The Challenge:
Creating a thermal management system that reduces cost and development time for a Serbian power plant.
The Solution:
Using the NI wireless sensor network (WSN) platform and an existing NI LabVIEW software application to reduce development time and avoid complicated, costly cable installations.
Author(s):
Dr. Aleksandar Nikolic -
Nikola Tesla Institute of Electrical Engineering, Belgrade, Serbia
Dr. Aleksandar Zigic - Nikola Tesla Institute of Electrical Engineering, Belgrade, Serbia
Nikola Miladinovic - Nikola Tesla Institute of Electrical Engineering, Belgrade, Serbia
A generator power transformer is the largest unit in a power plant. It can have a capacity of up to 1,400 MVA. There are two ways to thermally manage this type of transformer: One expensive solution involves sensors mounted in the transformer windings, while the other is based on calculating the transformer’s highest temperature (for example, the hot-spot temperature) using the measured transformer’s top oil temperature and load current.
Quick System Setup
We needed to configure our system for the largest generator transformer in Serbia (725 MVA) within a short time period. We chose NI CompactRIO and LabVIEW to replace a custom digital signal processing system. We used an NI cRIO-9073 system with NI 9217 modules and Pt100 sensors. We developed a LabVIEW algorithm with real-time deployment based on the IEC 60076-7 standard and fine-tuned it to perform all measurements, hot-spot estimations, and transformer cooling system control (oil pumps and fans). We developed the whole system based on CompactRIO in less than two months.
WSN Application
For another thermal power plant where cabling costs were too high due to complicated installation requirements, we improved two generator transformers (210 MVA and 100 MVA) based on an NI WSN. This sensor network consists of one NI WSN-3202 and three NI WSN-3226 nodes per transformer. The WSN-3202 performs transformer current measurements on one analog input and controls four cooling units via digital outputs. The WSN-3226 nodes are connected to Pt100 sensors that measure top oil temperature, cooling unit input and output temperatures, and ambient temperature. Nodes that collect sensor and relay signals from one transformer are placed in one enclosure (see Figure 1).
Data Collection
One transformer uses an NI WSN-9791 Ethernet gateway to collect node measurements and the other uses an NI 9792 programmable gateway. We also used an NI 9792 in LabVIEW application deployment. We significantly reduced our software development time, to no more than two days, because we used a LabVIEW application previously implemented on an NI 9073 CompactRIO system with only minimal changes (see Figure 2).
Web Development
The WSN-9791 and NI 9792 are connected via Ethernet to a PC installed in the power-plant control room. We developed the application in Microsoft Visual Studio and run it on a PC communicating with gateways via MODBUS TCP/IP. It displays all necessary operator data and stores values in a MySQL database. To provide remote supervision and application modifications, the whole system is connected to the Internet via the power plant’s local area network (LAN).
Conclusion
For a generator power transformer thermal management system, we created an application based on an NI WSN and LabVIEW. We selected NI hardware for its simple installation and mounting options and easy configuration and scalability. A modular system also gives us the flexibility to make changes not only during the development stage, but also during installation. Compatibility between NI platforms (CompactRIO and WSN) significantly reduced our development time. Because the WSN is connected via Ethernet to the power plant’s LAN, we can also remotely program and supervise the system via the Internet, reducing maintenance time and costs. We installed this system last summer when temperatures outside were more than 40◦ C (104◦ F). The system has never stopped working even in the worst conditions during the winter when temperatures drop below -25◦ C (-13◦ F).
Author Information:
Dr. Aleksandar Nikolic
Nikola Tesla Institute of Electrical Engineering, Belgrade, Serbia
Koste Glavinica 8a, 11000
Belgrade
Tel: +381-64-8259751
anikolic@ieent.org
Print
PDF
