Multi-objective design optimization of a high efficiency and low noise blower unit of a car air-conditioner
A multi-objective design optimization system, consisting of a Kriging-assisted genetic algorithm and a self-organizing map, is used to optimize a car air-conditioner blower unit towards high efficiency and low noise. Viscovery SOMine is used to visualize the design space and to analyze trade-offs between different Pareto-optima.
Multi-objective optimization for resin transfer molding process
Multi-point injection resin-transfer molding is optimized by applying finite element analysis and a genetic algorithm. Viscovery SOMine is used to visualize trade-offs and select favorable Pareto solutions for gate positions on flat-plate and rib structure models.
Clustering approach for multidisciplinary optimum design of cross-linked polymer
Chemical and mechanical properties of epoxy resin are optimized by molecular-dynamics simulations and data mining. Viscovery SOMine is used to visualize the complex structured polymers and find clusters of favorable properties.
Design knowledge exploration for the conceptual design of tire contours
Multi-objective design exploration is used to optimize the contour design of fuel-efficient car tires. Viscovery SOMine is used to visualize Pareto solutions obtained by a genetic algorithm in design space and analyze the interdependencies between design parameters and objective functions.
A humid electronic nose based on pulse voltammetry: A proof-of-concept design
A humid electronic nose consisting of four metal electrodes in contact with a fabric mesh damped with a NaCl solution is developed. Viscovery SOMine is used to analyze the electrochemical data obtained in experiments with different scent-emitting substances and provide a model to classify them.
TNT detection using a voltammetric electronic tongue based on neural networks
A voltammetric electronic tongue consisting of several metal electrodes is developed and used to detect TNT in an acetonitrile-water mixture. Viscovery SOMine and feed-forward neural networks are used to detect TNT traces and to design the electronic tongue for optimal detection properties.
A reuse design decision support system based on self-organizing maps
This article aims at a decision support system for design reuse. Viscovery SOMine is used to create a model of existing hood lock systems of an automotive supplier. By applying the model to specifications for a new hood lock system, the most similar existing solutions can be identified and their visualized properties can be used to support design decisions.
Composite panel damage detection using ultrasonic testing and neural networks
The focus of this article is to create a structural health-monitoring system for composite components used in modern aircrafts and wind turbines. Viscovery SOMine is used to analyze the high dimensional ultrasonic sound measurements obtained during operation or from different testing methods to detect structural damages.
Clustering of pressure fluctuation data using self-organizing maps
The pressure fluctuations in a microchannel traversed by an aqueous liquid and an organic liquid are analyzed with respect to the operating conditions of the microchannel. Viscovery SOMine is used in a multi-level approach to first cluster the time series of pressure measurements and then to order the operating conditions according to the distribution of the associated time series in the cluster model. This procedure is repeated once to restrict the model to the interesting cases.
Application of hybrid evolutionary algorithms to low exhaust emission diesel engine design
A diesel engine is optimized with a hybrid algorithm using particle swarm optimization and a genetic algorithm. Viscovery SOMine is used to analyze the relationship between exhaust characteristic and combustion chamber geometry.
Making sense of sensor data
This article examines tools and algorithms for analyzing sensor data. Clustering and exploration with Viscovery SOMine is compared to principle component analysis, singular value decomposition and other techniques.
Classification of metal ions according to their complexing properties: a data-driven approach
Factor, cluster and self-organizing map analyses are applied to the stability constants of complexes of metal ions and hydrogen with 3960 ligands. Both direct clustering and clustering on the basis of factor analysis established the existence of six different classes of similar cations. The self-organizing map created with Viscovery SOMine visually represents that similarity.