George was conducting an annual safety inspection at a dam, looking at the condition of the rip-rap. The dam impounded water for a hydroelectric facility that was one in a series, all owned by the same company, on a large river system. In the course of the inspection, George casually asked Tom, the plant manager, how the system was working. Tom said overall everything was fine. However, he stated that the company’s new power plant at Slate Canal was “twitchy.” George asked what he meant by this, and Tom explained that a minor system disturbance would cause the single 40-mw unit to drop off line on a fault.
Further questioning by George revealed that the turbine-generating unit at Slate Canal had been operating for slightly less than a year. During this time, the facility experienced several minor frequency deviations of less than one quarter cycle due to events such as lightning strikes, wind storms overturning trees onto the transmission lines, and other distribution disturbances. During every incident of this nature, the unit at Slate Canal dropped off line a few minutes after the frequency had returned to normal. After the second incident, company personnel carefully inspected the unit controls and governor and determined all parameters were correctly set.
This situation puzzled George. He was familiar with the system because he had worked on several of the company’s power plants. However, nothing in his experience with this system could explain the events at Slate Canal. George asked about the reason the unit dropped off line, and Tom said he would ask for the supervisory control and data acquisition (SCADA) report.
A few days later, Tom advised George that the report indicated the unit always tripped on low governor oil. After thinking about the circumstances for a few minutes, George told Tom that he had encountered a similar problem on a much smaller remote system a few years before. The previous situation involved a new Kaplan turbine and was the result of incorrect setting of the Kaplan blade timing (see “Lessons Learned,” July 2001, page 36). But this could not be the cause of the problem with Tom’s system because all the power plants contained Francis units. Tom then replied that the Slate Canal unit was the company’s only Kaplan!
Armed with this new information, George thought he could solve the problem. George told Tom he suspected that the Kaplan blade timing at Slate Canal had been set at close to the same timing for the wicket gates. He believed this combination usually resulted in “hunting” of the blades and wicket gates, causing the governor oil to be rapidly used up. Consequently, the unit could trip off line on low governor oil. A small system load change that called for a new wicket gate opening could trigger the incident.
Tom would have to determine the timing of the Kaplan blade movement and the governor wicket gate open-close times at Slate Canal. Before leaving the site, George asked Tom to call him with the blade and gate timings.
A few months later, after an extensive search of the commissioning data, Tom called George to provide the timings. The wicket gates at Slate Canal were set to open in 15 seconds and close in ten seconds, and the Kaplan blade timing was about 13 seconds. Based on this information, George advised Tom to change the blade timing to four times the longest wicket gate time, or 60 seconds. Tom said his company had established a committee to look into the problem, and he would advise them of George’s suggestion.
About a year later, Tom e-mailed George to update him on the situation at Slate Canal. The company had changed the blade timing as suggested, and the unit had been operating for several months with no further twitchy incidents.
For unknown reasons, commissioning data rarely mentions Kaplan blade timing, so the turbine erectors usually set the time to fall somewhere between the wicket gate times. The company spent several days looking through commissioning records to determine the blade timing at Slate Canal. Even then, this data was not certain because it had not been verified by the plant operators. Re-setting of the blade timing had to wait until the next annual dewatering for maintenance.
In a Kaplan unit, gate-blade hunting occurs when the gate and blade timings are close. The situation happens when there is a call for a new gate opening (say larger). The governor moves the gates to a larger opening, and the blades also move open to a more efficient operating point. This causes the wicket gate opening to overshoot the correct setting. The governor tries to correct with a smaller opening on the wicket gates, and hunting begins.
By slowing down the blade movement, the Kaplan unit operates temporarily as a propeller with the wicket gates rapidly moving to the new larger opening. The blades slowly move open to the new, more efficient operating point, and the governor has time to slowly close the wicket gates, avoiding hunting.
Always make sure that Kaplan blade timing is about four to six times the effective wicket gate movement timing. The effective time is measured in the dry and is taken as twice the time recorded for the wicket gates to move from 75 percent open to 25 percent open when closing from a full open position. The Guide to Hydropower Mechanical Design features a brief discussion of Kaplan blade timing (page 18).
— By James L. Gordon, B.Sc., hydropower consultant, 102 St.-John’s Boulevard, Pointe Claire, Québec H9S 4Z1
Canada; (1) 514-695-2884; E-mail: [email protected]