This thesis describes how system parameters in a manufacturing environment can be used to make judgements about the condition of the monitored machines. The two main areas of research described are firstly how the machine parameters are measured and processes and can then be used as variables to assess the condition of the machines. Secondly how to present this information to the machine operators and engineers who maintain these machines, in order to help them make decisions about the state of health of the equipment they are responsible for. The research is concerned with an investigation into the feasibility of offering a generic, reliable and practical approach to Supervisory or Condition Monitoring Systems. the system addressed has the ability to acquire data from a multitude of sensors, but simultaneously releases the engineer from the task of having to deal with the resulting data load. It provides user-friendly interfaces for human-machine interaction which support a user with the data analysis task and introduces new methods to automate and control the data acquisition and data validation process. Sensor priorities are a method to quantify the data acquisition time used to monitor a machine component. The approach introduced here allows the system to adapt to varying operating conditions by modifying the sensor's acquisition priority, hereby increasing the level of alertness through increased data acquisition from the sensors with higher priority within the system. If the machine component is in a healthy state the priority of its sensor is low, if tolerable levels for the parameters that indicate the health have been exceeded, the priority will be high. this sensor priority is then related to the data acquisition frequency to increase the total time that a machine component is monitored. Using this method, the 'attention' that a component receives from the the condition monitoring system is directly related to its condition. The research presented in this document shows that it is viable to implement dynamic sensor priorities into the data acquisition module of a condition monitoring system. Furthermore it can be shown that an object oriented graphical user interface, built in accordance with guidelines recommended by the Human-Computer Interface research community can simplify the task of a new user getting acquainted with a system and encourages the user to use the system with confidence.
|Date of Award||2000|
- Nottingham Trent University