Comment

January 2009

System Integration

Balance of power

Nickel producer generates power sharing, reduces blackouts with new control system

By Kevin Geraghty

In the central part of Sulawesi, Indonesia, in the middle of the jungle, sits an integrated mining and smelting nickel producer, PT Inco. A subsidiary of Vale INCO, the company’s key competitive advantage over other nickel producers is it owns and operates its own hydro power stations to produce the large quantity of electric power the process requires. With planned increases in production, the company realized it would have to update the control systems. The new unit would have to allow reactive power sharing. The company decided to replace the governor and control system on the plant’s hydro-electric generator.

The company operates an isolated power grid. Water shed from three large lakes supply the reservoir.

The generation system originally consisted of the Larona power station, commissioned in 1978. The station was equipped with three turbines coupled to 65MVA (megavolt ampere) generators. Due to the lay of the land, a seven kilometer-long intake canal has two radial intake control gates. The head pond at the end of the canal feeds three 1.3 kilometer long penstocks. The Larona units’ bypass valves protect the penstocks during load rejections. In 1999, the commissioned Balambano power station added two turbines coupled to 80.6MVA generators. In 2002, we upgraded the Larona 3 unit with a 68MVA generator. The new unit has significantly higher inertia than the original unit.

These five units comprise the dirty power supply, which supplies the four electric furnaces and other loads that can handle the bad power. The dirty power frequency will swing between 47 and 52 Hertz in what is typically a three-minute cycle. Recent recordings showed swings between 45 and 55 Hertz over a two-day period, without a trip.

Five 8MW diesel generators, 32 1MW diesel generators, and a 28MW steam generator provide the clean power. Two DC links enable the transfer of up to 38MW between the clean and dirty power grids.

The Larona units had electro-hydraulic governors, with relay-based unit controls. Raising and lowering the power setting and speed setting devices helped control the operating point of the generators. This would determine the opening position of the wicket gates. The Balambano units had system automation technique (SAT) controls for unit automation and a MIPREG digital governor. The main function of the SAT system at Larona was to monitor the status of the equipment.

Intelligent control system needed

The new unit control system would have to allow reactive power sharing, and the additional reactive loads would be pushing the generators closer to their maximum operating curves. Without an intelligent control system, one unit may be providing a disproportionate share of the reactive load. For improved water management, the system would have to be able to take direction from a higher level control. There would be no room for wasting water.

The system would have to integrate with advanced load and generation rejection schemes to reduce the likelihood of grid blackouts. Other requirements included reduced restoration time following unit trips or system blackout, enhanced monitoring, and intelligent alarming to assist operators in unit restoration after equipment failures or system disturbances, and a sequence of events (SOE) recorder for improved event diagnosis. The intent of the SOE recorder was to improve event diagnosis leading to the correction or elimination of preventable trips.

The governor system would have to protect the 1.3km penstock against over-pressure conditions. A failure of the penstock would be catastrophic. The governor must also protect the canal and head pond from over-topping under blackout conditions. When the units trip, seven kilometers of water come down the canal. For water management under unusual operating conditions, we need to bypass water from the Batubesi reservoir to the Balambano reservoir.

Shutdown, construction

A 21-day shutdown allowed for a furnace transformer replacement. Of the 21 days, there were 12 days for construction, commissioning, and startup. In conjunction with the PLC replacement, other required work included repairing cavitation damage on the runner, repairing the wicket gate bushings, modifying the intake gate controls, and upgrading the condition monitoring equipment. The remainder of the shutdown was to replace the SAT unit control PLCs on the Balambano.

Pre-shutdown was completed by working closely with operations and maintenance, installing new panels, cable trays, and cables, and terminating all cables in the new panels. We also prepared new cables for termination in existing panels. It installed HMI hardware and connected it to the network and identified equipment to be demolished as well as tie-ins for the hydraulic system.

The demolition had two stages: shutdown and post-shutdown. This involved removing equipment that needed existing real estate during the shutdown and disconnecting other pieces during the shutdown that we had to remove at startup.

One obstacle was a lack of communication, causing us to have to re-install items that had been demolished. We removed some wiring related to the synchronizer for Units 1 and 2 and had to hastily reconnect these wires when we discovered we could not bring the old units back online after stopping them.

Commissioning

In preparing detailed commissioning procedures, we included I/O designation, reference drawings, testing procedures, interlock requirements, HMI indications, HMI alarms and a sign-off for the electrical and instrumentation terminations.

During pre-shutdown, we brought several systems online and verified the Ethernet communications between the PLC and the HMI. We checked the hardwired connections and Modbus communications between the PLC and the governor as well as the Modbus communications between the PLC and the generator protection relay.

Once the shutdown began, we checked the hardwired connections as they became available and performed discrete system checks on the governor oil pumps, the generator lift pump, the cooling water systems, the wicket gate hydraulics, the bypass valve hydraulics, and the excitation system.

When it was time for startup, we only had to commission the integration of the discrete systems previously checked as stand-alone systems.

SNAFU in startup

Unit startup began with dry tests whereby we stroked the wicket gates and bypass valve to set the timing nuts on the distributor valves. The settings determine the maximum speed of the servo motors. We adjusted them to meet the existing speeds. During this time, the vendor provided hands-on training for the operations and maintenance personnel. Although the retrofitted hydraulics package is much simpler than the previous arrangement, this was all new equipment.

The next stage was to simulate the generator circuit breaker interlocks. Even though the majority of these interlocks were hardwired and had not been touched, we checked each one to verify there were no hidden traps. And even though we had gone to great lengths to update the drawings, several undocumented connections caused problems.

We filled the penstock to begin the no-load tests, which allowed us to check the penstock intake gate interlocks. With the penstock full, we manually started the auxiliary systems and began to ramp up the generator speed. Soon after the unit started rotating, we heard an unexpected sound. Another project had installed vibration monitoring probes but failed to check the clearances. A series of bolts on the shaft sheared the probes and sent them flying. Once we were satisfied we had found all the pieces, we once again began to ramp up the generator speed, stepping uneventfully through 25%, 50%, 75%, and 100% of rated speed, waiting 30 minutes at each speed to allow the machine to stabilize.

With the units up to speed, we performed the over-speed trip test. When the speed exceeded the limit, the wicket gates began to close, and the bypass valve began opening. We had configured it to open at a rate of 5% per second when the penstock pressure exceeded 16.3Bar and to close at a rate of 2.5% per second when the pressure dropped below this limit. This caused oscillations in the bypass valve, which resulted in pressure spikes over 20.5Bar, causing one of the rupture discs to let go. We now faced draining the penstock, replacing the rupture disc, and re-filling the penstock. This is typically a 12-hour process. We finally decided to revise the governor logic to control the bypass valve based on wicket gate closing rate.

To bring the unit online, we placed the unit control PLC in automatic and started the system. The PLC brought all the auxiliary systems while the governor brought the generator up to rated speed. The generator then manually synchronized and held at minimum load for 30 minutes. There was no indication of heat or vibration problems, so the load gradually stepped up to 60MW. The unit was very slow to take up load. We decreased the permanent droop, and the unit picked up a greater share of the load. We then took the unit off line to test the auto synchronizer. After some minor adjustments to the voltage settings and timing, we finally synchronized the unit. We then handed the unit over to operations and to the Balambano project. We had made it with just hours to spare.

Lessoned learned, future plans

Other than a few communication glitches, the project was a success. But we learned to check for a few things before the next project.

• First, know what you want from the system. Spend the time to do a thorough job of the functional description, and be sure to have operations involved. Drag them to the meetings if you have to.
• Second, remember to define the communication between the different vendor equipment. Do not just expect different systems to be able to communicate intelligently. Go to the factory to test the equipment before it is delivered. It is much easier to make changes in the factory.
• Third, give the operators a chance to get their hands on the system so they will gain experience.
• Fourth, do not be surprised when you discover existing equipment is not in good condition. Plan for it, and check the status of spares in the warehouse.
• Finally, spend the time to do as-built drawings before and after the project. This effort pays off in reduced construction time and reduced downtime.

Future plans

We plan to move the utilities operators into the same control room as the process plant operators to enhance communication. This project is planned to coincide with the upgrade of the Larona 1 & 2 units. In 2010, we plan to add two more units downstream of Balambano to increase our operating and maintenance flexibility. We are still in the planning stages of developing a high-level water management control system. The system will not be practical until we upgrade the remaining Larona units. We’ve also completed a preliminary functional description for a high-level volt ampere reactive sharing control system, which will also be practical after upgrading excitation systems on the Larona units.

ABOUT THE AUTHOR

Kevin Geraghty is a senior maintenance specialist at PT International Nickel Indonesia.

Hydropower Worldwide, hydropower plants produce about 24% of the world’s electricity and supply more than 1 billion people with power. The world’s hydropower plants output a combined total of 675,000 megawatts, the energy equivalent of 3.6 billion barrels of oil according to the National Renewable Energy Laboratory. There are more than 2,000 hydropower plants operating in the U.S., making hydropower the country’s largest renewable energy source. A power grid consists of a set of large power plants such as hydropower or nuclear power plants, all connected together by wires. One grid can be as big as half of the U.S. A grid works well as a power distribution system because it allows a lot of sharing. If a power company needs to take a power plant or a transmission tower off line for maintenance, the other parts of the grid can pick up the slack. The power grid cannot store power anywhere in the system. At any moment, millions of users could be consuming megawatts of power. At that same moment, dozens of power plants are producing exactly the right amount of power to satisfy demand. Transmission and distribution lines are also sending power from the power plants to the consumers. While this system is highly reliable, at times of high demand, the interconnected nature of the grid makes the entire system vulnerable to collapse. When that plant disconnects from the grid, the other plants connected to it have to spin up to meet the demand. If they are all near their maximum capacity, then they cannot handle the extra load. To prevent themselves from overloading and failing, they will disconnect from the grid as well. That only makes the problem worse, and dozens of plants eventually disconnect. That leaves millions of people without power. SOURCE: http://science.howstuffworks.com/blackout.htm