Browsing by Author "Kallava, Tomi"

This thesis is done for Wärtsilä, which is a big global actor in marine and energy markets. This thesis aims to test the feasibility of machine learning models in estimating total power- and fuel consumption of vessels' main engines and thus help in recognizing the effect of different factors on the energy consumption of vessels. This, on the other hand, helps to optimize routes and machinery concepts among other things. Another goal is to compress the engine sensor data utilizing wavelet transformation. After the introduction to the topic in the second chapter, we introduce the data we are using in this study. These include vessel location data, engine sensor data and technical specifications of the vessels. In the third chapter, we go through the mathematical formulations of the used methods. Finally, we will perform the calculations with real data and analyze the results. We'll test the performance of compression methods applied to the time series data coming from sensors. After that, we'll test different regression methods for consumption estimations and see what gives the most accurate results.