5. OPERATION OF GATED SPILLWAYS

Reliability of gate operation is mandatory to assure dam safety when passing extreme floods. As noted earlier in this text, a gated spillway is inherently less reliable than an ungated spillway. Thus, it is necessary to take special note of the reliability requirement when designing gated spillways and the equipment used to operate the gates. In general, there are two major reasons for which a gated spillway may fail to meet its necessary operating requirements.

-The operator opens the gates too late or fails to open the gates at all, even though the gates are operable.
.
-The operation of the gates fails because of structural failure of the gate itself or failure of one or more of the various components of the operational equipment.

5.1. FAILURE OF THE OPERATOR TO OPEN THE GATES

In general, failure of the operator to open the gates when necessary even though the gates are operable is a result of one or more of the following:

-The operator is not present at the control point when needed. Severe storms that produce extreme floods may also make access roads impassable at exactly the time when access to the dam and control facilities is most critical. Advance planning regarding access to the spillway site is critical to insure reliable operation of gated spillways during passage of extreme floods. This aspect is particularly important since many dams are now operated remotely and frequently no operator present at the dam when a flood begins.
.
-The operator depends upon instructions from a higher authority and communication has been cut off. Storms that produce extreme floods often do devastating damage to communication facilities and to the data collection and transmission facilities of a real-time forecasting system. Unfortunately, it is at exactly this time that communications to the dam operator are most important (R.17).

-The operator has not received sufficient training in order to understand what is required of him when an extreme flood is entering the reservoir, or may not have the necessary tools which can help him make a decision about when or if to start opening gates. Q. Shaw and W. Hakin point out this item can seriously reduce reliability in a country where the national economy has suffered or internal unrest has occurred (R.18). Total reliance on black-box programs for safe operation of spillway gates without backup training of the dam operator could jeopardize the safety of the dam if no back-up system is in place for the operator to use if the black-box program system fails (R.17).

5.2. FAILURE OF THE GATES TO OPERATE

There are many reasons why gates may not be operable when their operation is most critical. This concern was addressed by more responses than any other item (R.1, 3, . 9, 10, 11, 12, 14, 18, 20, 26, 33, 34, 35, 36, 37, 40, 43)This large response is clearly indicative that the problem of gate reliability is one of concern to all those who design, operate, or own dams with gated spillways. There are two broad reasons for which spillway gates are inoperable when needed (R.20, R.33):

-Design Deficiencies

-Damage Induced by Aging

5.2.1. Design deficiencies

In most cases design deficiencies are due to failure to recognize loading possibilities for which the gates and their structures should have been designed.

Seismic Loads

Lack of consideration of possible seismic loading may be the most subtle of these deficiencies. Attention is drawn to seismic design considerations in responses R.3 and (R.32). The need for consideration of seismic loading was pointed out by W. Daniell and C. Taylor (R.9) who noted that failure to analyze seismic loading is most likely to occur for dams in zones of low seismicity. First consideration of seismic loadings frequently occurs when the design of gated spillways is revisited during a process of rehabilitation of the spillway (R.3). For older dams seismic loads may have been considered to be unimportant at the time of their original design, but more recent seismic studies have shown that the probability of severe earthquakes is greater than first thought. Frequently the rehabilitation is the result of mandates by regulatory authorities or the development of more stringent standards for design.
Earthquake damage to gates is not unusual due to the large added-mass forces that can act on the gates during an earthquake. In 1990 a magnitude 7.3 earthquake in Iran caused serious damage to the radial gates of the intermediate spillway of the 106-meter high Sefidrud buttress dam but did not damage the bottom outlet gates or the irrigation-outlet gates (R.37). The epicenter of the earthquake was only 10 km from the dam. Accelerations of 6.0 9 were estimated to have been experienced which was significantly greater than that used in the design. The dynamic load exerted on the radial gates caused trunnion arms to buckle. The spillway gates were damaged even though the water level in the reservoir was 6 meters below maximum normallevel.

Vibrations

The second most common case of design errors is failure to consider the possibility of severe vibrations of a gate (R.1). Although guidelines have been developed which, if followed, will lead to a trouble-free design for the most common gate designs, there will always be special cases in which more detailed considerations should be made. Those special cases usually involve unusual approach flow conditions which may not seem important to the average designer. Radial gates may be subject to vibration at small gate openings due to instability that occurs as a result of seal problems; for that reason R. Berridge and M. MacDonald recommend that the angle between the upstream gate face and the spillway be made as large as possible (R.40).

It should be pointed out there is some speculation that vibrations played an part in the 1995 failure of the radial spillway gate at Foisom Dam in California [8]. If so the vibrations would have occurred with agate opening of 0.73 meters and a head on the gate of 12.2 meters. Stiffening of the remaining gates was increased significantly and the replacement gate was designed to be much stiffer than the failed gate. Field tests conducted in 1998 showed no tendency for vibration of the stiffened gates.

Design Details

Another common error in the design of gates is in failing to recognize opportunities for corrosion in the design of the gates or to detect serious corrosion (R.20, 21). It is of particular importance to recognize that ail parts of the gate must permit inspection and that care should be taken not build in opportunities for water to accumulate and accelerate corrosion. Corrosion is one of the chief problems encountered as the project ages (R.29. 34, 35). It is particularly important for gates which will operate in cold climates not to have any pockets where water can accumulate (R.20). Often times inspection fails to detect corrosion as quickly as desirable. When the radial gate at Foisom dam failed in 1995, inspection of the failed gate showed that several of the rivet heads placed during the initial fabrication of the gate had corroded to the point where they no longer had any structural strength [10]. In the period immediately after the failure and before the more detailed forensic studies were made, this structural corrosion was thought to be the root cause of the failure. The fact that this structural corrosion problem went undetected in a national organization with an established maintenance program, is testimony to the fact that special attention needs to be given to inspection and maintenance if the reliability of gated spillway operation is to be assured.

Fabrication and/or construction errors apparently were the cause of failure of a large radial gate (15m by 13.5m) on Singur Dam in India. The lapse was thought to be due to a lack of inspection at all levels (R.33). Obviously the choice of qualified contractors and sub contractors is vital in the fabrication and construction of all projects. The need for nothing less than complete quality in the construction of large dams was addressed extensively in F.M. Budweg's "General Report for Question 75" at the Nineteenth Congress [11]. It should be of considerable concern to the dam building profession that, at least in the United States, there has been a growing tendency to make inspection a responsibility of the contractor. Although this practice may offer administrative advantages, it places the guarantee of quality control in the hands of the contractor who will almost always have conflicts of interest because of his financial involvement. The practice is not conducive to good quality control

.
5.2.2. Damage induced by aging

As was pointed out in many of the responses, no single element is more important in insuring reliable operation of gated spillways than quality maintenance performed on a regular basis. Several of the responses discussed the fact that problems affecting operation of gated spillways often occur as a result of aging and should be expected. The importance of maintenance was discussed in several responses (R.20, 29, 35, 40, 43). These responses listed items of maintenance that were particularly important and described problems that have arisen with gate operation as a result of aging. Regular detailed inspections of the gated spillway after construction are needed in order that aging problems can be detected and corrected in order to assure reliable operation.

Problems due to aging

Many problems which occur as result of aging occur as a result of corrosion. These problems include deterioration of structural steel from which the gates have been fabricated and corrosion of moving parts. R. Berridge and M. MacDonald describe a failure due to corrosion of structural members of the 18.3m long vertical lift gates on a barrage across the Indus River (R.40). Corrosion was so severe that 56 of the original 66 gates on the barrage were replaced. It was also necessary to replace 55 of the gates on the canal headworks.

Both cables and chains used in gate hoists are subject to corrosion as a result of being exposed to the elements of weather. Sometimes the deterioration of hoisting cables is not detectable from surface inspection and is only discovered during operation when reliable operation is of the most concern. Such problems were discovered by the U.S. Army Corps of Engineers during flood operations in the spring of 1993. Hoisting cables broke at several dams in the Mississippi River Basin when the gates were being operated under load. Inspection of the failed cables showed that damaging corrosion had attacked the interior strands of the wire rope reducing the strength of the cable to the point where it was inadequate for the hoisting loads encountered. B. Sagar particularly recommends that stainless steel wire rope be used for cable hoists (R.34). However, corrosion can occur rather rapidly on stainless steel if oxygen-depletion cells are formed. Since the corrosion of wire rope occurs on the center strands of hoisting cables, it is unlikely that the use of stainless steel wire rope will eliminate the need for regular inspection of hoisting cables.

Accidents involving hoists are not unusual and where chains are used they require special inspection to detect corrosion and points of potential failure. A chain on the hoist for the spillway at Picote Dam in Portugal broke in February 1966 and dropped the gate. The dynamic load caused the gate to fail and it was then swept away by the flow (R.11). The chains on the gate hoists at Foisom Dam were found to be corroded and stiff after the failure of the radial spillway gate in 1995 [10]. Although none of the chains operating the gates had failed, rotation around the chain pins was restricted by corrosion and caused jerking of the gate when it was raised or lowered. The jerking was originally thought to have been a possible triggering mechanism for gate vibrations which might have contributed to the failure of the gate.

Corrosion of trunnion pins is another concern that was brought to the dam engineering community by the failure of the radial gate at Foisom Dam in 1995. In that case lubrication did not reach the entire surface area of the pin and coefficients of friction ranging from 0.270 to 0.672 were computed from tests performed on two of the remaining gates and on trunnions of the failed gate [12]. A finite element structural analysis showed that a coefficient of friction of 0.25 would have been sufficient to cause buckling of a critical member in the gate. Recommendations have been made that spherical bail bearings be used on the trunnions of radial gates (RAO). Such spherical trunnion bearings would be expensive and would certainly require regular maintenance. However, they would be an aid in moving gates where a small misalignment of the trunnion pins had occurred as a result of movement of one or both of the piers.

A unique problem involving the 1994 failure of the operating rod for a hydraulic cylinder occurred on one of the large radial gates for Itaipu Dam (R.43). The rod had fractured during lowering of the gate when upstream bulkheads were in place and the gate was not under load. Close inspection showed that almost imperceptible corrosion of the small end of this rod had resulted in a condition which produced small cracks in the cylinder. The cracks could only be detected with dye penetrant. A finite-element analysis of the rod, proved that the strength of the rod had been significantly reduced. However, static analysis indicated that the hoisting load was not sufficient to cause the rad to fail even with the minute cracks. Field tests done on the gate showed that the gate did not move smoothly when it was lowered or raised when not subject to load. Apparently the rubbing side seals caused vibrations and racking of the gate which in turn generated a dynamic overload that was
sufficient to stress the corroded rod to the point of fracture. Since 22 of the cylinder rods on the remaining gates were similarly corroded, special rehabilitation and maintenance measures were instituted in an attempt to remove the danger to reliable operation. The 22 rods were machined and polished to remove corrosion and cracks. Future measures will include coating the rods with a corrosion protection chemical on a regular basis. Restrictions include not moving the gates when bulkheads are in place without spraying a mixture of water and liquid vaseline on the seals. Apparently, a very similar failure occurred one of the spillway gates at Tucurui Dam in 1995. The engineers involved in the rehabilitation of the hydraulic cylinders at Itaipu Dam recommend that hydraulic cylinder rads should have a ceramic coating in order to eliminate the possibility of damaging corrosion (R.43).

Unusual loading

Spillway gates are apt to incur unusual loading when operated to pass extreme floods. The experience of Hydro Quebec during the flood on the Saguenay River during July 1996 brought to light an array of operational problems that can occur when the largest flood of record must be passed. The Saguenay flood surpassed the 1000-year exceedance value and, when the actual 1996 flood flows were included in the peak-flow data base, estimates of the 10,000-year flood were tripled [13]. Hydro Quebec found that many of their problems with gate operation were the result of inadequate maintenance (R.26). Problems encountered with gate operation included:

-Hoists broke or had insufficient power to operate gates. Enormous amounts of debris lodged against the gates and stop logs and caused loading well above that for which the hoists had been designed. In many cases the gates were overtopped.

-Stop logs jammed in their slots and could not be removed.

-Power failed. In several cases the failure was due to saturation of power cables that were submerged by the flood.

-Access to the gates was cut off and in several cases key operational people were not available or could not get to the dam site.

Legislation

The Saguenay experience has resulted in recommendations for dam safety legislation being made to the provincial government of Quebec, Canada. Prior to the Saguenay flood, the Province had no dam safety laws. Legislation recommended by the Canadian Dam Safety Association indudes:

-Spillway capacity for all dams must be reviewed.

-All wooden stop logs must be replaced with metal or inflatable gates.

-Alternative power supplies must be provided for all gated spillways.

-Special consideration must be given to the effects of floating de bris and sunken logs.

-Because of the need to pass large quantities of debris during an extreme flood, a minimum gate width of 4 meters is to be used.

-Access to gate structures must be available at all times.

Similar guidelines for the improvement of dam safety in France have been developed by the French National Committee of ICOLD. Their recommendations include (R.29):

-Backup power must be properly maintained.

-Hand cranks for the raising of gates should be eliminated since they are too slow and require more effort than one operator can exert.

-Primary and backup power cables leading to hoist motors should be routed along separate paths.

-Gate controls should be available in at least two locations.

-Instruments for data acquisition should be established and carefully maintained.
They should be inspected, tested, and maintained on a regular basis.

-Inspection must be done on a regular basis and should include observations for local subsidence, erosion of moving gate parts by friction, erosion due to cavitation and energy dissipation, and obstruction of gates by bed load.

-Both gate slots and the gates themselves should be inspected to detect and remove any debris lodged in them which could prevent operation of the gate and accelerate corrosion of structural members or operating equipment.

-Evidence of floating debris.

6. INSPECTION AND MAINTENANCE

Several points in the text of this General Report have mentioned inspection and maintenance of spillways as a vital part of assuring that spillways will operate reliably and safely. Inspections are required to insure that the gates and the spillway are in good operating condition and to detect any conditions that may be or may become a threat to reliable operation. Different levels of inspections should be held regularly (at least annually) by operators and engineers familiar with the operation and the operating history of the project. In addition, it is wise to have outside experts inspect the spillway and appurtenant equipment at least every five years. The outside experts provide experience from other projects and may observe items that the owner's operators and engineers may not be aware of. The inspection should always review the maintenance program as well as the spillway and related equipment themselves.

Maintenance of spillway gates and their operating equipment is of critical importance. Recent incidents involving corrosion of trunnion pins on radial gates have shown the possibility that maintenance may have been deficient on such projects. In the United States, the Federal Energy Regulatory Commission, which regulates all hydropower dams, requires that all spillway gates be periodically operated under load to assure that they can be operated when necessary. It is of course quite important that the gates be able to operate under load. However, it may be equally important that the gates be periodically raised fully to insure that lubricant applied to the trunnion pins covers the entire bearing surface (R.34, 20).

It is frequently inconvenient and expensive to test gates through their full range particularly if devices for use of stop logs have not been installed on the spillway. However, the critical importance of assuring that operation of the gates must be reliable is overwhelming and measures should be devised to do the required testing for all high-hazard dams.

After the failure of the Foisom spillway gate in 1995, the U.S. Army Corps of Engineers devised a laser device that can be used to measure the deflection of radial gates when they are opened under load. Using this deflection measurement they can estimate stresses on structural members of gates which are tested. The device is easily attached to the trunnion of a radial gate and provides the means for a quick assessment of overall conditions affecting the operation of the gate. The Foisom accident also triggered the Division of Safety of Dams in California to call for inspection of all spillway gates on California dams. That inspection is still under way.

7. CONCLUSIONS

1. The use of uncontrolled spillways will always be the safest design. However, even it will not adequately protect a dam if the design flood is not carefully chosen.
2. Gated spillways because of their flexibility in operation and economy in fabrication and construction will continue to be widely used.
3. Radial gates, because of wide experience ln their use, and their economy of fabrication, will continue to be the gate of choice in spillway design.
4. The importance of careful regular inspections and maintenance of spillways, gates, and operating equipment cannot be overemphasized.
5. Regular maintenance is absolutely essential to assure reliability in the operation of spillways.
6. Periodic inspection of spillways and related equipment by outside experts is highly desirable. Intervals between these inspections should not exceed five years.

8. REFERENCES
[1 ]
Bulletin 99, Dam Failures Statistical Analysis, ICOLD 1995.
[2]
Bulletin 82, Selection of Design Flood, Current Methods, ICOLD 1992.
[3]

Bulletin 49a, Operation of Hydraulic Structures of Dams, ICOLD 1986.
[4]

Bulletin 58, Spillways for Dams, ICOLD 1987.
[5]
Bulletin 102, Vibrations of Hydraulic Equipment for Dams, ICOLD 1996.

[6]
Bulletin 105, Dams and Related Structures in Cold Climate, ICOLD 1996.
[7]
"International Symposium on Dams and Extreme Floods, Vol. 1 Design, Vol. \1 Operation, Vol. III Discussion," Spanish National Committee, ICOLD, Granada, Spain 1992.
[8]
"Spillway Radial Gate Failure at Foisom Dam, California" Proceedings of the Nineteenth Annual USCOLD Lecture Series. Atlanta. Georgia. Mav 17-21. 1999, R.V. Todd.
[9]
Pugh, C.A. "Hydraulic Model studies of Fuse Plug Embankments," REC-ERC­85-7. U.S. Bureau of Reclamation, Denver, Colorado, Dec. 1985.
[10] "Forensic Report on Spillway Gate 3 failure Folsom," U.S. Bureau of Reclamation, Mid-Pacifie Region, Sacramento, California, November 18, 1996.
[11] F.M. Budweg, General Report. Question 75, Vol. IV, Transactions of the
Nineteenth Congress of ICOLD, 1997.
[12] R.V. Todd, "Trunnion Friction Report Analyses of Test Data From Spillway Gates No. 1 and 4," U.S. Department of the Interior, Bureau of Reclamation, Denver, Colorado, January 1996.
[13] "Lesson of Saguenay," Water Power and Dam Construction. Vol. 50, No. 5,
May 1998.

Previous <1 2 3>