Fault detection in a cold forging process through feature extraction with a neural network

This paper investigates the application of neural networks to the recognition of lubrication defects typical to an industrial cold forging process employed by fastener manufacturers. The accurate recognition of lubrication errors, such as coating not being applied properly or damaged during material handling, is very important to the quality of the final product in fastener manufacture. Lubrication errors lead to increased forging loads and premature tool failure, as well as to increased defect sorting and the re-processing of the coated rod. The lubrication coating provides a barrier between the work material and the die during the drawing operation; moreover it needs be sufficiently robust to remain on the wire during the transfer to the cold forging operation. In the cold forging operation the wire undergoes multi-stage deformation without the application of any additional lubrication. Four types of lubrication errors, typical to production of fasteners, were introduced to a set of sample rods, which were subsequently drawn under laboratory conditions. The drawing force was measured, from which a limited set of features was extracted. The neural network based model learned from these features is able to recognize all types of lubrication errors to a high accuracy. The overall accuracy of the neural network model is around 98% with almost uniform distribution of errors between all four errors and the normal condition.

Improving an inverse model of sheet metal forming by neural network based regression

An inverse model for a sheet meta l forming process aims to determine the initial parameter levels required to form the final formed shape. This is a difficult problem that is usually approached by traditional methods such as finite element analysis. Formulating the problem as a classification problem makes it possible to use well established classification algorithms, such as decision trees. Classification is, however, generally based on a winner-takes-all approach when associating the output value with the corresponding class. On the other hand, when formulating the problem as a regression task, all the output values are combined to produce the corresponding class value. For a multi-class problem, this may result in very different associations compared with classification between the output of the model and the corresponding class. Such formulation makes it possible to use well known regression algorithms, such as neural networks. In this paper, we develop a neural network based inverse model of a sheet forming process, and compare its performance with that of a linear model. Both models are used in two modes, classification mode and a function estimation mode, to investigate the advantage of re-formulating the problem as a function estimation. This results in large improvements in the recognition rate of set-up parameters of a sheet metal forming process for both models, with a neural network model achieving much more accurate parameter recognition than a linear model.

Building agent-based hybrid intelligent systems

Many complex problems including financial investment planning, foreign exchange trading, knowledge discovery from large/multiple databases require hybrid intelligent systems that integrate many intelligent techniques including expert systems, fuzzy logic, neural networks, and genetic algorithms. However, hybrid intelligent systems are difficult to develop because they have a large number of parts or components that have many interactions. On the other hand, agents offer a new and often more appropriate route to the development of complex systems, especially in open and dynamic environments. In this paper, it is argued that agent technology is well snited for constructing hybrid intelligent systems (especially loosely coupled hybrid intelligent systems) through a successful case study. A great number of heterogeneous computing techniques/packages are easily integlated into the experimental system under a unifying agent framework, which implies that agent technology can greatly facilitate the construction of hybrid intelligent systems.

Knowledge extraction from a mixed transfer function artificial neural network

One of the big problems with Artificial Neural Networks (ANN) is that their results are not intuitively clear. For example, if we use the traditional neurons, with a sigmoid activation function, we can approximate any function, including linear functions, but the coefficients (weights) in this approximation will be rather meaningless. To resolve this problem, this paper presents a novel kind of ANN with different transfer functions mixed together. The aim of such a network is to i) obtain a better generalization than current networks ii) to obtain knowledge from the networks without a sophisticated knowledge extraction algorithm iii) to increase the understanding and acceptance of ANNs. Transfer Complexity Ratio is defined to make a sense of the weights associated with the network. The paper begins with a review of the knowledge extraction from ANNs and then presents a Mixed Transfer Function Artificial Neural Network (MTFANN). A MTFANN contains different transfer functions mixed together rather than mono-transfer functions. This mixed presence has helped to obtain high level knowledge and similar generalization comparatively to monotransfer function nets in a global optimization context.

Building agent-based hybrid intelligent systems : a case study

Many complex problems (e.g., financial investment planning, foreign exchange trading, data mining from large/multiple databases) require hybrid intelligent systems that integrate many intelligent techniques (e.g., fuzzy logic, neural networks, and genetic algorithms). However, hybrid intelligent systems are difficult to develop because they have a large number of parts or components that have many interactions. On the other hand, agents offer a new and often more appropriate route to the development of complex systems, especially in open and dynamic environments. Thus, this paper discusses the development of an agent-based hybrid intelligent system for financial investment planning, in which a great number of heterogeneous computing techniques/packages are easily integrated into a unifying agent framework. This shows that agent technology can indeed facilitate the development of hybrid intelligent systems.

Modeling the flow behavior, recrystallization, and crystallographic texture in hot-deformed Fe-30 Wt Pct Ni Austenite

The present work describes a hybrid modeling approach developed for predicting the flow behavior, recrystallization characteristics, and crystallographic texture evolution in a Fe-30 wt pct Ni austenitic model alloy subjected to hot plane strain compression. A series of compression tests were performed at temperatures between 850 °C and 1050 °C and strain rates between 0.1 and 10 s⁻¹. The evolution of grain structure, crystallographic texture, and dislocation substructure was characterized in detail for a deformation temperature of 950 °C and strain rates of 0.1 and 10 s⁻¹, using electron backscatter diffraction and transmission electron microscopy. The hybrid modeling method utilizes a combination of empirical, physically-based, and neuro-fuzzy models. The flow stress is described as a function of the applied variables of strain rate and temperature using an empirical model. The recrystallization behavior is predicted from the measured microstructural state variables of internal dislocation density, subgrain size, and misorientation between subgrains using a physically-based model. The texture evolution is modeled using artificial neural networks.

Eccentricity and hardness control in cold rolling mills with a dynamically constructed neural controller

The main objective of a steel strip rolling process is to produce high quality steel at a desired thickness.  Thickness reduction is the result of the speed difference between the incoming and the outgoing steel strip and the application of the large normal forces via the backup and the work rolls.  Gauge control of a cold rolled steel strip is achieved using the gaugemeter principle that works adequately for the input gauge changes and the strip hardness changes.  However, the compensation of some factors is problematic, for example, eccentricity of the backup rolls.  This cyclic eccentricity effect causes a gauge deviation, but more importantly, a signal is passed to the gap position control so to increase the eccentricity deviation.  Consequently, the required high product tolerances are severely limited by the presence of the roll eccentricity effects.
In this paper a direct model reference adaptive control (MRAC) scheme with dynamically constructed neural controller was used.  The aim here is to find the simplest controller structure capable of achieving an optimal performance.  The stability of the adaptive neural control scheme (i.e. the requirement of persistency of excitation and bounded learning rates) is addressed by using as the inputs to the reference model the plant's state variables.  In such a case, excitation is due to actual plant signals (states) affected by plant disturbances and noise.  In addition, a reference model in the form of a filter with a desired transfer function using Modulus Optimum design was used to ensure variance in the desired dynamic characteristics of the system.  The gradually decreasing learning rate employed by the neural controller in this paper is aimed at eliminating controller instability resulting from over-aggressive control.  The moving target problem (i.e. the difficulty of global neural networks to perfrom several separate computational tasks in closed -loop control) is addressed by the localized architecture of the controller.  The above control scheme and learning algorithm offers a method for automatic discovery of an efficient controller.
The resulting neural controller produces an excellent disturbance rejection in both cases of eccentricity and hardness disturbances, reducing the gauge deviation due to eccentricity disturbance from 33.36% to 4.57% on average, and the gauge deviation due to hardness disturbance from 12.59% to 2.08%.

Agent-based hybrid framework for decision making on complex problems

Electronic commerce and the Internet have created demand for automated systems that can make complex decisions utilizing information from multiple sources. Because the information is uncertain, dynamic, distributed, and heterogeneous in nature, these systems require a great diversity of intelligent techniques including expert systems, fuzzy logic, neural networks, and genetic algorithms. However, in complex decision making, many different components or sub-tasks are involved, each of which requires different types of processing. Thus multiple such techniques are required resulting in systems called hybrid intelligent systems. That is, hybrid solutions are crucial for complex problem solving and decision making. There is a growing demand for these systems in many areas including financial investment planning, engineering design, medical diagnosis, and cognitive simulation. However, the design and development of these systems is difficult because they have a large number of parts or components that have many interactions. From a multi-agent perspective, agents in multi-agent systems (MAS) are autonomous and can engage in flexible, high-level interactions. MASs are good at complex, dynamic interactions. Thus a multi-agent perspective is suitable for modeling, design, and construction of hybrid intelligent systems. The aim of this thesis is to develop an agent-based framework for constructing hybrid intelligent systems which are mainly used for complex problem solving and decision making. Existing software development techniques (typically, object-oriented) are inadequate for modeling agent-based hybrid intelligent systems. There is a fundamental mismatch between the concepts used by object-oriented developers and the agent-oriented view. Although there are some agent-oriented methodologies such as the Gaia methodology, there is still no specifically tailored methodology available for analyzing and designing agent-based hybrid intelligent systems. To this end, a methodology is proposed, which is specifically tailored to the analysis and design of agent-based hybrid intelligent systems. The methodology consists of six models - role model, interaction model, agent model, skill model, knowledge model, and organizational model. This methodology differs from other agent-oriented methodologies in its skill and knowledge models. As good decisions and problem solutions are mainly based on adequate information, rich knowledge, and appropriate skills to use knowledge and information, these two models are of paramount importance in modeling complex problem solving and decision making. Follow the methodology, an agent-based framework for hybrid intelligent system construction used in complex problem solving and decision making was developed. The framework has several crucial characteristics that differentiate this research from others. Four important issues relating to the framework are also investigated. These cover the building of an ontology for financial investment, matchmaking in middle agents, reasoning in problem solving and decision making, and decision aggregation in MASs. The thesis demonstrates how to build a domain-specific ontology and how to access it in a MAS by building a financial ontology. It is argued that the practical performance of service provider agents has a significant impact on the matchmaking outcomes of middle agents. It is proposed to consider service provider agents' track records in matchmaking. A way to provide initial values for the track records of service provider agents is also suggested. The concept of ‘reasoning with multimedia information’ is introduced, and reasoning with still image information using symbolic projection theory is proposed. How to choose suitable aggregation operations is demonstrated through financial investment application and three approaches are proposed - the stationary agent approach, the token-passing approach, and the mobile agent approach to implementing decision aggregation in MASs. Based on the framework, a prototype was built and applied to financial investment planning. This prototype consists of one serving agent, one interface agent, one decision aggregation agent, one planning agent, four decision making agents, and five service provider agents. Experiments were conducted on the prototype. The experimental results show the framework is flexible, robust, and fully workable. All agents derived from the methodology exhibit their behaviors correctly as specified.

Predicting amount of saleable products using neural network metamodels of casthouses

This study aims at developing abstract metamodels for approximating highly nonlinear relationships within a metal casting plant. Metal casting product quality nonlinearly depends on many controllable and uncontrollable factors. For improving the productivity of the system, it is vital for operation planners to predict in advance the amount of high quality products. Neural networks metamodels are developed and applied in this study for predicting the amount of saleable products. Training of metamodels is done using the Levenberg-Marquardt and Bayesian learning methods. Statistical measures are calculated for the developed metamodels over a grid of neural network structures. Demonstrated results indicate that Bayesian-based neural network metamodels outperform the Levenberg-Marquardt-based metamodels in terms of both prediction accuracy and robustness to the metamodel complexity. In contrast, the latter metamodels are computationally less expensive and generate the results more quickly.

Construction and optimisation of prediction intervals for neural network models

This thesis makes contributions to basic and fundamental research in the field of prediction interval construction using neural network models. It proposes novel methods for objective assessment, rapid construction, and optimisation of neural network-based prediction intervals for uncertainty quantification.

Hybrid neural-evolutionary model for electricity price forecasting

Evolving artificial neural networks has attracted much attention among researchers recently, especially in the fields where plenty of data exist but explanatory theories and models are lacking or based upon too many simplifying assumptions. Financial time series forecasting is one of them. A hybrid model is used to forecast the hourly electricity price from the California Power Exchange. A collaborative approach is adopted to combine ANN and evolutionary algorithm. The main contributions of this thesis include: Investigated the effect of changing values of several important parameters on the performance of the model, and selected the best combination of these parameters; good forecasting results have been obtained with the implemented hybrid model when the best combination of parameters is used. The lowest MAPE through a single run is 5. 28134%. And the lowest averaged MAPE over 10 runs is 6.088%, over 30 runs is 6.786%; through the investigation of the parameter period, it is found that by including future values of the homogenous moments of the instant being forecasted into the input vector, forecasting accuracy is greatly enhanced. A comparison of results with other works reported in the literature shows that the proposed model gives superior performance on the same data set.