Please use this identifier to cite or link to this item:
https://hdl.handle.net/10321/2358
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Chetty, Maggie | en_US |
dc.contributor.author | Carsky, M. | en_US |
dc.contributor.author | Kormuth, E. | en_US |
dc.date.accessioned | 2017-03-10T06:57:04Z | |
dc.date.available | 2017-03-10T06:57:04Z | |
dc.date.issued | 2015 | - |
dc.identifier.citation | Chetty, M.; Carsky, M. and Kormuth, E. 2015. Application of neural network in chemical and non-chemical engineering problems. South African Journal of Chemical Engineering. 20(2): 1-17. | en_US |
dc.identifier.issn | 1026-9185 | - |
dc.identifier.uri | http://hdl.handle.net/10321/2358 | - |
dc.description.abstract | Use of a neural network approach to the solution of complex tasks is demonstrated on first, the identification of a food chemical specification (“Chemical Engineering case”) and second, the prediction of survival and prognosis for leukaemia patients (“Non-Chemical Engineering case”). In the first case product colour was identified as an important quality feature, which significantly affects marketing of food grade antioxidants used for preservation of edible oils. The product quality is specified in terms of the “Lovibond colour index”. Production of phenolic antioxidant in a well-established operation exhibited for an extended period of time a variation of the product colour. Due to the complexity of the technological process it is impossible to pinpoint a simple reason for variations. The product is synthesised in a batch reactor, the crude product is then recovered from post-reaction batch via crystallisation and refined in a complex purification process to get the final product, which meets specification. It is unclear, whether the undesirable colour is formed during the synthesis itself or results from underperformance of the purification process. Data from the manufacturing process were collected over 140 batches, where the impurity profile of crude product was recorded along with some other purification parameters. Impurity profile included an occurrence of 32 different compounds, some of them occurring on a random basis. The neural network model predicts the final product quality on the basis of a crude product impurity profile. Thus the process management decision to premeditate the product can be made well in advance during the process. Should the prediction show that the final product would not meet the required specification, corrective measures can be implemented well in advance to rescue the final product. In the second case the neural network model attempts to predict survival rate of leukaemia patients over two and three year periods based on 38 medical factors of patients and treatment procedures chosen, Which make it possible to apply the right treatment at an early stage. | en_US |
dc.format.extent | 17 p | en_US |
dc.language.iso | en | en_US |
dc.publisher | South African Institution of Chemical Engineers | en_US |
dc.relation.ispartof | South African journal of chemical engineering | en_US |
dc.title | Application of neural network in chemical and non-chemical engineering problems | en_US |
dc.type | Article | en_US |
dc.dut-rims.pubnum | DUT-005240 | en_US |
item.grantfulltext | open | - |
item.cerifentitytype | Publications | - |
item.fulltext | With Fulltext | - |
item.languageiso639-1 | en | - |
item.openairecristype | http://purl.org/coar/resource_type/c_18cf | - |
item.openairetype | Article | - |
Appears in Collections: | Research Publications (Engineering and Built Environment) |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Chetty_SAJCHE_Vol20No2_2015.pdf | 473.61 kB | Adobe PDF | View/Open |
Page view(s)
470
checked on Dec 22, 2024
Download(s)
130
checked on Dec 22, 2024
Google ScholarTM
Check
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.