Husqvarna AB has as of today an extensive research and development department.This department serves to control the active product as well as the upcoming ones.The way that is done is through two different sets of tests. The first one being a longterm endurance test with aimed to unveil the durability of a product. Second and finalsort of test is a more one dimensional one. The aim is to determine different specificunits of interest like for example Newton (N). Today the R&D department has a great knowledge within normal distributed data andsomewhat less when it comes to the opposite, so called none normal distributed data.When endurance is of interest the likelihood of that to be of the latter sort is morecommon than not. For now no complete method has been appointed to make iteasier to process a situation of this kind. Studying ever unique case individually, bylooking at the data, has been the way to go. This causes an inconsistency in theanalysis and makes it purely based on which individual that has done it. Lastly it mayalso, unintentionally, ignore the large picture of how a product has progressed. To solve these problems this thesis work was put together to propose and conduct amethod. To form this method was an ongoing process throughout the whole thesisperiod. Ideas and thoughts were put forward to be reviewed and discussed. After aseries of tweaks to steer it towards the overall goal the method was finalized. Themethod that was put forward was firmly tested. Also a wide laboration in what themethod actually meant was done. The result was a method to be applied on none normal distributed data. This methodhas three parts. The first being the report where everything is embraced. The secondpart is a short manual for an operator to use. Last part is an example where themethod is put to use.
|Publisher||Uppsala universitet, Matematiska institutionen|
|Source Sets||DiVA Archive at Upsalla University|
|Type||Student thesis, info:eu-repo/semantics/bachelorThesis, text|
Page generated in 0.0064 seconds