Pandu Ranga Vundavilli

Fuzzy logic-based expert system for prediction of depth of cut in abrasive water jet machining process

The development of an expert system for abrasive water jet machining (AWJM) process is considered in the present work. The expert system has been developed by using fuzzy logic (FL). It is to be noted that the performance of AWJM in terms of depth of cut depends on various process parameters, such as diameter of focusing nozzle, water pressure, abrasive mass flow rate and jet traverse speed. Three approaches have been developed to predict the depth of cut in AWJM using FL system. The first Approach deals with the construction of Mamdani-based fuzzy logic system. It is important to note that the performance of the FL depends on its knowledge base. In Approach 2, the data base and rule base of the FL-system are optimized, whereas in the third Approach, the total FL-system is evolved automatically. A binary-coded genetic algorithm has been used for the said purpose. The developed expert system eliminates the need of extensive experimental work, to select the most influential AWJM parameters on the depth of cut. The performances of the developed FL-systems have been tested to predict the depth of cut in AWJM process with the help of test cases. The prediction accuracy of the automatic FL-system (i.e. Approach 3) is found to be better than the other two approaches.

Soft computing-based gait planners for a dynamically balanced biped robot negotiating sloping surfaces

Dynamically balanced gait generation problems of a biped robot moving up and down the sloping surface have been solved utilizing soft computing-based approaches. The gait generation problem of a biped robot is difficult to model due to its inherent complexity, imprecision in the collected data of the environment, which are the characteristics that can be the best modeled using soft computing. Two different approaches, namely genetic-neural (GA-NN) and genetic-fuzzy (GA-FLC) systems have been developed to solve the ascending and descending gait generation problems of a two-legged robot negotiating the sloping surface. Two modules of neural network (NN)/fuzzy logic controller (FLC) have been used to model the gait generation problem of a biped robot using the GA-NN/GA-FLC system. The weights of the NNs in the GA-NN and knowledge bases of the FLCs in the GA-FLC systems are optimized offline, utilizing a genetic algorithm (GA). Once the GA-NN/GA-FLC system is optimized, it will be able to generate the dynamically balanced gaits of the two-legged robot in the optimal sense.

Dynamically balanced optimal gaits of a ditch-crossing biped robot

This paper deals with the generation of dynamically balanced gaits of a ditch-crossing biped robot having seven degrees of freedom (DOFs). Three different approaches, namely analytical, neural network (NN)-based and fuzzy logic (FL)-based, have been developed to solve the said problem. The former deals with the analytical modeling of the ditch-crossing gait of a biped robot, whereas the latter two approaches aim to maximize the dynamic balance margin of the robot and minimize the power consumption during locomotion, after satisfying a constraint stating that the changes of joint torques should lie within a pre-specified value to ensure its smooth walking. It is to be noted that the power consumption and dynamic balance of the robot are also dependent on the position of the masses on various links and the trajectory followed by the hip joint. A genetic algorithm (GA) is used to provide training off-line, to the NN-based and FL-based gait planners developed. Once optimized, the planners will be able to generate the optimal gaits on-line. Both the NN-based and FL-based gait planners are able to generate more balanced gaits and that, too, at the cost of lower power consumption compared to those yielded by the analytical approach. The NN-based and FL-based approaches are found to be more adaptive compared to the other approach in generating the gaits of the biped robot.

DYNAMICALLY BALANCED ASCENDING AND DESCENDING GAITS OF A TWO-LEGGED ROBOT

The present paper deals with dynamically balanced ascending and descending gait generations of a 7 DOF biped robot negotiating a staircase. During navigation, the foot of the swing leg is assumed to follow a trajectory, after ensuring its kinematic constraints. Dynamic balance margin of the gaits are calculated by using the concept of zero-moment point (ZMP). In the present work, an approach different from the well-known semi-inverse method has been developed for trunk motion generation, in which it is initially generated based on static balance and then checked for its dynamic balance. The joint torques are determined utilizing the Lagrange–Euler formulation, and the average power consumption at each joint is calculated. Moreover, variations of the dynamic balance margin are studied for both the ascending as well as descending gaits of the biped robot. Average dynamic balance margin and average power consumption in the ascending gait are found to be more than that of the descending gait. The effect of trunk mass on the dynamic balance margin and average power consumption for both the ascending and descending gaits are studied. The dynamic balance margin and average power consumption are found to decrease and increase, respectively with the increase in the trunk mass.

Modeling and Analysis of Resin Bonded Sand Mould System Using Design of Experiments and Central Composite Design

Abstract In this paper an attempt has been made for linear and non linear modeling of resin bonded sand mould system using full factorial design of experiments and response surface methodology, respectively. It is important to note that the quality of castings produced using the resin bonded sand mould system depends largely on properties of moulds, which are influenced by the characteristics of sand, like type of sand, grain fineness number, grain size distribution and quantity and type of resin, catalyst, curing time etc. In the present study, percentage of resin, percentage of hardener, number of strokes and curing time are considered as input parameters and the mould properties, such as compression strength, shear strength, tensile strength and permeability are treated as responses. In the present work, phenol formaldehyde is used as the resin whereas tetrahydrophthalic anhydride as the hardener. A two level full factorial and three level central composite designs are utilized to develop input-output relationships. Surface plots and main effects plots are used to study the effects of amount of resign, amount of hardener, number of strokes and curing time on the responses, namely, compression strength, tensile strength, shear strength and permeability. Moreover, the adequacies of the developed models have been tested using analysis of variance. The prediction accuracy of the developed models have been tested with the help of twenty test cases and found reasonably good accuracy.

Multi-Objective Optimization of Squeeze Casting Process using Evolutionary Algorithms

The present work focuses on determining optimum squeeze casting process parameters using evolutionary algorithms. Evolutionary algorithms, such as genetic algorithm, particle swarm optimization, and multi objective particle swarm optimization based on crowing distance mechanism, have been used to determine the process variable combinations for the multiple objective functions. In multi-objective optimization, there are no single optimal process variable combination due to conflicting nature of objective functions. Four cases have been considered after assigning different combination of weights to the individual objective function based on the user importance. Confirmation tests have been conducted for the recommended process variable combinations obtained by genetic algorithm GA, particle swarm optimization PSO, and multiple objective particle swarm optimization based on crowing distance MOPSO-CD. The performance of PSO is found to be comparable with that of GA for identifying optimal process variable combinations. However, PSO outperformed GA with regard to computation time.

Weighted Average-Based Multi-Objective Optimization of Tube Spinning Process using Non-Traditional Optimization Techniques

Tube spinning is an effective process for producing long and thin walled tubes. It is important to note that the quality of parts produced in tube spinning process, namely internal surface roughness, external surface roughness, change in diameter and change in thickness depends on the right combination of input process parameters, such as mandrel rotational speed, feed rate of rollers, percentage of thickness reduction, initial thickness, solution treatment time and ageing treatment time. As the 2024 aluminum tube spinning process contains four objectives, it is very difficult to achieve a set of optimal combination of input process parameters that produce best quality product. This paper presents a weighted average-based multi-objective optimization of tube spinning process using non-traditional optimization techniques, namely genetic algorithm, particle swarm optimization and differential evolution. Multiple regression equations developed between the control factors and responses have been considered for optimization.

Design of genetic fuzzy system for forward and reverse mapping of green sand mould system

Abstract A genetic fuzzy system has been developed to solve the forward and reverse mapping problems of green sand mould systems. The performance of a fuzzy logic (FL) system depends on its knowledge base, which consists of a database and a rule base. A binary coded genetic algorithm (GA) has been used to optimise the knowledge base for the FL based approaches. Two approaches have been developed for each model (i.e. forward and reverse modelling). In the first approach, a manually compiled database and rule base of the FL system are optimised by GA, whereas in the second approach, the GA is used to evolve the optimal FL system automatically. The membership function distributions of the FL system are assumed to be asymmetric triangular. The first approach is found to perform better than the latter in terms of accuracy in prediction of the responses.

Neural network-based expert system for modeling of tube spinning process

The present paper deals with the development of neural network (NN)-based expert system for modeling of 2024 aluminum tube spinning process. Tube spinning is a highly nonlinear thermo-mechanical process for producing large-diameter thin-walled shapes. It is interesting to note that the performance of the process depends on various process parameters, such as wall thickness, percentage of thickness reduction, feed rate, mandrel rotational speed, solution treatment time and aging time. Therefore, an NN-based expert system is necessary for modeling the tube spinning process. The input layer of NN consists of six neurons corresponding to the inputs of the tube spinning process. Moreover, the output layer consists of four neurons that represent four responses, namely change in diameter, change in thickness, inner and outer surface roughness. It is to be noted that the performance of NN depends on various factors, such as number of neurons in the hidden layer, coefficients of transfer functions and connecting weights, etc. In the present paper, three algorithms, such as back-propagation, genetic and artificial bee colony algorithms, are used for optimizing the said variables of NN. Further, the developed approaches are tested for their accuracy in prediction with the help of some test cases and found to model the tube spinning process effectively.

Modeling of ECDM micro-drilling process using GA- and PSO-trained radial basis function neural network

Electrochemical discharge machining (ECDM) is a non-traditional manufacturing process potentially used to machine electrically non-conductive materials, such as ceramics and glass. The present paper explains the modeling of multi-input–multi-output ECDM micro-drilling of silicon nitride ceramics using radial basis function neural network (RBFNN). To establish the model, the process parameters such as applied voltage, electrolyte concentration and inter-electrode gap are treated as inputs and the important machining criteria namely material removal rate, radial overcut and heat affected zone are considered as outputs. A batch mode of training has been implemented to tune the developed RBFNN by utilizing a genetic algorithm (GA) and particle swarm optimization (PSO) methods, separately. Once, the optimal RBFNN is obtained, the performances of GA-trained RfBFNN (GA-RBFNN) and PSO-trained RBFNN (PSO-RBFNN) are compared with the help of experimental test cases. It has been observed that PSO-RBFNN is found to perform marginally better than GA-RBFNN.