It is a high time for Nepali hydropower plants to maintain a proper database, especially that of old ones. This includes complete drawings, circuit diagrammes, data log from the sensors and documentation of any event in the plants. Archiving such data improves possibility of carrying out research works in the future
The condition of any machine degrades with time. Just as our body needs regular checkup, the health of any machine also needs regular monitoring so that any fault in it can be detected in the early stage.
Hydropower is one of the major industries of Nepal, where the application of condition monitoring is relevant. The turbines in Nepali hydropower plants, for example, undergo continuous wear, mainly because of the erosion due to excessive sand in the rivers.
This further aggravates other electro-mechanical components of the plant, leading to premature failure of the machines. By incorporating various sensors in the plant, the lifespan of its components can be extended through preventive maintenance. Besides, it can also prevent any accidents in the power plants caused by gradual degradation of the machines.
Some basic condition monitoring tools can be observed in most of the existing power plants. The most common one is the temperature sensor, which identifies any fault in the machines when the temperature exceeds the normal range. The temperature of the bearing, for example, increases when the shaft is misaligned or when vibration occurs in the turbine due to any fault. Similarly, pressure sensors can be installed at several regions around the turbine to measure the pressure of the flow.
Some of the other signals may include water level, vibration level, power output, speed of the turbine's rotation, leakage and position.
These days, more emphasis is given to digitalisation of the output signal, so that the condition of the machines can be remotely monitored and stored for future reference.
In cases where any fault occurs in the power plants, alarming signals are received, which are developed through pre-programmed sensors. In the case of analog signals, which are installed in most power plants of Nepal, the signals need to be continuously and manually monitored by the operators. The actions taken by the operators in cases of fault are normally shutting down the machines and inspecting the possible reasons for the fault signals. Apart from some periodic maintenance of the turbines, it is normal for the hydropower plants to adapt run-to break strategy, which means that any component runs until the fault reaches its final stage.
A state-of-the-art in hydropower plants worldwide is using Artificial Intelligence to detect the fault in the machine at an early stage, so that necessary interventions can be taken to minimise the damage to the components. This is again a perfect analogy to diseases of the human body.
The discussion on the need of condition monitoring and digitalisation of hydropower plants has come up favourably in the recent days in Nepal as well. New hydropower plants, as well as those that are being refurbished, have been readily implementing condition monitoring tools of the turbine components. However, there is still a long way to go for utilising the data obtained from these sensors to conduct a fault diagnosis using AI.
In the present context, the digital data are being used more or less for the same purpose as those produced by the analog sensors.
Nevertheless, since data from the sensors are stored, there is a good possibility to conduct a trend analysis of the signals in various conditions, so that various case scenarios can be developed using AI algorithms.
Undoubtedly, applied engineering studies and problem-oriented research in universities can have a significant impact on bringing these concepts into reality. The Turbine Testing Lab of the School of Engineering at Kathmandu University (KU) is currently conducting a project in collaboration with Nepal Electricity Authority (NEA), where the aspects of condition monitoring and fault diagnosis of the turbines are included.
Similarly, a MoU was recently signed between KU and Independent Power Producers' Association, Nepal (IPPAN) to uplift the academic-industrial collaboration in the hydropower sector. This is a sign that both problem-oriented research and research-oriented solution are welcomed by academicians and professionals respectively.
It is a high time for Nepali hydropower plants to create/maintain a proper database, especially that of old power plants. This includes complete drawings, circuit diagrams, data log from the sensors and documentation of any event/incident that occurred in the power plants. Archiving such data improves the possibility of carrying out research works and solving any problems that may occur in the future. For old power plants, where the drawings of the turbines are not available, it is also possible to re-design the components by using reverse engineering techniques, or knowledge of the turbine design obtained from the university. This will ultimately ease the process of fault diagnosis, since proper modelling of the power plant, with know-how of its design, structure, location, orientation and so on can be studied in advance.
Research works related to real-time condition monitoring of hydropower plants for early fault detection and diagnosis can be largely relevant in the context of Nepal, which faces significant operational and maintenance challenges.
These challenges are the results of low quality products, lack of skilled manpower and spare parts for repair/maintenance, excessive sediment-contained flow and extreme geographical and climatic challenges of the country in terms of flood and landslide.
Owing to high sediment content in the rivers and the fact that most of the turbine components are damaged due to its impact, some additional condition monitoring tools might be appropriate, such as sensors to detect sediment concentration, or analysis of the sediment size and mineral content in sand. If we could start to think of conducting useful research in the field of surveillance of the power plants, a lot can be done to minimise their effects.
Chitrakar and Lama are with the Turbine Testing Lab, Kathmandu University
A version of this article appears in the print on April 25, 2022, of The Himalayan Times.