Rule engines

Rules are generally implemented in something called a rules engine, which provides a basic framework for writing rules and then for running them in the manner described above. 

Ghosts Double images in the output of a grating, the result of imperfections in the ruling engine used in its preparation. 

There are several advantages to using rule engines in expert systems. 

In any situation, where no conventional programming approaches exist to solve the problem, rule engines might be an alternative. Here are some typical reasons to decide for a rule-based approach. 

Each JESS rule engine can incorporate a series of facts, which are stored in the working memory. Rules react to additions, deletions, and changes to the working memory. JESS distinguishes between pure facts, which are defined by JESS, and shadow facts, which are connected to Java objects. 

With all these subtle considerations in mind, a designer can hopefully pick a more appropriate inductor current rating for his or her application. Clearly, there are no hard and fast rules. Engineering judgment needs to be applied as usual, and perhaps some further bench-testing may also be needed to validate the final choice of inductor. 

The interface between the DCS and Systeml server (rules engine) uses the Dynamic Data Exchange (DDE) protocol. On the other hand, the interface between the rules engine and the Mark V as well as that between the rules engine and the monitoring server uses OPC (Open Process Control) whieh is based on a server-client architecture protocol. The rules engine can also interact with input files via API (Application Programming Interface) to read manually input data that is not available from either the Mark V or DCS. 

The monitoring system acquires and stores data from its sources at the specified rate of once per minute unless a higher sampling rate is needed (up to once per second). Certain rules, as we will discuss later, require this to process a buffer of data within a minute or several minutes. The rules engine is so powerful that it has tens of mathematical functions stored in libraries in the software. In addition, it can interact with DLLs compiled using a programming language such as C++. 

When a spread is deteeted on a multi-can gas turbine, it is not straightforward to judge the source of the problem (faulty combustor). This is the case, because the thermocouples are not placed adjacent to the combustor cans. Hence, a rule is developed to trace back the anomaly at the exhaust to the faulty combustor. This was done using an experiment during a shut-down period to simulate faults in one combustor and record the thermocouple readings at different speeds and discharge pressures. A correlation was obtained, which can be used in the rule engine to identify the faulty combustor in the event of a real spread. 

Instead of blindly looking at the shaft vibration and comparing it to a fixed threshold, we propose to identify the parameters influencing this vibration. Three significant parameters were identified oil temperature, speed, and load. Using historical data of the machine to be monitored or, if not available, of other machines of the same type, we can correlate this vibration with the three parameters. This results in a regression formula that can be coded in the rule engine to be applied on-line. This correlation study will be the focus of a future paper. 

---