Neural network modelling and multiobjective optimization of creep feed grinding of superalloys

The grinding process is a very complex system for which analytical and empirical models have been developed to pursue a control strategy. This paper utilizes a new approach to model the creep feed grinding of superalloys, Ti-6Al-4V and Inconel 718, by using a neural network. A back-propagation learning algorithm is adopted to capture the system behaviour. The neural network learns to associate the inputs (feed rate, depth of cut and wheel bond type) with the outputs (surface finish, force and power) and predicts the systems outputs within the working conditions. Mathematical formulation of a multiobjective optimization problem is then carried out by utilizing the network models. The optimization study results are presented in the form of decision tables and value path diagrams to assist the decision-making process     

Computational intelligence modelling of pharmaceutical tabletting processes using bio-inspired optimization algorithms

In pharmaceutical development, it is very useful to exploit the knowledge of the causal relationship between product quality and critical material attributes (CMA) in developing new formulations and products, and optimizing manufacturing processes. With the big data captured in the pharmaceutical industry, computational intelligence (CI) models could potentially be used to identify critical quality attributes (CQA), CMA and critical process parameters (CPP). The objective of this study was to develop computational intelligence models for pharmaceutical tabletting processes, for which bio-inspired feature selection algorithms were developed and implemented for optimisation while artificial neural network (ANN) was employed to predict the tablet characteristics such as porosity and tensile strength. Various pharmaceutical excipients (MCC PH 101, MCC PH 102, MCC DG, Mannitol Pearlitol 200SD, Lactose, and binary mixtures) were considered. Granules were also produced with dry granulation using roll compaction. The feed powders and granules were then compressed at various compression pressures to produce tablets with different porosities, and the corresponding tensile strengths were measured. For the CI modelling, the efficiency of seven bio-inspired optimization algorithms were explored: grey wolf optimization (GWO), bat optimization (BAT), cuckoo search (CS), flower pollination algorithm (FPA), genetic algorithm (GA), particle swarm optimization (PSO), and social spider optimization (SSO). Two-thirds of the experimental dataset was randomly chosen as the training set, and the remaining was used to validate the model prediction. The model efficiency was evaluated in terms of the average reduction (representing the fraction of selected input variables) and the mean square error (MSE). It was found that the CI models can well predict the tablet characteristics (i.e. porosity and tensile strength). It was also shown that the GWO algorithm was the most accurate in predicting porosity. While the most accurate prediction for the tensile strength was achieved using the SSO algorithm. In terms of the average reduction, the GA algorithm resulted in the highest reduction of inputs (i.e. 60%) for predicting both the porosity and the tensile strength.     

Computer modelling of hot-working processes

Abstract

In this paper, the development of process modelling within the Department of Metallurgy at The University of Sheffield is traced. This was initially concerned with predicting temperature distributions generated during hot-extrusion and hot-rolling operations, but the principles of finite-difference modelling have also been applied to upset forging and compression testing. The modelling of microstructural evolution during multipass working operations from the basis of laboratory test data was developed separately. These two aspects have been combined, and the effects of microstructure, temperature, and other engineering variables on flow stress have been incorporated to produce a model of hot rolling which predicts microstructural changes at each stage of the process and takes into account their effect on rolling-load and power requirements.

MST/168     

Computer modelling of dynamic processes

We demonstrate the capabilities of developed software products for modelling oscillatory processes. A digital method assures efficiency for modelling with given probabilistic characteristics. Analog simulation makes it possible to solve systems of differential equations with different types of nonlinearities and inputs. Translated from Izmeritel'naya Tekhnika, No. 3, pp. 16–19, March, 1996.     

低温冷风磨削加工参数优化试验

绿色加工是国际国内金属加工行业的趋势,采用自行设计的低温冷风装置进行绿色磨削试验研究,测试低温冷风温度(范围:- 40至- 80℃)、压力、磨削进给量、砂轮粒度等因素对试件表面加工质量的影响,发现存在磨削最佳参数组合,且低温冷风下的磨削质量可以达到传统磨削的精磨标准.     

Closed system of equations for analysis of grinding processes

The grinding of a magnesium-containing mineral, brucite, in a Fritsch Pulverisette-5 planetary ball mill has been studied in a semiempirical approach. An expression has been derived for evaluating the particle size distribution density of the ground material. It is shown that the distribution density can be used as a process control parameter.     

Energy and cement quality optimization of a cement grinding circuit

This study aimed at optimizing both the energy efficiency and the quality of the end product by modifying the existing flowsheet of the cement grinding circuit. As a general application, mill filter stream is sent to the air classifier owing to its coarser size distribution than the desired product. However, the study proved that some further evaluations i.e., quality tests and chemical assays, could make it possible to treat this stream as a final product. Consequently, directing this stream to the final product silo could be considered. Within the study, sampling survey was undertaken initially that was followed by the modelling and simulation works. The calculations implied that the production rate increased by 4.45% that corresponded to energy saving of 4.26%. As the plant decided to change the flow sheet, another sampling campaign was arranged to validate the outputs of the simulation studies. In that case, the real data showed that the increase in production rate was 3.68% and 28-Days strength of the cement improved by 2.9%. As a result, the simulation outputs were found to be in agreement with the real data hence the efficiency of the cement production, both quality and energy, for a given circuit was improved.     

Syntactic user modelling with stochastic processes

An approach to user modelling with discrete stochastic processes is presented, which aims at the dynamic individualization of user interfaces on the syntactic layer. The state of the art of syntactic user modelling is surveyed. The mathematical background of simple Markov Chains and the 'classic' Hidden Markov Model is presented. Furthermore, dynamic Bayesian Networks are introduced, which generalize these Markovian Models. A case study of the simulation experiments uses a multimodal user interface for supervisory control of advanced manufacturing cells. A corresponding simulation model is created and exploited to generate interaction cases, which are the empirical basis for the evaluation. Six topologies of dynamic Bayesian Networks are evaluated for 100 interaction cases and 50 replications each: (1) Markov Chain of order 1, (2) Hidden Markov Model, (3) autoregressive Hidden Markov Model, (4) factorial Hidden Markov Model, (5) simple hierarchical Hidden Markov Model, and (6) tree structured Hidden Markov Model. The dependent variable is the prediction accuracy for a single prediction lead. In a first step, a one-way analysis of variance in conjunction with Tukey's post-hoc test demonstrated a significant superiority of the simple hierarchical Hidden Markov Model. In a second step, an additional two-way analysis of variance also indicated a significantly better prediction accuracy of the simple hierarchical Hidden Markov Model compared to the Hidden Markov Model, but the number of interaction cases also had a significant effect. Hence, the modeller has to take both factors--model topology and number of interaction cases--into account when designing syntactic user models with stochastic processes.     

CFD modelling of condensers for freeze-drying processes

The aim of the present research is the development of a computational tool for investigating condensation processes and equipment with particular attention to freeze-dryers. These condensers in fact are usually operated at very low pressures, making it difficult to experimentally acquire quantitative knowledge of all the variables involved. Mathematical modelling and CFD (Computational Fluid Dynamics) simulations are used here to achieve a better comprehension of the flow dynamics and of the process of ice condensation and deposition in the condenser, in order to evaluate condenser efficiency and gain deeper insights of the process to be used for the improvement of its design. Both a complete laboratory-scale freeze-drying apparatus and an industrial-scale condenser have been investigated in this work, modelling the process of water vapour deposition. Different operating conditions have been considered and the influence exerted by the inert gas as well as other parameters has been investigated.     

Multicriterion optimization of plasma processes

A new method of automatic multicriterion synthesis of plasma processes and apparatuses is proposed. The method is applied to multicriterion optimization of reduction plasma production with the recovery of blast-furnace gas in a direct steel-making process.     

Energy and material flow modelling of additive manufacturing processes

Additive manufacturing (AM) processes allow fabrication of three-dimensional complex parts. Due to the exact amount of material used during the manufacturing step, these new manufacturing processes offer great opportunities for sustainable manufacturing. However, existing studies on these processes focus mainly on energy consumption and information about resources consumptions and waste flows are still lacking. This study aims to quantify with accuracy inventory data of AM processes during the manufacturing step of the life cycle of products. In order to accurately assess the environmental impact of a product, a generic method for acquisitions and characterisation of inventory data for parts made by AM processes is proposed. This methodology focuses not only on the electrical energy consumption but also on material consumption. This paper also describes the development of a parametric process model, which provides to an operator, an accurate estimation of the environmental performances of the fused deposition modelling process.     

On modelling thermoelastic processes in layered composites

Summary The main aim of this contribution is to propose a 2D-mathematical model of thermoelastic layered composites. This model is based on a 3D-linearized thermoelastic continuum theory with internal kinematical and thermal constraints. A fundamental system of field equations for 3D-layered composites is reduced to a system of field equations for 2D-thermoelastic material continua, describing thermoelastic processes in composite plates and plate-like bodies.This paper has been presented at the 28^th Polish Solid Mechanics Conference, September 4–8, 1990, Kozubnik, Poland     

Evolutionary optimization of material properties of a tropical seed

Here, we show how the mechanical properties of a thick-shelled tropical seed are adapted to permit them to germinate while preventing their predation. The seed has evolved a complex heterogeneous microstructure resulting in hardness, stiffness and fracture toughness values that place the structure at the intersection of these competing selective constraints. Analyses of different damage mechanisms inflicted by beetles, squirrels and orangutans illustrate that cellular shapes and orientations ensure damage resistance to predation forces imposed across a broad range of length scales. This resistance is shown to be around the upper limit that allows cracking the shell via internal turgor pressure (i.e. germination). Thus, the seed appears to strike an exquisitely delicate adaptive balance between multiple selection pressures.     

Evolutionary reliable regional Kriging surrogate for expensive optimization

This work proposes a sequential optimization algorithm, EORKS, combining a Kriging surrogate from an adaptive sampling and an iterative constrained search in the dynamic reliable regions to reduce the sampling size in expensive optimization. A surrogate established from small samples is liable to limited generality, which leads to a false prediction of optimum. EORKS applies Kriging variance to establish the reliable region neighbouring the learning samples to constrain the evolutionary searches of the surrogate. The verified quasi-optimum is used as an additional sample to dynamically update the regional model according to the prediction accuracy. A hybrid infilling strategy switches between the iterative quasi-optima and the maximum expected improvement from Kriging to prevent early convergence of local optimum. EORKS provides superior optima in several benchmark functions and an engineering design problem, using much smaller samples compared with the literature results, which demonstrates the sampling efficiency and searching robustness.     

Predictive modelling for optimization of textile composite forming

Wrinkling often occurs during textile composite forming and is a major problem for manufacturers. The prediction of this defect is, therefore, of major importance for the design and optimization of textile composite structures. Numerical simulations of forming for textile composites over a hemisphere have been conducted using a rate/temperature-dependent hybrid FE model. The hybrid FE model incorporates a fully predictive multi-scale energy model which determines the shear resistance of the textile composite sheet. The effects of varying the normal force distribution across the edges of the blank and blank size, together with the effect of changes in forming temperature on the final fibre pattern and wrinkling behaviour, are investigated. Predictions are evaluated against press-formed components. The results from the simulation and the experiments have good correlation and show that wrinkling can be minimized by optimizing the force distribution around the edge of the manufacturing tool and by careful choice of forming temperature.     

Life cycle cost modelling using marked point processes

This article presents a life cycle cost (LCC) model of a repairable system. The model is based on a marked point process and allows for non-constant failure intensity as well as stochastic nature of costs associated with system's failures. The model is applied to failure data from computer numerically controlled (CNC) machines.