Local real-time monitoring for data collection is becoming very common in today's information age.
Gathering data in the work place either by intelligent device or by human interface is used for setting control limits, trending, long term records, regulatory compliance, book keeping and others.
And now advances in technology are allowing that data to be interpreted by computers. B.C.
Hydro is currently using this kind of expert system programming for simple and complex tasks
performed by the utility.
Expert Systems
Expert systems are a practical application of Artificial Intelligence (AI). AI is the branch of computer science which is concerned with the use of computers to solve problems normally thought to require human intelligence, such as the comprehension of language or complex decision making.
The expert system is a specialized area of AI in which computer software is used to tackle the problems normally executed by highly educated human experts, such as medical diagnosis or weather forecasting. In the beginning, expert systems were used quite successfully to assist geologists in locating commercially viable mine sites. Another early expert system was used in the food processing industry to provide expert assistance in troubleshooting food product sterilizers. These expert systems will often outperform the average worker, and sometimes even the best experts. They have shown a significant reduction in down time.
A number of electric utilities use the expert system technology because it reduces the number of
brownouts. Expert systems work, and save money, due to the fact they can accumulate and
organize the expertise needed to keep operations running, and make that knowledge immediately
available even to less experienced workers when problems occur. In a sense, an expert system is a
great "leveling tool", since it gives all workers access to a knowledge base that enlarges his or her
personal resources to fit the requirements of the tasks at hand. The biggest difference between the
diagnostic process for a person or a machine is that the oppressive monotony of the task is not a
factor for the machine.
An Expert System Based PLC Training Course
In 1995 B.C. Hydro and a local educational institution developed an in-house Programmable Logic Controller (PLC) training course which is expert system based. It is currently being enhanced to allow this training course to be delivered via Internet with a remote chatting teaching application.
Within B.C. Hydro the need of PLC training for electrical trades was identified by trade skill development professionals. A computer based PLC tutor using the expert system technology for the air blast circuit breakers air dryer system was adopted for individualized on-site training. This enables any student to learn whenever ones schedule permits him or her to use the host computer. It is low cost and practical, allowing students to review and practice their skills after the training is completed. The training is thorough enough to transfer knowledge for specific equipment application.
The PLC replaced the electro-mechanical relays for air-dryer control and monitoring functions, which is a portion of the air supply system for air blast circuit breakers in the substation.
The objective of the training course is to allow very experienced utility electrical trades people an opportunity in trade skill development with this standardized equipment. With this training, trades people moving from one location to another do not need to be retrained, and in an event of equipment difficulties, all trades people can help each other out. The expert technology is used to complete the simulation of B.C. Hydros modern air dryer functions. It also allows the student to learn the PLC operation and to acquire simple knowledge of programming.
The simulation program interacts with the student, and during the exercise session it will inform the user when student designed and implemented air dryer system functions are incorrect. The student can learn from previous mistakes and progress through the course with the personal computer. The structure of PLC training program includes PLC simulation software and expert system overview block diagrams.
The need for the human instructors knowledge is essential in any learning process, to guide and
teach students in problem solving, to monitor progress and to evaluate the results. In this context,
the Internet might be a solution for remote off-site interactive teaching employing a low cost
delivery method. This would resolve the training site selections problems, since utility substations
are spread out in a vast geographic area. With some means for remote learning, the instructor
could simultaneously correspond with many sites in the province on-line in real-time. The Internet
could also be used for downloading PLC training software to the users, and for directing E-mail
to an instructor when students have problem doing assignments. Setting up a chat session, in
which all the students can chat over the net, would promote group dynamics. This will eliminate
or minimize the expenses for such things as traveling, lodging and meals since there is no need to
the entire group to be in one classroom. The down side of this approach is we lose the basic
human contact which is essential in team work.
An Expert System Based Machine Vibration Monitoring System
Machinery specialists for large rotating equipment are usually not readily available when required.; many operating plants are in remote location and do not have on-site equipment specialists. Expert systems can automatically perform some of the functions normally done by human experts. Commercially available computerized diagnostic expert systems, when implemented, can reduce the time and effort required by equipment specialists by comparing data with rule sets generated by human equipment experts, incorporating many years of machinery diagnostics expertise.
B.C. Hydro in co-operation with Bently Nevada have done an eight month machine audit on a 150 MW steam turbine generating unit at the Burrard Thermal Generating Plant. The turbo generator was originally equipped with simple moving coil vibration detectors, as part of the original Turbine Supervisory Instrumentation (TSI) package. The detectors' signals are used for chart recorder trending, general annunciation, vibration indication and conditional unit trip on excessive vibration. The old TSI provided adequate unit protection, but it was not capable of providing sufficient information for a condition dependent operation and maintenance strategy that must utilize continuous real-time monitoring and diagnosis. In 1994 the machine was retrofitted with proximity probes in XY configuration for each bearing and at various locations of the machine train for all units. Computerized permanent TSI systems were installed to collect real-time data on vibration, temperature and other pertinent information.
The audits were performed remotely by dialing an on-site computer configured with customized expert software. They were performed on data stored within the on-site expert program computer, which was connected via network interface to the plant data acquisition system network that collects real-time data from the units. The machine train monitoring points are shown in figure 3. The expert system computer automatically acquires data from the real-time machine monitoring computer daily at a specified time. The expert program evaluates the machine conditions for magnitude of the vibration, frequency of the vibration, form of the vibration, amplitude/phase angle, rotor centre line position. It processes variable information, machine geometry information, trend file information and machine history file.
Custom rule sets were used to evaluate the information, perform the off-site diagnostics and
produce reports. The structure of knowledge bases is shown in figure 4. The overall flow of data
through the expert system is shown in Figure 5, it consists of functions for data acquisition,
information extraction, machine parameter, expert program and remote dial-up connection.
Reports are generated at various levels to meet the different needs of the users (from the
operators to the managers as well as the machinery specialists). With this time saving process to
digest and interpret the information, it enables the human to make the final decision and take
action at a much higher confidence level, resulting in more efficient use of the work force and
achieve higher productivity.
Case Example
On 27 February 1995 the unit tripped due to high temperature at bearing number five at approximately 2:30 p.m.. Reference to the expert system audit performed early the same morning indicated that no machinery malfunctions were present, and that vibration and position measurements showed no significant changes during the four weeks prior to the trip. This lent support to the possibility that the trip was caused by a faulty instrument wire, diagnosed off-site by a human expert. Subsequent on-site investigation by an instrument technician confirmed the conclusion drawn with the expert system generated records.
Other potential utility expert system applications include transformer condition monitoring and diagnostic, circuit breaker performance evaluation and information management.
A low cost simple expert system can faithfully perform any simple task, they can be created by engineers, electricians and managers for daily use. Basically, the monitoring devices have to be programmed to collect the right data, this data would have to be transmitted to a site with the resources (human and otherwise), to use that data to solve problems and then solution would have to be transmitted back to the site with the problem and acted upon by the local people.
The low cost simple expert system allows us to move data and not people. It is cost effective because we are changing from data collection and off-line investigation to automatic continuous machine or equipment condition monitoring. This is a key factor for a successful condition dependent maintenance program and provides faster response to the problem. Deployment of this low cost technology is practical and feasible.
In the 1990's business environment, utilities are one of many industries being faced with
downsizing, re-organizing, and to do more with less concepts. In real life, our customers
demand low cost, better quality and reliable service. Expert systems may be one of the solutions
to support the utilitys corporate objectives and satisfy the customers needs, when human experts
are disappearing or no longer available within the work place. They increase our effectiveness,
both individually and collectively, by preserving the expert know-how and distributing the
knowledge effectively by modern technology almost without boundary.
Alex Lam, P.Eng. is a senior engineer at B.C. Hydro and Power Authority