responses are tested for their prediction capacity with the help of test cases. The predictive non-linear models, developed for the process resulted in average deviation of less than 5%. The optimization (GA, PSO, DFA and TLBO) tools are applied to optimize the process for conflicting requirements in sand mould properties. Six case studies with different combination of weight fractions assigned to sand mould properties are considered. The optimum solution correspond to highest composite desirability value is selected. TLBO outperformed other optimization tools (i.e. GA, PSO, and DFA) while determining the highest desirability value and resulted in optimized sand mould properties. Experiments are conducted for the optimized and normal (i.e. lowest desirability) sand mould conditions. Castings are prepared by pouring molten LM20 alloy to the prepared moulds. The casting obtained for the optimized sand mould condition resulted in a better casting quality.

The results of Electrical Discharge Machining (EDM) are characterized through many parameters. These include, material removal rate, surface finish, geometrical accuracy, tool wear, and kerf width. The three main types of EDM, wire, sinker, and micro EDM all have similar characteristics in relation to input parameters and their effects on the results. The typical EDM system is too complex to accurately model the effect of all the parameters together. Therefore, it is necessary to create an optimization algorithm to predict the results of specific input parameters. Various techniques such as Taguchi robust design, grey relational analysis, desirability, genetic algorithm, and neural network etc. have been used for optimization of EDM based processes. This chapter first briefly introduces all the aforementioned optimization processes and comprehensively discusses their implementation and effect for optimization of EDM based processes.

This chapter presents analysis and optimization of machinability of Mild steel grade 350 while high speed drilling operation. Taguchi design of experiments (DoEs), analysis of variance (ANOVA) and other traditional methods were applied to optimize the input variables in order to minimise the circularity, cylindricity, diameter error and surface roughness of drilled holes. It was found that point angle was the highest contributor for the circularity, cylindricity and surface roughness of drilled holes. The circularity error was minimum at the low speed (584 rpm), low feed (0.15 mm/rev) and moderate point angle (125°). The cylindricity error of holes was minimised at the high speed (849 rpm), moderate feed (0.2 mm/rev) and moderate point angle (125°). The moderate speed, low feed and moderate point angle minimised surface roughness considerably. The interaction between speed and point angle had the maximum contribution to the diameter error of drilled holes. The diameter error was minimum at the moderate speed, low feed and moderate point angle.

Laser metal deposition process is one of the important processes of additive manufacturing technology which is used for the production of end-use parts as well as repair of worn-out high valued engineered parts. The functional performance of laser metal deposition process is greatly dependent on its process parameters; therefore, considering the type of job and nature of material, they need to be adequately optimized before a job can be successfully carried out and with the desired properties. The processing parameters that govern the laser metal deposition process include: the laser power, the scanning speed, the powder flow rate and the gas flow rate. A lot of interactions exist among these processing parameters that make the careful optimization of the processing parameters an important task. In this chapter, modelling of laser metal deposition process of metal alloys and composites is presented. The chapter consist of an in depth review of literature on this subject in the introduction (Sect. 1). Optimization of process parameters for laser metal deposition of titanium alloy is presented in Sect. 2. A case study on statistical modelling of titanium alloy composite and process parameters optimization is presented in Sect. 3. The chapter ends with the summary in Sect. 4.

Fused Deposition Modeling (FDM) is universally used 3D printing technology, to manufacture prototypes as well functional parts due to its capability to create components having any geometric complexity in shorter duration, without any specific tooling requirement or human intervention. FDM fabricated parts have found many promising application in various industries such as aerospace, automobile, medical, customizable products etc. However, the application of FDM parts has been restricted by poor mechanical performance. The mechanical properties of the FDM fabricated part are largely affected by selection of various build parameters. Optimal selection of various build parameters can help to achieve better mechanical strength. The Adaptive network-based a fuzzy Interference System (ANFIS) is uses both neural networks and fuzzy logic to generate a mapping between inputs and response. In ANFIS, the parameters for fuzzy system has been identifying using a neural network. Hybrid learning rule can be used for creating a fuzzy set of IF-THEN rules with the appropriate membership functions and generating previously defined Input/Outputs pairs. Initially, a detailed experimental investigation was conducted to understand the impact of different build parameters on the tensile strength of printed PLA. Using experimental data, an optimized model of ANFIS was developed to anticipate the tensile strength of printed parts.

Traditional machining of polymer matrix composites (PMCs) possesses difficulties as they exhibit excellent specific strength and stiffness. Superior properties led PMCs parts were extensively used in structural, aviation, construction and automotive applications. The advanced machining process abrasive water jet machining (AWJM) has been explored to machine PMCs. The AWJM factors namely abrasive grain size, working pressure, standoff distance, nozzle speed, and abrasive mass flow rate affect the final outcome of surface quality (i.e. surface roughness, SR) and productivity (i.e. material removal rate ‘MRR’ and process time ‘PT’) are studied. Taguchi L₂₇ orthogonal array of experimental design is employed for conducting practical experiments. Taguchi method limit to optimize multiple conflicting outputs (maximize: MRR, and minimize: PT and SR), simultaneously. In general, multiple outputs may have many solutions and are dependent on the tradeoff (relative importance or weights) assigned to each output. Traditional practices such as engineer judgement, expert suggestion and customer requirements may lead to local solutions (i.e. superior quality for one output, while compromising with the rest). Principal component analysis (PCA) method overcomes the said shortcomings of traditional practices and determines weight fractions for each output based on the experimental data. Multi-objective optimization on the basis of ratio analysis (MOORA), Grey relational analysis (GRA), Technique for order preference by similarity to ideal solution (TOPSIS) and Data Envelopment Analysis based Ranking (DEAR) are the four methods employed for the purpose of multi-objective optimization. MOORA, GRA and TOPSIS methodologies require assigning weight fractions for each output by the problem solver. Note that, solution accuracies vary with the weight fractions assigned to each output. The aggregate (composite values of all responses) values determined by PCA-MOORA, PCA-TOPSIS, PCA-GRA and DEAR method were used for determining optimal factor levels and their contributions. DEAR method determined optimal levels resulted in better machining quality characteristics.

This chapter explores the application of a hybrid approach namely multi-objective optimization based on ratio analysis (MOORA) in fuzzy context to obtain the best parametric combination during machining of commercially pure titanium (CP-Ti) Grade 2 with uncoated carbide inserts in dry cutting environment. A series of experiment was performed by adopting Taguchi based L₂₇ orthogonal array. Cutting speed, feed rate, and depth of cut were selected as three process variables whereas cutting force, surface roughness and flank wear were selected as three major quality attributes to be minimized. The minimization was exploited using fuzzy embedded MOORA method and hence an optimal parametric combination was attained. The results of the investigation clearly revealed that, the fuzzy coupled with MOORA method, was capable enough in acquiring the best parametric setting during turning operation under specified cutting conditions.

Multi-objectives Genetic Algorithm (MOGA) is one of many engineering optimization techniques, a guided random search method. It is suitable for solving multi-objective optimization related problems with the capability to explore the diverse regions of the solution space. Thus, it is possible to search a diverse set of solutions with more variables that can be optimized at one time. Solutions of MOGA are illustrated using the Pareto fronts. A Pareto optimal set is a set of solutions that are non-dominated solutions frontier. With the Pareto optimum set, the corresponding objective function’s values in the objective space are called the Pareto front. The conventional methods for solving multi-objective problems consist of random searches, dynamic programming, and gradient methods whereas modern heuristic methods include cognitive paradigm as artificial neural networks, simulated annealing and Lagrangian approcehes. Some of these methods are managed in finding the optimum solution, but they have tendency to take longer time to converge so that need much computing time. Thus, by implementing MOGA approach that based on the natural biological evaluation principle will be used to tackle this kind of problem. In this chapter authors attempts to provide a brief review on current and past work on MOGA application in few of the most commonly used manufacturing/machining processes. This chapter will also highlights the advantages and limitations of MOGA as compared to conventional optimization techniques.

This chapter introduces various types of manufacturing systems and different types of traditional and modern optimization techniques. Chapter briefs about evolutionary techniques such as Particle swarm optimization and Genetic algorithm to optimize the various kind of manufacturing system with an objective to overcome the limitations of traditional optimization techniques and to enhance the optimality of objective function. Besides that, customary methodology is to utilize an ordinary least squares relapse investigation for building up the machinability models. In recent decade, the utilization of evolutionary calculation techniques, or additionally called the Genetic strategies, in light of impersonation of Darwinian characteristic choice has turned out to be across the board. This is because of truth that numerous frameworks are excessively complex, making it impossible to be effectively enhanced by the utilization of traditional deterministic calculations. Despite what might be expected, the evolutionary algorithms (EA) include probabilistic tasks. The current chapter also presents brief details about stepwise procedure of implementation of genetic algorithm and particle swarm optimization to solve various problems associate with manufacturing systems.