As part of the FERC relicencing process, hydraulic design changes were necessary at the Salt Springs dam in California, US. Michael C Johnson, James E Pearman and Rod Lubben report on the results of a physical hydraulic model study to find an acceptable fixed cone diversion valve and stationary hood combination for use at the project
Salt Springs dam is in an integral part of the Mokelumne river project, which is owned and operated by Pacific Gas and Electric Company (PG&E). The project consists of 12 dams and four hydroelectric power houses. Salt Springs power house is the uppermost generating facility and is located at the base of Salt Springs dam. The power house includes two turbine units. Unit 2 produces power utilising water from Lower Bear river reservoir and Cole Creek. Unit 1 produces power utilising water from Salt Springs dam.
Salt Springs dam was constructed on the North Fork Mokelumne river to store water and provide head for hydroelectric power generation. The dam is a concrete face rockfill dam (CFRD) and was built in 1931. The crest of the dam is at elevation 1176m above mean sea level (asl) and the low level outlet (LLO) is at elevation 1118.5m asl.
Mokelumne project licence
The Mokelumne project is operated under a Federal Energy Regulatory Commission (FERC) licence. In June 1999, PG&E, the USDA Forest Service (FS), the USDI Fish and Wildlife Service (FWS), the California Department of Fish and Game (CDFG), the USDI Bureau of Land Management (BLM), the California State Water Resources Control Board (SWRCB), East Bay Municipal Utility District (EBMUD), American Whitewater (AW), Friends of the River (FOR), Foothill Conservancy (FC), and Natural Heritage Institute (NHI) agreed to engage in a public, collaborative process with the goal of executing a Settlement Agreement by 30 June 2000, which would resolve all streamflow issues among the parties in support of the FS issuing its Final 4(e) Conditions and FERC issuing a new project licence. The settlement agreement was developed and executed with an effective date of 21 July 2000. The settlement agreement was incorporated in the new licence issued by FERC on 11 October 2001.
Appendix A of the Settlement Agreement identifies a complex set of flow release management and facility requirements, including minimum stream flows, pulse and recreation flows, ramping rates, streamflow and reservoir gauging, flow information, and associated adaptive management provisions.
Prior to the new licence, flow requirements below Salt Springs dam ranged from 0.6m3 per second (m3/sec) to 0.8m3/sec. Requirements of the new licence range from 0.4m3/sec to 63.7m3/sec, although flows above 30m3/sec will be from the spillway and not subject to ramp rates. Thus, controlled flow release from the Salt Springs facility range from 0.4m3/sec to 30m3/sec. Flow releases are subject to ramp rates as shown in Table 1.
llo flow release capability prior to modifications
The dam includes an intake structure located in the reservoir near the dam. Water travels through the intake, through a rock tunnel under the dam and connects to two penstocks. Penstock A feeds unit 1 and has a branch to a 198cm LLO valve. Penstock B feeds unit 1 by-pass valve and has a branch to a 198cm LLO valve. Both outlet valves sit side by side at 1118.5m asl and discharge into the Mokelumne river. They are both motor operated butterfly valves and are about 70 years old. Their intended purpose is to provide emergency dewatering of the reservoir.
Stream flows from the facility and into the river could have been provided from the following sources:
• Flow from an existing 76.2cm Instream Flow Release (IFR) valve.
• Spill from over the power house tailrace canal side-spill wall.
• Release from the two 198cm LLO Valves.
• Spill from the reservoir spillway radial gates.
The IFR valve is located on the power house’s tailrace canal side-spill wall. It is automated and has a controllable capacity up to 0.85m3/sec with the Tiger Creek canal (canal) at normal operating level.
The power house tailrace canal side-spill wall is located upstream of the canal cross gate. The power house discharge and the canal cross gate control the water surface at the spill section. The spill section has a capacity of 7m3/sec at normal canal water surface elevation but cannot control flow well enough to meet new licence ramp rates.
The two 198cm diameter LLO valves were intended for possible emergency dewatering of reservoir and were found inadequate to provide controlled releases. Therefore, they are only operated if emergency conditions warrant.
In order to meet new licence requirements the following modifications were made:
• Two automated 1.5m sluice gates were added to tailrace canal. These can provide up to 8.5m3/sec assuming water is available from Unit 1, Unit 2, or the Unit 1 by-pass valve.
• One 198cm LLO valve was removed and replaced with parallel 61cm and 198cm fixed cone dispersion valves (FCDVs) supplied by the Rodney Hunt Company (RHCO). These can provide all controlled flow release required by the licence. However, it is more desirable to use flow from the tailrace since that water has been used to generate power.
Fixed cone dispersion valve backgound
FCDVs (Figure 2) were first developed and introduced for dam outlet works service in the early 1930’s by C. H. Howell and Howard P. Bunger who were with United States Bureau of Reclamation at the time. The original FCDV developed by Howell and Bunger was acquired by the S. Morgan Smith Company in the mid 1930’s and promptly trademarked as the Howell-Bunger valve and remains as such today. The S. Morgan Smith Company was acquired by the Allis-Chalmers Corporation in the late 1950’s and the Howell-Bunger valve remained with the Allis-Chalmers Corporation until the late 1980’s when RHCO acquired it in the fall of 1990. This represents over seven decades of continuous history of FCDV and Discharge Hood applications covering a multitude of applications and installations.
When the application requires restriction of the discharge spray when the large spray pattern of the Howell-Bunger is objectionable (perhaps to contain overspray moisture to protect nearby equipment, especially electrical equipment or prevent icing on nearby structures, roadways, etc.), a discharge hood is often employed to contain the discharge spray. There are three basic discharge hood configurations utilised for this purpose;
• Hood integrally attached to Howell-Bunger valve sliding outer gate which travels to open/close the valve. This is the traditional Ring Jet valve developed by S. Morgan Smith Company and Allis-Chalmers Corporation and acquired by RHCO. The Ring Jet valve discharge hood is limited to approximately 53m of net operating head and cannot be used on submerged discharge applications. The Ring Jet valve hood does not create any air demand.
• Separately mounted uniform diameter stationary discharge hood often configured as a steel or stainless steel liner contained in a concrete discharge tunnel. These discharge hoods are required to be an absolute minimum diameter of 2-1/2 times the nominal valve diameter and preferably 2-3/4 times the nominal valve diameter and extend back a minimum of 30cm behind the end of the Howell-Bunger valve. These discharge hoods are typically subject to large amounts of ‘back splash’ and typically result in large air demands requiring large amounts of ‘free air’ typically in the form of large vents behind the discharge hood. These discharge hoods should not be used on submerged discharge service.
• Separately mounted constrained (tapered conical diameter) stationary discharge hood. These discharge hoods are separately mounted, but typically not encased in concrete. The Howell-Bunger valve is inserted into this discharge hood and this discharge hood begins with a diameter slightly larger than maximum gate outside diameter and opens up to a uniform diameter required to be an absolute minimum diameter of 2-1/2 times the nominal valve diameter and preferably 2-3/4 times the nominal valve diameter. These discharge hoods more closely resemble the traditional Ring Jet valve hood, however, they are stationary and do not travel with the sliding gate. Because the initial expanding Howell-Bunger valve discharge jet is being constrained by these discharge hoods, there is very little air demand typically not requiring any additional ‘free air’ venting as required by the separately mounted uniform diameter stationary discharge hood. Again, these discharge hoods should not be used on submerged discharge service.
LLO modifications were needed to meet several criteria including:
• A valve or valves would be needed to accurately control from about 0.85m3/sec to 30m3/sec.
• The LLO capacity could not be reduced.
• Facility noise must be kept to a minimum due to environmental concerns.
• Cost, safety and maintenance.
In order to meet these criteria, FCDVs were selected. However, using FCDVs leads to further constraints:
• Piping immediately upstream of the valve must be straight for at least five pipe diameters.
• The valve discharge is into an existing plunge pool downstream from the dam. The discharge could not hit the pond’s banks due to erosion and turbidity.
• Valve geometry could not interfere with release from the remaining 198cm butterfly valve.
• Valve discharge must be free and not at all submerged by the pond elevation.
Original valve specifications called for a detached stainless steel hood to direct the discharge straight into the pool and away from the banks. The hood size was specified at twice the valve diameter.
RHCO was first contacted by PG&E regarding the LLO discharge valve requirements in the fall of 2003, long after many of the critical preliminary design issues had already been addressed and committed to.
These decisions included recommended valve types, required valve sizes and discharge hood configurations. Many of these decisions were arrived at during the difficult relicencing and design review process with the multiple reviewing agencies previously mentioned. These decisions placed some formidable restraints on the discharge valve and discharge hood requirements which RHCO was challenged to address in the design, manufacture, supply and testing of the LLO discharge valves. During the bidding process, it was discovered that hoods are typically 2.5 times valve diameter.
During the extensive design review process of the LLO discharge valve requirements, PG&E ultimately settled on the Howell-Bunger type FCDV complete with a discharge hood. Because the net operating heads were near 91m, a separately mounted discharge hood was required and the separately
mounted constrained (tapered conical diameter) stationary discharge hood was selected as best suited for the existing site conditions and restraints for the application.
However, existing site conditions limited the maximum discharge hood diameter to approximately twice the nominal valve diameter. This would severely restrict the ability of the discharge hood to serve the Howell-Bunger discharge requirements without substantial adverse effects.
Coupled with this, to the best of RHCO’s knowledge, the separately mounted constrained (tapered conical diameter) stationary discharge hood had never been utilised anywhere near this relatively large 198cm diameter valve before. Since no hood smaller that 2.5 times had been constructed, model testing was needed to see what hydraulic impacts might arise if hood diameter was lessened.
The objectives of the model study were:
• The required discharge hood configuration was based on twice the nominal valve diameter requirement including required tapered conical section angle and overall length of hood. Site conditions restricted both maximum diameter and overall length of valve/hood combination.
• The required valve insertion into hood to maximise valve opening capacity as well as minimising effects of back splash at large openings while minimising overspray at small openings.
• The free air demands, thereby determining minimum inlet opening diameter to satisfy those free air demands.
The model was constructed and operated in the Utah Water Research Laboratory in Logan, Utah. Water was supplied from a combination of a 100hp and 200hp pump or under gravity flow from First dam located on the Logan river.
Flow for the model was measured using a calibrated venturi meter installed upstream from the model. Flow and pressure to the model was controlled using a pump by-pass valve and an in-line butterfly valve upstream from the model. The piezometric pressure was measured one diameter upstream from the inlet of the FCDV with a precision pressure gauge.
Piping that was used to supply flow to the model included approximately 3.7m of 8-inch pipe upstream from the FCDV. There were 20 diameters of straight piping upstream from the venturi meter and six diameters of straight piping downstream from the flow meter. Figure 1 shows a drawing of the test setup used for the testing.
Figure 2 shows the model FCDV that was used for this study free discharging to atmosphere. It has a 90-degree cone angle (total angle) and a 20cm cone diameter at the exit. The fixed cone is supported with six struts welded to the cone and the valve body. The length ratio, or scale of the model, was defined by the ratio of the diameter of the prototype valve to the model valve. For this study, the prototype conditions prevailing at the 152.4cm valve which is to be located at another PG&E site (Lower Bear) was used as the prototype for setting flows and pressures. At the time of the field testing, the 152.4cm valve was not installed, so the model data was scaled to the 183cm valve for comparison. In some cases, the model hood/valve was tested far beyond the requirements of the Salt Springs site in order to assess the performance under extreme
conditions and to broaden the potential operating range.
In order to facilitate variations that could be investigated, the hoods were mounted to a moveable carriage that was placed on rails over a large channel in the laboratory. This installation made it possible to move the hood relative to the valve while the valve was operating. Dynamically changing the location of the valve relative to the hood while the valve was operating aided in assessing the performance of the FCV/hood combination. The developmental phase of the hood design resulted in ten different hoods being evaluated. Personnel from RHCO and Utah State University agreed upon the various designs tested in the laboratory. The model study demonstrated the difficulty in getting the correct hood/valve configuration to contain the back splash. Figure 3 shows the hood that was first tested and the hood that was finally considered acceptable. Ultimately the ratio of the valve to hood diameter was approximately 1:2.2.
Water was supplied from a combination of a 100hp and a 200hp pump connected in parallel. Flow was measured from a calibrated 30cm venturi flow meter. The flow, pressure upstream from the valve, decibel level (dBA), and jet trajectory were measured for each condition evaluated. The total head upstream from the valve was computed by adding the velocity head to the pressure head (Ht = V2/2g +P/ symbol here) where V is the average velocity, g is gravity, P is the pressure upstream from the valve and SYMBOL HERE is the unit weight.
In order to hit specific conditions of interest, it was necessary to open the gate on the FCV and simultaneously throttle the flow with an upstream valve until the proper flow and pressure were established at the valve. The process was iterative and the conditions established closely correspond to those expected to prevail at the 198cm prototype FCV installation. Due to laboratory flow and pressure limitations, conditions corresponding to the smaller valve sizes could not be replicated. Table 2 shows the conditions that are expected to occur at the Salt Springs site.
Model study results
For this study, the primary focus was that of developing a hood that would operate over a wide range of flows without experiencing significant back splash. Given the reduced ratio of hood diameter to valve diameter, this proved a difficult task. Each of the hoods evaluated contained the conical spray; however the back splash performance of each hood varied between the various hoods investigated. To demonstrate the problem back splash can produce, Figure 4 shows hood configuration #1 operating at 71.3m3/sec with a total head of 50m. Figure 5 shows the final hood configuration (#9) operating at 39m3/sec with a total head of 82.5m. Tables 3 and 4 summarise the data for hoods numbered 1 and 9 respectively.
During the course of the study it became evident that reaching the absolute maximum flow without excessive back splash was probably not possible with the hood diameter constraints dictated by the licence. Therefore, the acceptance criteria and diameter constraints were relaxed and the hoods were simply required to pass the normal maximum flow without excessive back splash.
On Wednesday 12 January 2005, PG&E performed full flow field testing of the RHCO 198cm Howell-Bunger valve and constrained (tapered conical diameter) stationary discharge hood. Because of ongoing work on the dam, the reservoir elevation was approximately 36.6m above the centerline of the valve, significantly lower than the normal pool elevation. The objective of the field verification was to assess how well the valve/hood combination passed the flow and to determine to what extent the valve could be opened before back splash became excessive. The results of the testing not only met, but exceeded PG&E’s specification requirements significantly with no adverse operating effects. Furthermore, the valve/hood exceeded the expectations identified in the model study.
Figure 6 shows the 198cm valve/hood in operation at an opening of 53cm with 41.6m3/sec passing through the valve and at approximately 36.6m of head. Compared to the results given in Table 4 where, at an opening of 37.8cm, back splash started at 39m3/sec at a head of 82.5m, the valve operated exceptionally well. The noise levels at the inlet of the hood and approximately 61m away from the hood on the banks of the stilling pool were measured to be 104 dbA and 82 dbA respectively.
Experience in the laboratory has demonstrated that the large valve openings at relatively low total energies result in the most difficult operating conditions for hooded FCDVs. This is due to the velocity being considerably reduced at the exit of the valve to pass the same flow were a higher energy (head) available, resulting in smaller valve openings for the same flow. Therefore, the conditions prevailing at the time of the test were considered to be a worst case scenario with the valve required to pass the maximum flow at the minimum head. The authors are confident that with higher heads the performance of the valve/hood
combination will only improve, since the valve will not be required to open as far to pass the same quality of flow. Figure 7 shows a drawing of the final valve/hood combination.
This study was completed to find a design of a hood that would work in conjunction with a FCDV subject to formidable design constraints. Generally hoods have a diameter of approximately 2.5 times the valve diameter and have relatively few problems with back splash. The hoods of this study were required by agreement by multiple collaborative organisations to be nominally twice the diameter of the valve, in order to meet the requirements to relicence the hydro project at Salt Springs on the Mokelumne river in California. RHCO was awarded the contract to supply the valves and hoods and a model study was completed at the Utah Water Research Laboratory. Various hood configurations were investigated and a hood design was found that would satisfactorily operate under the conditions expected without excessive back splash.
The final hood configuration evaluated proved to minimise back splash over all the other hoods tested (ten in total). Additionally, the hood safely and adequately contained and turned the conical spray in a way that directed the spray away from structures and placed it where desired. The model valve was evaluated at several total heads and stroked over ranges of flows representative of prototype conditions. Based on the data collected and observations made during the witness visit, personnel from the RHCO and PG&E accepted the performance of hood.
Upon completion of installation of the valves and hoods at the Salt Springs facility, a field test was conducted to verify the performance of the FCDVs coupled with the hood geometry determined from the model study. The field testing was completed and demonstrated that the 198cm valve/hood combination exceeded operational expectations required by the relicence.
Michael C Johnson, Research Scientist, Utah State University Research Foundation (Tel +1 435 797 3176; Janes E Pearman, Senior Valve Engineer, Rodney Hunt Company (Tel: +1 978 544 7204); Rod Lubben, Consulting Engineer (Tel: +1 925 254 5507)
This paper is printed with kind permission of the Association of State Dam Safety Officials. For further information, please visit www.damsafety.org.