Hemant Ramaswami

Minimum-zone form tolerance evaluation using particle swarm optimisation

Robust and accurate evaluation of form tolerances is of paramount importance in todays world of precision engineering. Present-day Coordinate Measuring Machines (CMMs) and other optical scanning machines operate at high speed and have a high degree of accuracy and repeatability which are capable of meeting the stringent measurement requirements. However, the evaluation algorithms used in conjunction with them are not robust and accurate enough, because of the highly non-linear nature of the minimum-zone form tolerance formulation. Evolutionary Algorithms (EAs) have proved effective in solving non-linear optimisation problems. In this paper, Particle Swarm Optimisation (PSO) is employed to evaluate various minimum-zone form tolerances. An unconstrained formulation of the minimum-zone form tolerance is used for the optimisation. The methodology is validated by testing on several datasets from published literature and yields equal or better results than other existing minimum-zone algorithms. It is also extremely robust and the quality of the results is not affected by the number of points in the dataset.

A multivariate statistical analysis of sampling uncertainties in geometric and dimensional errors for circular features

Traditional research on sampling strategies for circular features has focused on reducing the error in estimation of form error for determining the optimal sample size. However, circular features have more than one geometric dimensioning and tolerancing (GD&T) error (form, position, and size) that need to be estimated simultaneously using a single inspection procedure. The inspection method used should, therefore, have a good overall performance in simultaneously assessing all of the GD&T parameters of interest. In this paper, a multivariate statistical model using Exploratory Factor Analysis has been presented for simultaneously analyzing the effect of sampling strategies on the sampling uncertainty of the GD&T errors. Realistic circular profiles have been generated incorporating lobes, eccentricity, and roughness, and a minimum zone fitting algorithm has been used to determine the form error. The error in position and size are computed based on the minimum circumscribed circle of the profile, in accordance with ANSI guidelines. A six-dimensional performance metric vector quantifies the difference between the true value of the errors and the value evaluated using the sample. This vector is used as a basis to recommend a sample size and technique that performs well in simultaneously evaluating the form error, position, and size of a circular profile. The results obtained from the simulated profiles are validated using data collected from manufactured parts.

Selective Zone Search Method for Evaluation of Minimum Zone Sphericity

Accuracy, robustness and speed are essential components of every precision engineering procedure. With the availability of high speed inspection machines and the ability to generate large datasets with minimal effort and time, the evaluation algorithm becomes a critical component of the inspection time. This paper presents a new approach for evaluation of minimum zone sphericity tolerance using a selective zone search method. The method uses geometric constructs to identify the five extreme points required to generate the two minimum zone spheres. Four different models have been developed to investigate all possible combinations for the tolerance evaluation, viz. 4-1, 3-2, 2-3 and 1-4 models. The robustness of the algorithm has been tested using various simulated and reported datasets and the results have been found to be comparable to existing methods. The presented algorithm has also shown considerable savings in time compared to a nonlinear optimization formulation for minimum zone sphericity evaluation.© 2009 ASME

Minimum-Zone Circularity Evaluation Using Particle Swarm Optimization

Robust and accurate evaluation of form tolerances is of paramount importance in today’s world of precision engineering. Present-day Coordinate Measuring Machines (CMMs) operate at high speed and have a high degree of accuracy and repeatability which are capable of meeting the stringent measurement requirements. However, the evaluation algorithms used in conjunction with them are not robust and accurate enough, because of the highly non-linear nature of the minimum-zone circularity formulation. Evolutionary algorithms have proved effective in solving constrained non-linear optimization problems. In this paper, Particle Swarm Optimization (PSO), which is one of the most recent and popular evolutionary algorithms, is employed to evaluate the minimum-zone circularity. The PSO approach imitates the social behavior of organisms such as bird flocking and fish schooling. It differs from other well-known Evolutionary Algorithms (EA) in that each particle of the population, called the swarm, adjusts its trajectory toward its own previous best position, and toward the previous best position attained by any member of its topological neighborhood. The constrained nonlinear model is rewritten as an unconstrained non-linear model using the penalty-function approach. The methodology is validated by testing on several simulated and experimental datasets and yields better results than other existing minimum-zone algorithms.Copyright

Self-Organized Hebbian Inference of Environment Topology by Distributed Sensor Networks

Ad hoc wireless sensor networks are emerging as an important technology for applications such as environmental monitoring, battlefield surveillance and infrastructure security. While most research so far has focused on the network aspects of these systems (e.g., routing, scheduling, etc.), the capacity for scalable, in-field information processing is potentially their most important attribute. Networks that can infer the phenomenological structure of their environment can use this knowledge to improve both their sensing performance and their resource usage. These intelligent networks would require much less a priori design, and be truly autonomous. This paper presents a distributed algorithm for inferring the global topological connectivity of an environment through a simple self-organization algorithm based on Hebbian learning. The application considers sensors distributed over an environment with a network of tracks on which vehicles of various types move according to rules unknown to the sensor network. Each sensor infers the local topology of the track network by comparing its observations with those from neighboring sensors. The complete topology of the network emerges from the distributed fusion of these local views.

Predictive Surface Generation and Characterization for Cylindrical Turning Processes

This paper presents a generalized simulation based approach for generation and characterization of turned surfaces based on process parameters and manufacturing errors. The presented model shows that with proper analytical modeling along with appropriate process monitoring system (force signals, vibration signals, spindle motion error signals etc.,) a comprehensive surface generation model can be developed. First, the tool nose geometry and cutting-force induced vibrations are superimposed to obtain the cutting tool path. Next, the information obtained from spindle motion errors is used to analytically formulate the position of each point on the machined surface. Regression models are fit to establish the relationship between form error / surface roughness and input parameters. The simulation-based approach presented here provides a quantitative bridge between process parameters/manufacturing errors and surface characterization metrics. Such a scheme would allow manufacturing engineers to pre-select processes, parameters, and capable machines to achieve design specification. This model will allow engineers to proactively control the influence of machining parameters on product quality through computer simulation, and, thus, “do things right the first time.”Copyright

Multiple Parting Surfaces for Sand Casting

Multiple parting surfaces are frequently used in sand casting, die casting and injection molding processes. However, most research in this area has focused on die casting and injection molding. Parting surfaces for die casting and injection molding are relatively easier to compute compared to sand casting because their orientations and shapes are less restricted. In sand casting, the parting surfaces have to be parallel to each other and perfectly flat to permit the use of flasks with more than two pieces. The concepts of visibility and object illumination can be used to divide an object into two parts using a single parting surface. These methods, however, cannot be directly used for multiple parting surfaces. In this paper, a methodology to generate multiple parting surfaces for sand casting is described. The method uses Gauss maps to identify potential casting directions, and global accessibility cones to determine which faces can be cast in the same part of the pattern. The pattern is sliced using parallel planes such that each slice can be withdrawn from the mold in at least one direction. After the object is sliced, the number of parting surfaces is reduced by combining adjacent middle sections depending on their accessible directions.Copyright