How Stuff Really Works
A 4-Day workshop
(Bookings not being taken at this time)
The overall objective of the z-STRATEGIES—How Stuff Really Works (HSRW) class is to learn how to rapidly diagnose machine performance and reliability behaviour by using convergent strategies and tactics. This is done by hierarchically decomposing behaviour into those functions and properties that have leverage for improvement. We teach strategies and tactics that exploit the basic physical laws which govern behaviour. This is in sharp contrast to diagnosis that employs statistical analysis in the search for a ‘root cause’. Diagnosis based on physics is faster, more efficient and ultimately more effective.
A progressive search carried out using these tactics turns out to be easier to execute, and gives far more insight, than statistically designed experiments. However, we do recognise value in statistical analysis, and show how to combine probabilistic with deterministic evidence during diagnosis, and integrate both into the causal explanation.
The methodology requires simple functional models to be documented, and therefore we teach how to put together these models from basic building blocks. The models act as maps of the systems being studied. Class participants learn the essence of diagnostic strategy; how to converge from behaviour to causal explanation by exploiting the workspace defined by conjugate pairs of state variables. They will learn what to measure and/or manipulate in different energy domains, and how to empirically make use of generalized Thévanin & Norton equivalent systems.
We also teach how to extend the value of functional model-building beyond diagnosis into product validation. The class is about strategy as well as data collection and data analysis tactics. The key to improving performance and reliability lies in small data sets carefully chosen to answer specific questions. Actual behaviours of lumped functions identified in the models are characterised. This is easier, safer, faster and more revealing than employing accelerated life testing to failure.
To cement the key learning points, the z-STRATEGIES—HSRW class has been designed to be approximately 50% lecture, and 50% lab-based. In the lab sessions, attendees work in groups of 3 to apply the principles taught to diagnose the behaviour of some small1px desktop electro-mechanical and pneumatic machines. Data acquisition hardware and software is also used. The examples cover all types of behaviour observable in real devices.