Javascript must be enabled for the correct page display

Predictive maintenance: the feasibility of a non‐straight edge knife sharpness deterioration model

Tjabbes, Pieter (2018) Predictive maintenance: the feasibility of a non‐straight edge knife sharpness deterioration model. Research Project, Industrial Engineering and Management.


Download (3MB) | Preview
[img] Text
Restricted to Registered users only

Download (106kB)


In September 2013 the European Union(EU) announced that the sugar quota system would end in September 2017. As a result the sugar price in the EU declined nearly 50% in five years, forcing sugar production facilities to drastically increase their efficiency. SuikerUnie Vierverlaten scaled up its throughput and tried to optimize the sugar to water diffusion through increased cossette (sliced beets) quality. Optimized diffusion requires less water, consequently lowering the energy cost related to evaporation. Determining the optimal knife operation time increases cossette quality. A deep learning algorithm was implemented to decide which slicer should have its knives changed. Due to the nature of an organic product and the available data provided to the model, an accuracy of only 60% was realized. The aim of this research is to prove the feasibility of a knife sharpness deterioration prediction model through the analysis of non-straight edge factory knives that experienced deterioration under specific input settings while side-liningthe effect of external factors. While knowledge of knife sharpness related to cutting soft solids is wildly available, no benchmark is mentioned for knives with a non-straight edge blade geometry. Three knife sharpness measurement methods were customized, applied and verified. It was found that two out of three methods could be successfully implemented for non-straight edge blades.Butwere not usable as standalone variables due to the accuracy limits required to quantify the input variables at Vierverlaten. However, constants acquired from the knife analysis observed at steady state cutting proved that quantifying the input settings is possible and multiple findings lead to further research opportunities to accurately quantify variables.

Item Type: Thesis (Research Project)
Supervisor name: Jayawardhana, B.
Degree programme: Industrial Engineering and Management
Thesis type: Research Project
Language: English
Date Deposited: 18 Jul 2018
Last Modified: 18 Jul 2018 10:10

Actions (login required)

View Item View Item