"George, Dyke 2 at Bear Rapids just washed out. Can you come fix it?" Ulrich was calling from a mining town in the northern part of his country, where electricity was provided by a series of hydro plants on the Bear River. George knew the system because he had worked on one plant many years ago and for the last few years had undertaken dam safety inspections on all four facilities. He had warned Ulrich that the crest of Dyke 2 on the third plant downstream from the main reservoir was low, but repairs had been postponed because of budget constraints.
George arrived and assessed the situation. Water was pouring through a breach, and it appeared that the glacial till foundation was being eroded deeper with each passing hour. Unfortunately, the flow could not be shut off by lowering the headpond because it was controlled by a long concrete weir and the power plant was operating at full capacity.
When the mine was started many years earlier, a dam was built where the river dropped off a plateau, creating a large reservoir. A stoplog spillway was included to discharge floods, but it was rarely used. As power demand at the mine increased, other plants were added downstream.
A long series of dry years, coupled with high power demand, meant the stoplog spillway was not used. However, recent warmer temperatures had resulted in a wetter climate, slowly filling the reservoir. The last year was particularly wet and, when the spring flood arrived, the stoplog spillway had to be opened.
Instead of employing a contractor experienced in hydro, the mine owner used mine staff and rented equipment to build Bear Rapids. Thus, costs had exceeded the budget. Shortcuts were instituted to save money. For example, because Dyke 2 was mostly a freeboard dyke, it was decided to simply build up the access road at the dyke location instead of first excavating down to an impervious foundation.
George was not aware of this history, but he had observed some settlement of the dyke crest during the safety inspections. Because of intermittent permafrost around the area, George assumed the settlement was caused by melting of permafrost in the foundation as warm headpond waters encroached on the foundation.
The Bear Rapids headpond spillway is a long weir. With no flood waters being released from the reservoir for years, the weir had only overflowed by a few centimeters for a few days each spring, discharging the small local flood. The dyke foundation was at about weir crest level, and there were no provisions to measure seepage at a downstream weir. Brush had not been cleared from the downstream slope, so it was difficult to inspect for seeps. And George's inspections occurred in the summer, after the spring flood had passed. Thus, he had not seen any seepage.
So why did the dyke wash out? It became apparent that when the headpond was spilling over the years, water had flowed through the pervious road gravel in the dyke foundation and eroded the underlying till. With time, the annual erosion became more extensive, until the water path was below the spillway crest and the erosion became more continuous and severe. The dyke breached when the large flood from the upstream reservoir raised the headpond water level to about a meter above the spillway weir crest, creating enough extra head to institute a piping failure.
Repairs consisted of placing rockfill in a semi-circle upstream from the breach, dumping glacial till on the upstream face of the rockfill until the leakage subsided enough to be handled by pumps, removing the road fill and overburden down to a till foundation, and building a homogeneous dyke section on the original road centerline.
The plant staff is into the third generation, with the second generation having retired only a couple of years ago. The new generation is young, with minimal hydro experience. Unfortunately, maintenance and safety management procedures were not passed on to the new staff. Thus, they did not appreciate that dams and dykes need to be constantly monitored and signs of distress — such as settlement, cracks, or seepage — assessed immediately. Also, downstream slopes on dams and dykes need to be kept clear of brush, and seepage weirs should be installed to assist monitoring and inspections.
There have been several cases where dyke failures occurred after operators retired and the new operator was not aware of past practices (see "Lessons Learned: The Case of the Persistent Sinkhole", HRW, November 1999).
Hence, it would be prudent for anyone preparing to retire to develop a report on operations and maintenance practices and surveillance procedures, with emphasis on unusual observations or conditions.
— By James L. Gordon, B.Sc., hydropower consultant