Abstract
Throughout history, gears have been used in many applications, however, it has only been in the last 160 years that manufacturing techniques and materials have been studied to optimize processes to yield smaller, lighter yet stronger and more cost-effective gears. One of the critical items for mechanical power transmission are gears in various mechanical systems. Gear optimization plays a vital role in gear design and manufacturing in relation to the reduction of weight and volume from a design perspective and reduction of surface irregularities i.e. micro-geometry errors and surface roughness, and improvement in productivity from a manufacturing perspective. Miniature gears are essential components in many devices such as timer mechanism, miniature pumps, harmonic devices, and miniature robots etc., and determine their functional performance characteristics. Gear transmission error is significant cause of dynamic excitation through complicated transfer function which leads to noise production in the powertrain system. High surface irregularity contributes to transmission error, noise, and gear failure. Both conventional and advanced manufacturing processes are used to manufacture gears. Process optimization is one of the important research and innovation strategies by which the best product quality and process performance can be ensured. In this work, two advanced machining processes i.e. wire electric discharge machining (WEDM) and laser machining have been optimized for manufacturing of miniature gears to achieve the best gear quality for high performance characteristics and high machining (or gear cutting) rate for the best productivity. Considering the effectiveness of the Fuzzy-MOORA, Fuzzy-TOPSIS, DEAR, GRA and SRM optimization techniques for solving multi-criteria decision making problems with interests’ conflict in nature, their effectiveness for gear manufacturing has been evaluated in this work.
WEDM optimization was performed for the minimization of micro-geometry errors, that is total profile error (Fa) and accumulated pitch error (Fp) and surface roughness that is average surface roughness (Ra) and maximum surface roughness (Rt), and the maximization of gear cutting rate (GCR). Multi-objective optimization considering the surface roughness parameters simultaneously with the micro-geometry and GCR responses was done using Fuzzy-MOORA, Fuzzy-TOPSIS, DEAR, GRA and SRM approaches. The current study provides multi-objective optimization with conflicting objectives ranging from the maximization of the process productivity of WEDM which is the gear cutting rate, the minimization of micro-geometry errors and surface roughness. There is a distinct set of WEDM parameters for each response with its optimized value. Any one set of parameters was therefore not able to yield minimum surface roughness and micro-geometry errors with the highest productivity. As such, to fulfil the criteria simultaneously, a fair trade-off was necessitated.
Optimal factor levels for WEDM process were found through the application of multi-objective optimization tools i.e. DEAR, GRA and SRM. It was observed that the DEAR and GRA although gave the same results, the optimal factor levels are different to those obtained by Fuzzy-MOORA and Fuzzy-TOPSIS. The difference could have been due to the varying procedure to find composite responses. The DEAR and GRA methods outperformed the hybrid Fuzzy-MOORA and Fuzzy-TOPSIS methods in finding the optimal levels that give the required values of minimum micro-geometry errors and surface roughness and maximum value
vii
of gear cutting rate. The identified performance factors herein are of paramount significance in that they influence both the production rate and cost to a notable extent. The non-specificity of machining characteristics and interaction effects of the chosen process variables has rendered the process of determining optimal parametric combinations challenging. Thus, the fuzzy set theory was proposed for application to perform an effective assessment of the machining characteristics of interest.
The laser machining optimization is done for the minimization of mean roughness depth (Rz), average surface roughness (Ra), dimensional deviation (DD), and the maximization of material removal rate (MRR) which is the indictor of process productivity. The multi-objective optimization has been attempted using Fuzzy-MOORA, Fuzzy-TOPSIS, DEAR, GRA and SRM approaches, as in the WEDM process. Similarly, the laser machining process with conflicting objectives among the process productivity, surface roughness and dimensional deviation, required multi-objective optimization of process parameters. There are different sets of laser machining parameters for each response with its optimized value where no one set of parameters was able to give minimum surface roughness and dimensional deviation with the highest productivity. As in the case of WEDM, to ensure that all the criteria simultaneously, a fair trade-off was inevitable.
Optimal factor levels for laser machining process were determined by the applying of multi-objective optimization tools that is DEAR, GRA and SRM. It was discovered that the DEAR and GRA although gave the same results, the optimal factor levels are different to those obtained by Fuzzy-MOORA and Fuzzy-TOPSIS as observed through WEDM. The DEAR and GRA methods still outperformed the hybrid Fuzzy-MOORA and Fuzzy-TOPSIS methods in this case also by determining the optimal levels that give the required values of minimum dimensional deviation and surface roughness and maximum value of material removal rate. The chosen performance characteristics here are of vital significance in that they influence both the production rate and cost to a notable extent. The non-specific machining characteristics and interaction effects of the laser machining process variables has rendered the process of determining optimal parametric combinations difficult.
The study reveals that the non-hybrid MCDM techniques DEAR and GRA (in other sense they are hybrid as they were applied with signal-to-noise ratio) are superior to the hybrid MCDM techniques Fuzzy-MOORA and Fuzzy-TOPSIS through examination of the experimental results given by the optimal parametric combinations of input parameters determined through these methods. The SRM method stood alone by giving optimal factor level of input parameters beyond the range of experiments carried out for WEDM and laser machining.