Yoke San Wong

A dynamic model for managing overlapped iterative product development

Intense competition in many industries impels firms to develop more products in less time. Overlapping of development activities is regarded as one of the most promising strategies to reduce project cycle time. However, the gain from overlapping must be weighed against the additional resource and time for rework. This paper presents a new product development (NPD) process model, termed Dynamic Development Process Model (DDPM), for managing overlapped iterative product development. We validated the model with data from a mobile phone development project. The DDPM was employed to identify appropriate policies for the overlapped iterative projects in the case study company. These identified policies were implemented in the company and led to marked improvement in project performance, thus demonstrating the viability of our model.

Optimal overlapping and functional interaction in product development

Overlapping of development stages and interaction between different functions are regarded as important strategies for reducing development lead time. However, overlapping typically requires additional costs for rework and functional interaction increases communication time. This paper presents an analytical model to improve project performance by balancing the positive and negative effects of overlapping and functional interaction. We first investigate the progress of downstream development, which is essential to derive the optimal overlapping policies. We find that the downstream progress increases over time when the upstream evolution is fast or linear, but it is indefinite when the upstream evolution is slow. Then, we present optimal overlapping policies taking into account the complexity of downstream progress. The impact of different project properties, such as the dependency between development stages and the opportunity cost of time, on overlapping policies is discussed. Finally, we derive the optimal functional interaction strategy when the optimal overlapping is followed. The methodology is illustrated with a case study at a handset design company.

A Hybrid Cutting Force Model for High-speed Milling of Titanium Alloys

Abstract In this paper, the Johnson-Cook (JC) strength model is used to describe the flow stress of Ti6AI4V and to estimate two important parameters in Oxleys model: the strain-rate constant and the angle made by the resultant force and the shear plane. The JC model is also incorporated into a finite element method (FEM) simulation for the deformation process of T16AI4V. Finally, a hybrid cutting force model based on the FEM simulation and Oxleys theory is proposed to predict cutting forces when machining Ti6AI4V. Experimental results are found to substantiate the developed model.

A diffusion wavelet approach for 3-D model matching

This paper proposes a new 3D shape retrieval approach based on diffusion wavelets which generalize wavelet analysis and associated signal processing techniques to functions on manifolds and graphs. Unlike current works on 3D matching, which are based either on the topological information of the model or its scatter point distribution information, this approach uses both information for more effective matching. Diffusion wavelets enable both global and local analyses on graphs, and can capture the topology of a surface with the diffusion map of its mesh representation. As a result, both multi-scale properties of the 3D geometric model and the topology among the meshes can be extracted for use in 3D geometric model retrieval. Tests using 3D benchmarks demonstrate that the approach based on diffusion wavelets is effective and performs better than those by spherical wavelet and spherical harmonics in 3D model matching.

A multi-objective disassembly planning approach with ant colony optimization algorithm

This paper proposes a multi-objective disassembly planning approach with an ant colony optimization algorithm. The mechanism of ant colony optimization in disassembly planning is discussed, and the objectives to be optimized in disassembly planning are analysed. In order to allow a more effective search for feasible non-dominated solutions, a multi-objective searching algorithm with uniform design is investigated to guide the ants searching the routes along the uniformly scattered directions towards the Pareto frontier; based on the above searching algorithm, an ant colony optimization algorithm for disassembly planning is developed. The results of a case study are given to verify the proposed approach.

A Comprehensive Identification of Tool Failure and Chatter Using a Parallel Multi-ART2 Neural Network

Tool failure and chatter are two major problems during machining. To detect and distinguish the occurrences of these two abnormal conditions, a novel parallel multi-ART2 neural network has been developed. An advantage of this network is more reliable identification of a variety of complex patterns. This is due to the sharing of multi-input feature information by its multiple ART2 subnetworks which allow for finer vigilance thresholds. Using the maximum frequency-band coherence function of two acceleration signals and the relative weighted frequency-band power ratio of an acoustic emission signal as input feature information, the network has been found to identify various tool failure and chatter states in turning operations with a total of 96.4% success rate over a wide range of cutting conditions, compared to that of 80.4% obtainable with the single-ART2 neural network.

3D CAD model retrieval with perturbed Laplacian spectra

This paper presents a novel approach to the 3D CAD model retrieval, whereby the 3D models are treated and matched as undirected graphs. While there is much success made in the matching of graphs based on their spectral decomposition, most of these approaches consider smooth surfaces and are not suitable for CAD models because of their complex topology and singular structure. In the proposed approach, the models are simplified based on the piecewise flat properties of the surfaces first, and a perturbed Laplacian spectrum approach is then applied to characterize the shape. These spectral values are used as samples for spectral distribution estimation. The perturbed spectral distributions of different models are then compared by their KL-divergence for model retrieval. The proposed approach is tested with models from known 3D CAD database for verification.

A manufacturing-oriented approach for multi-platforming product family design with modified genetic algorithm

With highly fragmented market and increased competition, platform-based product family design has been recognized as an effective method to construct a product line that satisfies diverse customer’s demands while aiming to keep design and production cost-effective. The success of the resulting product family often relies on properly resolving the inherent tradeoff between commonality across the family and performance loss. In this paper, a systematic multi-platforming product family approach is proposed to design a scale-based product family. In the light of the basic premise that increased commonality implies enhanced manufacturing efficiency, we present an effective platform decision strategy to quantify family design configuration using a commonality index that couples design varieties with production variation. Meanwhile, unlike many existing methods that assume a single given platform configuration, the proposed method addresses the multi-platforming configuration across the family, and can generate alternative product family solutions with different levels of commonality. A modified genetic algorithm is developed to solve the aggregated multiobjective optimization problem and an industrial example of a planetary gear train for drills is given to demonstrate the proposed method.

An interpolation scheme for tool-radius compensated parabolic paths for CNC

In general, NC systems provide only linear and circular interpolations. Complex shapes to be machined are first approximated with linear and circular segments to some predetermined tolerance. Parabolic curves are more suitable for piecewise approximation of higher-order curves because less segments than lines or circular arcs are required. In machining operations, the tool motion controlled by the interpolator does not move along the profile of the job. The tool path is offset from the profile by the tool radius. Tool-radius compensations for linear and circular tool paths involve simple offsets of the initial and final points. The tool paths remain linear or circular. For parabolic profiles, the offset tool paths are not parabolic; hence, parabolic interpolations cannot be used to generate the offset paths. This paper presents an algorithm to perform interpolations for the tool to follow parabolic curves as well as the offset paths passing through two user-defined end points. The parabolic interpolation ...

Towards enhancement of machinability data by multiple regression

Abstract An investigation of various multiple regression model-building techniques has been carried out on machinability data in order to study the suitability of the empirical equations employed. A comparative analysis of the full-form first-order regression mode and quadratic regression model has been performed also, to aim for a parsimonious and easily interpretable model to be used for the computation of various aspiration levels, which latter include the minimum unit production cost and the maximum unit production rate. In order to address the problem of the update of machinability data found in current systems, the raw data of the machining responses and machining variables are retained for future update, thus enhancing the models, which relate closely to the actual production process. This module can thus be incorporated into an expert system that has been discussed elsewhere, achieving progress towards an integrated machining system.