CS-020: Savannah River Remediation

CHALLENGE

The challenge was to simulate the Process Vessel Vent (PVV) system which removes toxic gasses that are released from the processing of nuclear waste and vents the gas out to atmosphere.  Therefore it is necessary to develop an engineering simulator for the PVV system.

BENEFITS

• Develop an engineering simulator for the PVV system and benchmark it against measured data.

• The simulator had a customized Human Machine Interface (HMI).

• Potential locations of leakages and blockages on the system were identified.

• Capability of using external calculations/correlations that are user-defined.

SOLUTION

The simulator developed proved to be a valuable tool to predict the location of leakages and blockages in the PVV system. Flownex® can now be used to predict the system conditions at different inputs and to evaluate the overall performance of the PVV system. 

INTRODUCTION

Savannah River Remediation manages the Liquid Waste Operations at the US Department of Energy’s Savannah River Site. These include waste solidification, during which by-products of nuclear weapons production are vitrified (glassified) and made ready for permanent disposal. In preparing the nuclear waste for vitrification, its slurry is chemically adjusted, stripped of mercury, and concentrated through evaporation. It is then compacted through the addition of glass frit. During these preparation processes, toxic and/or radioactive gasses are released. The PVV system maintains vacuum in process equipment in order to limit and mitigate leakage of highly contaminated vapors into the atmosphere.

SYSTEM DESCRIPTION

The PVV includes the blowers, heater, filter, valves, scrubbers, condensers, piping and other equipment necessary for (1) maintaining negative pressures in the process vessels, (2) filtering particulates and (3) discharging the gas flows.

Nuclear waste slurry is chemically adjusted and concentrated in the SRAT and SME prior to feeding the melter. The process, known as vitrification, produces borosilicate glass canisters. In the SRAT and SME processes toxic gasses, which include hydrogen, carbon dioxide, nitrous oxide as well as other contaminated particles, are generated as a by-product of chemical reactions. The tanks are thus purged with air to keep the levels of gas low. This gas flows through condensers, where the vapor within the gas is condensed and separated. One of the desired reactions is to reduce and strip mercury from the sludge. Ideally the mercury exits the condensers with the condensate while the gas flows to ammonia scrubbers. At the scrubbers, the gas flows through packing material whilst being sprayed with water to remove contaminated particles.  The gases exiting the SRAT and SME scrubbers are sent to another condenser, the FAVC. The gas is cooled further to 10 °C and filtered to remove finer particles. Blowers drive the circulation of the gas to the exhaust, where it is released into the atmosphere.

Due to the radioactive and toxic nature of nuclear waste processing, the PVV system and the equipment it serves are located in a secure and sealed off section within the DWPF building. Thus, any maintenance or work on the system has to be performed remotely and knowing the location of leaks and blockages can save valuable downtime.

OBJECTIVE OF SIMULATION

• Develop an engineering simulator of the PVV system.

• Benchmark the simulator against measured data, and

• Locate possible areas on the system where there may be leakages and/or blockages, and any other impediments to the flow through the system.

FLOWNEX MODEL

The Flownex® model of the PVV system is pictured below. One way to improve process efficiency and speed up the SRR process is to decrease concentration times by increasing boiling rates in the SRAT and SME. However, the SRAT and SME must be kept under vacuum, so boiling rates are limited by pressure restrictions in the PVV. At increased boiling rates, pressure drops across the condenser and scrubber are increased. This is caused by the increased vapor that is boiled off from the tanks at increased boiling rates. To fully account for this phenomenon, a mixture of incondensable air and vapor was used as the working fluid in the network from the boiling tanks to the exit of the condensers.

For the rest of the network, however, the gas composition is similar to that of dry air. Air was used as a working fluid. The yellow boxes in the picture are used for input data (Plant measurements can be added here), and the green boxes display the results. The computer icons are scripts that were used to perform external calculations, which were integrated into the model. These included the pressure drop correlation across the scrubbers and display of results in custom units that are familiar to end-users.

To many users, the engineering model pictured above would look daunting at first glance. So to make the model user friendly and to only display the required results, an HMI was developed. The HMI shows the different sub-sections of the PVV system schematically, as well as a summary of system conditions. Users can fully control the model, and investigate phenomena in the system without working on the engineering model. A few pages of the HMI are shown below:

RESULTS

The flow conditions and state of PVV system equipment were benchmarked against measured data. Particles that are boiled off from the boiling tanks were suspected of cooling and settling along the inside of piping and equipment. SRR engineers had found mercury and other deposits inside the system pipes. Another observation was that when the blower speed was increased, there was no corresponding decrease in pressure at one of the boiling tanks, the Sludge Receipt and Adjustment Tank (SRAT). Thus, a leak into the system was suspected along the SRAT’s flow path.

Through the use of the designer feature in Flownex, a number of leakages and blockages were introduced to the model. The results of the simulator compared very closely to the measured data from several tests. Areas that were found to have significant blockages were the scrubbers and the condensers. Thus, there may be significant deposits of particles on the inside of the condenser tubes and on the scrubber packing material. Another possibility, which cannot be confirmed due to limited data, is that the scrubbers are flooded, which leads to an increase in the pressure drop.

A sizable leakage into the system was introduced at the SRAT scrubber inlet. With the inclusion of the leakage, the pressure drop across the scrubber was matched with measured data from several tests. The engineering simulator was able to replicate the observation that an increase in the blower’s speed did not lead to a decrease in the SRAT tank’s pressure, but rather resulted in an increase in the leakage introduced. The simulator also showed an increase in the condenser pressure drop when the rate of boiling at the tank was increased.

CONCLUSION

As the engineering simulator was successfully benchmarked against measured data, it could now be used to predict the system conditions at different inputs. SRR is now able to evaluate the performance of the different components of the system on the engineering simulator. Suggestions to improve the overall efficiency of the PVV system may also be tested on the simulator.

Flownex® has proved to be an economical and time saving tool in developing an accurate simulator of this complex and integrated system.

DOWNLOAD PDF