A Hybrid Discrete Bat Algorithm with Krill Herd-based advanced planning and scheduling tool for the capital goods industry

Capital goods companies produce high value products such as power plant or ships, which have deep and complex product structures, with components having long process routings. Contracts usually include substantial penalties for late delivery. The high value of items can lead to substantial holding costs. Efficient schedules minimise earliness and tardiness costs and need to satisfy assembly and operation precedence constraints as well as finite capacity. This paper presents the first advanced planning and scheduling (APS) tool for the capital goods industry that uses a Discrete Bat Algorithm (DBA), modified DBA (MDBA) and hybrid DBA with Krill Herd algorithm (HDBK) to optimise schedules. The tool was validated using four data-sets obtained from a collaborating capital goods company. A sequential experimental strategy was adopted. The first experiment identified appropriate parameter settings for the DBA. The second experiment evaluated and compared the performance of the proposed HDBK algorithm with an Artificial Bee Colony, Krill Herd (KH), Modified KH, DBA and MDBA metaheuristics. The experimental results revealed that the HDBK performed best in terms of the minimum penalty cost for all problem sizes and achieved up to a 47.837% reduction in mean total penalty costs of extra-large problem size.     

Using a membrane reactor for the oxidative coupling of methane: simulation and optimization

An oxidative coupling of methane (OCM) is a promising process to convert methane into ethylene and ethane; however, it suffers from the relatively low selectivity and yield of ethylene at high methane conversion. In this study, a membrane reactor is applied to the OCM process in order to prevent the deep oxidation of a desirable ethylene product. The mathematical model of OCM process based on mass and energy balances coupled with detailed OCM kinetic model is employed to examine the performance of OCM membrane reactor in terms of CH₄ conversion, C₂ selectivity, and C₂ yield. The influences of key operating parameters (i.e., temperature, methane-to-oxygen feed ratio, and methane flow rate) on the OCM reactor performance are further analyzed. The simulation results indicate that the OCM membrane reactor operated at higher operating temperature and lower methane-to-oxygen feed ratio can improve C₂ production. An optimization of the OCM membrane reactor using a surface response methodology is proposed in this work to determine its optimal operating conditions. The central composite design is used to study the interaction of process variables (i.e., temperature, methane-to-oxygen feed ratio, and methane flow rate) and to find the optimum process operation to maximize the C₂ products yield.     

A mixed-integer linear programming model for integrated production and preventive maintenance scheduling in the capital goods industry

The scheduling literature is extensive, but much of this work is theoretical and does not capture the complexity of real world systems. Capital goods companies produce products with deep and complex product structures, each of which requires the coordination of jobbing, batch, flow and assembly processes. Many components require numerous operations on multiple machines. Integrated scheduling problems simultaneously consider two or more simultaneous decisions. Previous production scheduling research in the capital goods industry has neglected maintenance scheduling and used metaheuristics with stochastic search that cannot guarantee an optimal solution. This paper presents a novel mixed integer linear programming model for simultaneously solving the integrated production and preventive maintenance scheduling problem in the capital goods industry, which was tested using data from a collaborating company. The objective was to minimise total costs including: tardiness and earliness penalty costs; component and assembly holding costs; preventive maintenance costs; and set-up, production, transfer and production idle time costs. Thus, the objective function and problem formulation were more extensive than previous research. The tool was successfully tested using data obtained from a collaborating company. It was found that the company’s total cost could be reduced by up to 63.5%.     

Analysis of nanoparticle agglomeration in aqueous suspensions via constant-number Monte Carlo simulation

A constant-number direct simulation Monte Carlo (DSMC) model was developed for the analysis of nanoparticle (NP) agglomeration in aqueous suspensions. The modeling approach, based on the "particles in a box" simulation method, considered both particle agglomeration and gravitational settling. Particle-particle agglomeration probability was determined based on the classical Derjaguin-Landau-Verwey-Overbeek (DLVO) theory and considerations of the collision frequency as impacted by Brownian motion. Model predictions were in reasonable agreement with respect to the particle size distribution and average agglomerate size when compared with dynamic light scattering (DLS) measurements for aqueous TiO(2), CeO(2), and C(60) nanoparticle suspensions over a wide range of pH (3-10) and ionic strength (0.01-156 mM). Simulations also demonstrated, in quantitative agreement with DLS measurements, that nanoparticle agglomerate size increased both with ionic strength and as the solution pH approached the isoelectric point (IEP). The present work suggests that the DSMC modeling approach, along with future use of an extended DLVO theory, has the potential for becoming a practical environmental analysis tool for predicting the agglomeration behavior of aqueous nanoparticle suspensions.     

Practical Studies of the Electrolysis and Volatilization of the Bromide from Drinking Water to Minimize Bromate Production by Ozonation

In four recently published articles, a process for the oxidation of bromide to bromine and the volatilization of bromine from drinking water sources was presented. This process was shown to be able to remove up to 35% percent of the bromide found naturally in the California State Water Project. Although bromide itself is quite harmless, it has been shown to react with commonly used disinfectants to produce compounds or disinfection by-products (DBPs) of suspected carcinogens. Bromide reacts with ozone to form bromate. This article presents two studies of pilot scale, flow-through electrolytic reactors that oxidize bromide to bromine and volatilize bromine at <pH 3.5, which occurs at the anode as a result of the oxidation of water. One reactor had 14 anodes that were 91 cm deep and the other had 13 anodes 1.2 cm deep. The bromide removal rates were studied at several different water flows and power settings for different bromide concentrations for both reactors. The results show removal of bromide is impacted by water flows and power settings for different bromide concentrations. Effluent from the deep reactor did show some reduction in bromate concentration as compared to control samples but the results were inconsistent. This appeared to be caused by significant differences in the ozone demand produced by different experimental conditions, difficulty determining the concentration of chlorine, and the use of hydrogen peroxide as a dechlorinating agent. Using the shallow reactor, these difficulties were overcome by developing a more consistent determining chlorine concentration, using much larger ozone doses to overwhelm the ozone demand, and by using ascorbic acid instead of hydrogen peroxide. With these changes, it could be shown that the electrolytic reactor not only lowered the concentration of bromide in the water but when ozonated, the amount of bromate formed was reduced in direct proportion to the amount of bromide removed for an equal dose of ozone.