Basic Principles - Chapter 6
Shared by Miryam Strautkalns, updated on Mar 20, 2013
- Author(s) :
- George Vachtsevanos, Kai Goebel
This chapter described at a very high level some of the considerations that need to be made when designing algorithms for a vehicle health management application. The choices made here affect the quality of the diagnosis and prognosis (covered in Chapter 7). Therefore, the algorithmic design choices are made in conjunction with the design choices for diagnostics and prognostics to optimally support these tasks. Furthermore, additional considerations imposed by computational constraints, resource availability, algorithm maintenance, need for algorithm re-tuning, etc. will impact the solutions.
It should also be noted that technological advances, both in hardware and software, impose the need for new solutions. For example, as new materials and new sensors are being developed, the algorithmic solutions will need to follow suit.
In general, there seems to be a trend to have more sensor data available. While this is potentially a good thing, sensor data provides value only when it is being processed and interpreted properly, in part by the techniques described here. Testing of the methods, however, requires the “right” kind of data. Generally, there is a lack of seeded fault data which are required to train and validate algorithms. It is also important to migrate information from the component to the subsystem to the system levels so that health management technologies can be applied effectively and efficiently at the vehicle level. It may be required to perform elements described in this chapter between different levels of the vehicle architecture.
- Publication Name
- Publication Location
- Year Published
|719.5 KB||57 downloads|