| dc.description.abstract | It is a global challenge to produce enough healthy food to a growing world population. By moving industrial fish farming to exposed lo- cations, the farmers can possibly satisfy the dietary requirements of the future by expanding the production. The environment at exposed locations is rough, and conditions like harsh wind, large waves and strong currents are present. On the positive side, there are better wa- ter flow and distribution of waste. Exposed locations are also further away from natural salmons, which might reduce negative environmen- tal e acts.
The production plants for fish farming are designed to be flexible and adaptive to waves and sea currents. Monitoring of the health and condition of these structures will be more important regarded reducing cost of operations and maintenance. This data can be combined with historical data and expert knowledge to support the operators decision about acting upon a possible problematic situation.
Our goal in this MSc thesis is to study the method of implement- ing a decision support system for predictive maintenance of exposed aquaculture structures, based on a previously proposed architecture from the specialisation project. The architecture is based on the fields of Machine Learning, Structural Health Monitoring and Case-Based Reasoning.
This thesis describes the work done in order to achieve our goals, which include the work of researching the relevant domains, conduct- ing a data analysis, creating models with the resulting data sets, re- vising the previously proposed architecture and implementing a pro- totypical decision support system. At last we conclude and discuss the future work.
The work with this thesis shows that our prototype of a decision support system is able to support an operator with advice about what to do, if a situation similar to previously experienced situations occur. | |
