Civil experiences - review of Waterpower XIV3 August 2005
Waterpower XIV provided delegates with the opportunity to share experiences and investigate new technologies through a series of technical presentations, roundtable discussions and highly-focused symposia. Carrieann Davies reports on the paper presentations focusing on the rehabilitation and maintenance of civil structures, and discusses the experiences shared
THE vibrant city of Austin in Texas, US, which boasts such famous residents as sporting legend Lance Armstrong, musician Willie Nelson and the late Stevie Ray Vaughan, played host to this year’s Waterpower XIV conference and exhibition. With the city motto being ‘Keep Austin weird’, together with its abundance of tourist attractions and lively nightlife, the location promised to be an interesting venue for this event.
Held from 16-22 July at the city’s Hilton hotel, the conference offered delegates a number of different sessions to attend, including four symposia focusing on new developments in hydro, rehabilitation, enhancing environmental performance and ways to operate your business. It also featured poster galleries, technical presentations, interactive roundtables and a new ‘Hydropower Basics’ training programme developed by the Hydropower Training Institute.
During one particularly interesting session, speakers shared their experiences on repairing and maintaining civil structures. One paper – written by Richard Steffens and J Christopher Ey of Devine Tarbell & Associates, and Ernest D Brokman of Duke Power – talked about the dam stability and flood remediation programme being implemented by Duke Power on ten hydro stations on the Catawaba river, in North and South Carolina, US, with the aid of engineering staff at Devine Tarbell & Associates. Presented by Steffens, the paper focused on remedial work at Oxford hydro station – an important project as it not only addressed dam and public safety aspects, but it also achieved enhancements in environmental aspects and recreation access.
Oxford hydro facility
Located on the Catawba river in North Carolina, Oxford hydro power facility was first operated in April 1928. The project consists of a dam impounding Lake Hickory, a power house with an integral intake structure, a gated spillway, and concrete gravity non-overflow bulkheads at both ends of the spillway.
In 1991, Duke Power retained the services of an independent consultant to update the hydrologic and hydraulic analyses of the Catawba/Wateree river system basin projects, which were originally performed in 1968. The work – which consisted of a watershed analysis and the development of a calibrated and verified hydrologic/hydraulic model of the watershed – indicated that remedial work would be required at the Oxford dam where structures would be overtopped by approximately 3m during the probable maximum flood (PMF). The duration of the overtopping coupled with the depth would result in significant damage to the facility and possibly breach the dam.
An Inflow design flood (IDF) study, performed using the National Weather Service’s DAMBRK computer simulation model, indicated a full depth breach would produce significant incremental inundation along the river channel and could result in increased flood elevations at downstream dams.
DTA engineers analysed multiple scenarios for remediation. The recommended PMF routing scenario included structural modifications to the Rhodhiss dam immediately upstream and construction of flood walls on the south bulkhead and earth dike, and lowering of the north bulkhead at the Oxford facility.
As a result, site investigations were carried out in 1998, which included core drilling the concrete structures, foundations and the North and South abutments. Material properties tests were performed on selected core samples. The investigations showed that the concrete structures are founded on bedrock consisting of two major types of poor to good quality biotite gneisses separated by an inactive fault near the centre of the gated spillway. The fault does not affect the stability of the structures, however, the final design included installation of tendon anchors in the fault zone that were adjusted to keep the anchor bond zones in more favourable rock.
Stability analyses for each water-retaining concrete structure were performed by DTA staff using water loads from hydraulic model studies plus material properties and shear strength parameters from the site investigation programme.
According to the paper the final design for the Oxford hydro station flood remedial work consisted of the following:
• Multi strand, epoxy coated, double corrosion protected tendon anchors: 19, 58-strand anchors in the south bulkhead; 40, 58-strand anchors in the gated spillway (four per bay); 11, 58-strand and two 44-strand anchors in a new emergency spillway.
• A 436m3 concrete buttress, with three 40-strand anchors constructed against the downstream power house wall between the two draft tube openings in 10m of water.
• Emergency spillway created by removing 4.6m of the existing north bulkhead and constructing a concrete ogee on top of the remaining bulkhead. This included 8410m3 of concrete and approximately 7264m3 of earth excavation and fill.
• Training wall at the north end of the emergency spillway. A soil nail stabilised wall with a two layer reinforced shotcrete face 91m long with a maximum height of 16.8m and surface area of 883m2 that included 7646m3+ of excavation.
• 2.7m high reinforced concrete floodwall on the power house bulkhead and south bulkhead. Total length of 135.3m.
• Reinforced concrete floodwall and footing on the south abutment, 38m long and varying in height from 1.8m to 3.6m.
• Sheet pile floodwall extending 21.3m from the end of the abutment concrete wall to where the abutment is about 0.6m above the PMF elevation.
• New concrete overlay on the power house decks and south bulkhead surface plus new handrails.
The total cost of the project was approximately US$14M. The work was completed on schedule and is now in compliance with the FERC guidelines concerning stability during flood conditions.
In the paper ‘Self excited penstock pressure oscillations at Gordon power station in Tasmania and other similar events’, Keith Caney and Enes Zulovic from Hydro Tasmania described their experiences with self excited and forced resonance in hydro power penstocks, and discussed the causes of penstock resonance and the remedial measures that can be taken to manage this problem.
Presenting the paper, Caney talked about Hydro Tasmania’s experience in dealing with a major forced outage at the 450MW Gordon station due to severe penstock pressure pulsations. The 3 x 150MW station, the largest of Hydro Tasmania’s facilities, is an unmanned station operated remotely from Hobart. A serious self-excited pressure pulsation incident occurred in the penstock system in August 2002, shortly after the station had been completely shut down. As a result, the power station was immediately taken out of service to investigate this potentially catastrophic problem.
A comprehensive root cause analysis was conducted to determine possible causes of the event. Caney and Zulovic say that it was immediately obvious that the penstock had resonated at its first natural frequency, which can be observed as a damped 0.84hz oscillation after every machine shutdown. For this mode the pressure wave would have travelled from end to end of the penstock, affecting all three machines almost simultaneously and reflecting from the intake gate.
The authors write that early in the investigation the basic cause of self-excitation was attributed to the existence of a water leakage path from the penstock, however, it was very difficult to find the source, which could have been on any one of the turbine TIVs or even on a small auxiliary hydro generator also attached to the penstock. After a process of elimination the investigation finally focused on the possibility of: (i) leakage past one of the service seals at its main sealing face, (ii) leakage past one of the D-rings fitted to the sliding service seal assembly or (iii) malfunction of a complicated control valve controlling water supplies for opening and closing each service seal.
A series of site tests were done to find evidence of leakage due to (i) and (ii). Water leakage past the service seal on each machine was measured with the seal fully pressurised in its closed position, but no significant leakage was found. The D-ring seals were found to hold pressure very well. The control valves in (iii) were all tested and one of the control valves which gave evidence of very slight sticking was dismantled and inspected. However, no evidence of leakage or malfunction was found.
An attempt was made to reproduce the original penstock pressure pulsation under controlled test conditions. Tests were conducted by reducing pressure in the TIV service seal closing chamber. In case of any emergency, a contingency plan of action was developed.
A pressure pulsation was eventually excited, but not at the first natural frequency of the penstock. It occurred at a frequency of 2.12Hz and was identified from site recordings as a local resonance between units 1 and 2, which had little effect on unit 3 or elsewhere in the penstock system. According to the authors, this highlights the fact that many resonant frequencies exist within a complex penstock system and a resonance event may not be detected by a single point penstock pressure monitor. As such, it was decided that the Gordon machines should be protected on a per unit basis.
Because it was impractical to start dismantling the equipment in the hope of finding more evidence, it was decided to manage the problem by developing operational strategies to avoid and/or respond to a similar event; installing a suitable automatic penstock pulsation protection system (PPP); and planning future investigation and dismantling of equipment at a more convenient time.
Subsequently a leakage test was repeated on the service seal of unit 2 and a large leakage flow was detected on one test only, which, for the first time, suggested a cause for the original event. However, since the evidence was inconclusive, it did not change Hydro Tasmania’s approach to managing the problem as a means of returning the station to service, after which the suspect service seal could be dismantled for inspection when convenient.
After further investigation, it was found that the damaged sealing face on the Unit 2 service seal was the root cause of the original event. It appears that the seal control system had operated correctly and the service seal had closed with full holding pressure. However, the seal had seated badly on the rough surface and was effectively sitting on some high points. This allowed a leakage flow and also introduced enough flexibility to allow the service seal ring to vibrate in its closed position. The penstock pulsations eventually stopped of their own accord when vibration allowed the seal ring to move a little and find a better seating position, which reduced or stopped seal leakage flow and removed the excitation source.
Another paper featured in the session looked at a state-of-the-art grout curtain test project used for the Chicagoland Underflow Plan McCook Reservoir in the US. Written by Cary Hirner and Norman Holst of black-veatch, together with William A Rochford of the US Army Corps of Engineers and Faruk Oksuz of Black & Veatch, the paper ‘State-of-art grouting for a groundwater barrier’ explained how the McCook reservoir project included a fully automated grouting control and data collection system that allowed evaluation of water pressure testing and grouting operations in real time.
The project is one of three planned reservoirs in the Tunnel and Reservoir Plan (TARP), which incorporates 175km of tunnels in the Chicago area. Upon completion the McCook reservoir will store combined sewer overflow (CSO) conveyed from the TARP mainstream and Des Plaines deep tunnel systems until floodwaters recede and the CSO can be pumped to a nearby wastewater treatment plant.
The project design includes tunnels, shafts, chambers, buildings, gates and valves. It also features an overburden cut-off wall and a grout curtain around the reservoir parameter to control groundwater infiltration and CSO exfiltration.
The reservoir will be excavated primarily in the dolomitic Silurian formations that form the bedrock in the Chicago area. The underlying, more shaly formations of the Ordovician Maquoketa Group are present below the floor of the reservoir. The jointed Silurian dolomite formations have been subject to some solution by groundwater and form the regional shallow bedrock aquifer.
The primary objective of the McCook reservoir grouting test was to discover the effectiveness of grouting in this environment while determining the most appropriate drilling method to complete the perimeter grouting of the proposed reservoir. This included grouting of the upper bedrock and integration with the overburden cut-off wall. The project was also conducted to evaluate drilling methods and grouting parameters for construction of the reservoir perimeter grout curtain.
The project consisted of three grout curtain sections: a 61m double line section constructed along the north side of the planned reservoir using percussion drilling with water as the circulation fluid; a 61m double line section constructed along the south side of the planned reservoir using rotary drilling; and a 25m section constructed adjacent to an overburden cut-off wall demonstration section using rotary drilling. Each grout curtain section consisted of two lines of holes with each line drilled approximately 2.3m from the centreline of the existing overburden cut-off wall. The holes in each section were drilled through steel casings, which were installed through the overburden and seated into rock.
The paper talks in detail about the methods and procedures used during the grouting tests, and summarises the results in both table and figure form. The results of the project demonstrated that curtain grouting could reduce the permeability of the bedrock around the future McCook reservoir thus minimising the infiltration of groundwater and the exfiltration of CSOs. The automated grouting control and data collection system proved to be an invaluable tool to cost effectively track the progress of work.
The grouting test project, according to the authors, identified realistic cost efficiencies for the future construction of the reservoir perimeter grout curtain by determining the optimum curtain depth and appropriate amount of water pressure testing. In addition the test results demonstrated that the water driven, down-hole hammer drilling is the most cost effective drilling method for the deep holes require; treatment of the rock/overburden contact when grouting requires a more intensive effort than is necessary for the deeper bedrock; water pressure testing is not consistently a useful guide to grouting decision making in later sequence holes; and microfine cement grouts were not more effective than the highest viscosity Portland cement grout mix in penetrating the finer discontinuities at the reservoir site.
Scour hole remediation
The final paper included in this session focused on the work involved in repairing a scour hole discovered in the tailrace channel of Arkansas Electric Cooperative Corporation’s (AECC) HS 2 station. ‘Tailrace scour hole remediation at a HS 2 Hydroelectric Generating Station’ by Steve Metcalf of AECC and W B Smith of Hydropower International Services discusses the investigation into the cause of the scouring and examines the development and implementation of a plan to remediate the damage both underneath the power house and in the tailrace channel.
A 102.6MW low head facility located on the Arkansas river at the Wilbur D Mills dam, the project was completed in 1999 and features three bulb turbine units.
The project’s power house was designed and constructed on dense sand as a floating structure using structural slurry wall panels. Significant scouring was found in the sand bottomed tailrace from about 61m to 305m downstream of the power house during soundings taken in early 2000. Model studies performed during the design phase of the project had predicted rip rap plating of the tailrace in this area would not be necessary, due to low velocities, so the original design included rip rap for only about 46m to 61m downstream of the power house. The material which had been lost due to scour was replaced with quarry run stone and the entire tailrace from the end of the original rip rap to a low flow control weir approximately 427m downstream was plated with a minimum of 0.9m.
Soundings in the project tailrace taken during December 2003 indicated that there were areas where additional material may have been lost due to scour. AECC performed a more detailed survey of the area immediately downstream of the power house and verified the existence of a potentially significant scour hole. An underwater diving inspection was scheduled to determine if the scour had progressed, and the results found that the scour had begun to undermine the extended draft tube section. Operation of the facility was suspended until repair work could be completed , and Hydropower International Services (HISINC) was retained to assist in the evaluation and repair.
The cause of the scour was immediately investigated, with detailed mapping of the void area underneath the draft tube extension carried out by contractors Mainstream Divers. It was discovered that the scour hole extended downstream from the power house for approximately 37m, generally along the centreline of the tailrace channel, with depths up to 4.6m below the original grade.
The authors wrote that the cause of the scouring is not completely understood but it is assumed to be the result of turbulence from the power house discharge in combination with an eddy in the tailrace channel. Another concern as a possible cause was the existence of a piping route beneath the power house from the headrace tunnel into the tailrace channel. However, during the dive inspections none of the divers felt any evidence of flow from beneath the draft tube section.
Concurrent with these investigations, AECC and HISINC examined options for repair. It was the general consensus that the void area should be filled before any remediation was performed. Many options for this work was discussed but it was decided to use divers to place sand/cement bags and/or fabric cloth grout bags to sectionalise the void area. Grout or concrete mix could then be pumped into the various sections of the void area using a tremie method of placement.
Repair of the tailrace was also discussed and it was agreed that areas where material had been lost should be brought back to the approximate design grade using rip rap. In the area where the sand was exposed, the lower portion of the hole was filled with a quarry run stone containing a significant quantity of fines to replace the filter fabric. The remainder of the hole was to be filled with a larger gradation stone to provide additional protection. Luhr Brothers was selected as the dredging and rock contractor.
Implementation of the remediation plan began on 4 February 2004 with the support of Mainstream Divers. The paper discusses in detail each stage of the work involved and describes the methods and material used.
Work was completed on 24 February, with soundings of the tailrace taken immediately after the rock was placed to verify that it had been re-graded back to original design grades. Operation of the facility resumed following this inspection.
Soundings of the area were taken a week after operation resumed, with additional inspections after a further two weeks. These were repeated at monthly intervals until the end of 2004. Some minor levelling of the rip rap was detected but no additional scour has been found, demonstrating that the project was successful.
Sharing such experiences with an international audience at Waterpower XIV is a vital way to help ensure repair projects are undertaken effectively in the future. Discovering the lessons learned by those involved in rehabilitation projects will help avoid any pitfalls that could be encountered during such work. The papers mentioned in this article were just a small number of those presented focusing on repair and rehabilitation, demonstrating that the area is one of increasing importance.
Full copies of the papers featured in this article, together with other technical papers presented at Waterpower XIV, are available from the conference organisers, HCI Publications. www.hcipub.com