High-purity (99.99 %) fine-grained alumina samples were prepared using the colloidal processing method to minimise the formation of agglomerates during powder consolidation. In this research, the soft agglomerates were broken down by ultrasonic agitation whereas sedimentation at pH ~ 2.0 successfully stabilised the suspension and separated it from the unwanted hard agglomerates. Consolidation was achieved through slip casting directly from the suspension to avoid the reflocculation of the powders. The removal of the agglomerates from the initial starting material resulted in the increase of packing density during compaction and led to a higher green density A dense homogeneous microstructure was thus achieved with a uniform, fine grain size generated at low sintering temperatures. Submicron-grained alumina of about 0.5 um in grain size was produced at the lowest sintering temperature of 1310°C.

The effects of time and total loading on the grinding process of food materials are presented in this work. Results show that the efficiency of process: defined as the weight ratio of product produced over the feed material, increases as the time of grinding increases (product in this case is defined as particles with size less than 250 1 um). At any given time, a volume fraction of 45% of the total volume of the grinding machine is found to be the optimum loading to produced the highest amount of product. me work index (Bond 1951) and attrition index (ASTM D440) of the food materials were calculated and found to be related in inverse manner.

This paper presents the use of artificial intelligence for the purpose of learning power quality. A software package has been developed in the form of an object oriented expert system by using the KAPPA-PC software. The power quality problem is arranged by means of information rules such as types of power disturbances, symptoms, causes and their solutions. The rule based reasoning process is by means of backward chaining. In the objective oriented approach, the power disturbances are represented as. classes and objects and the methods are used for communication between the objects. The developed expert system is also made user-friendly for the purpose of power quality education. It is shown that the proposed expert system can significantly reduce design and implementation efforts.

A process selection methodology for an industrial waste water treatment was developed and later incorporated into an expert system shell (XpertRule) which allowed a selection to be made based on biological treatment process alternatives. The methodology deals not only with tangible parameters but also intangible or non-numerical parameters which need further quantification on a hierarchical manner. This paper discusses the procedure for ranking the intangible parameter by incorporating statistical elements. Principal Component Analysis (PCA) for biological waste water treatment process selection. The quantification of intangible non numeric data has been rationalised in the study. This has overcome the problem of subjective judgement given to the process selection criteria as encountered by previous workers.

A process selection methodology was developed for an industrial wastewater treatment plant. The starting point of the procedure was the categorisation of tho wastewater based on industry type, general pollution indicators. or contaminant removal processes giving a standardised compositional code which consisted of seven basic wastewater characteristics. A preliminary design assessment was undertaken by comparing the effluent parameters with the desired effluent quality or consent conditions; where an effluent parameter was not known, a minimum characterisation set of analysis data was used. A preliminary process selection was carried out in terms of maximum volumetric loading rates and depths for each process which gave the required footprint area for the reactor. In addition to the reactor. the overall land area requirement was determined by incorporating the potential ancillary equipment such as sedimentation and sludge processing tanks for each process. The process selection was further refined by the establishment of performance graphs for each process. based on the volumetric loading rates and the percentage removal of COD or BOD. Based on these graphs. each process Can be quantified as to whether the COD or BOD consent is met in relation 10 the respective volumetric loading rate. If there was not a performance envelope available for the process. data was sought from a treatability study in the laboratory. The methodology has been incorporated into an expert system shell (XpertRule), which runs on a PC and provides a simple user interface. Certain provisions are made available in the program for new information to be added into the knowledge base. The automation of the methodology currently allows the user to make a selection based on biological treatment process alternatives.

Heavy metals such as chromium, nickel, copper. zinc and lead, can be effectively removed from metal finishing wastewaters by hydroxide precipitation. Prior to precipitation, ferrous sulphate is added to reduce chromium from hexavalent to trivalent state and also to form stable complex ferrocyanides with cyanide. This batch study was carried out to evaluate the minimum pH range for the minimum solubility of metal hydroxide precipitation. All the metals (Cu. Ni. Zn. Pb and Cr) were present simultaneously in synthetic wastewaters. In the absence of cyanide, the maximum hydroxide precipitation for Cu. Ni. Zn. Pb and Cr occurred at pH ranges of 6.5-12. 9-12. 8.5-12. 8.5-12 and 8-12 respectively. When cyanide was added into the wastewater. only pH range of minimum solubility for Cu was shifted to pH 8.5-11. For other metals, their pH ranges were not affected by the presence of cyanide. In addition, stable complex ferrocyanides precipitated optimally starting at pH 9.