Exjobbsförslag från företag

Detta är ett uppsatsförslag hämtat från Nationella Exjobb-poolen. Klicka här för att komma tillbaka till samtliga exjobbsförslag.

Förslaget inkom 2010-10-07

Estimation of Key Product Qualities

Nynas AB (Nynas) upgrades crude oil to special products. This separates Nynas from most other oil companies, which typically produce fuels and heating oils. Nynas refineries produce naphthenic specialty (NSP) oils as well as bitumen. NSP oils are used as insulating and cooling elements in transformers, as softeners in plastic products, in lubricating oils and many other industrial applications. Bitumen is primarily used as a binding agent in asphalt road.
Nynas has a global marketing and sales organisation in 30 countries, as well as 9 refineries (fully owned and in partnership) which are situated in Nynäshamn and in Gothenburg in Sweden, as well as in Great Britain, Europe, USA and Venezuela.
More information can be found at http://www.nynas.com.

The Nynas product qualities are very important for our customers. There are strict quality specifications for all of Nynas AB's products. To continuously improve the product quality is always one of the main Nynas goals. Some of the product qualities, however, are difficult to measure continuously during product processing, but product samples are taken at least every 8 hours and analysed in the laboratory. Truthful and accurate quality predictions provide complement to the less frequent lab results and may indicate quality deviation in advance, which also sets good basis for advanced control.

To model the product quality properties using process recorded parameters.

Work Description
1. Select cause process variables and build models for product qualities.
2. Develop a general and clear modelling approach which should be easy to apply.
By using this general approach, process data should easily be conditioned to remove data outliers; process data at different operating points should easily be separated; models should easily be updated, e.g. for different time periods.
This approach (preferably with modelling templates) should be applicable for any operating points / modes / processes. Further more, the ability to apply the same or different quality models for different operation modes is to be investigated.
3. Set up criteria for model acceptance and for comparing different models. Develop model accuracy online measurements and model update criteria.
4. If interested and time allows, different modelling aproaches (e.g. PLS/PCA, GP…) can be applied to compare pros and cons.


Informationen om uppsatsförslag är hämtad från Nationella Exjobb-poolen.