I've chosen to start this blog, just to vent my thoughts, and keep some sort of diary! Furthermore, i'll use it when I go to USA, so family and friends can follow what i'm doing. The contents of the blog, will be a mixture of tech and phd related news, mixed with societal issues at hand - it'll be mixed up, like Jeremy Clarkson does it, in his car reviews.
But, here goes!
The goal of my PhD, is to develop a method that can detect, isolate and determine errors on a refrigeration container, you know, one of those that ensure that the fruit you buy at the store is ripe. To the companies who own the containers, it is really important to know the integrity of the system, as they can prepare maintenance accordingly. A system that malfunctions, destroys the cargo, and trust me on this, a container packed with luke-warm, rotten meat, does not smell good.
There are several methods that can be applied to solve such a problem, where most of them, apply some sort of model-based approach. For you, non-mathematicians, a model is a way to express how an input, influences an output. One of these model-based approaches, is to apply, what is known as a Kalman Filter to the measurements. This have several benefits, firstly, it filters the sensor measurements, and provides the end-user with a better idea of the different temperatures in the system.
When its working properly, you also gain the benefit if being able to remove sensors from your system, and keep the system running! The latter is a huge benefit, given the problem described above.
The tuning of the filter, however, is really a pain in the ass! As a small adjustment one place, can lead to huge changes in other places, cause and effect. Today, i've spent several hours trying to tune it, and as soon as I saw some progress, the next iteration, went two steps back. I weren't able to determine what caused which changes, but i'll get back to that tomorrow, and hopefully be able to go to weekend, with a functioning prototype.
Thats it for now!