| dc.description.abstract | The process plant at Kollsnes uses flares for emergencies and routine. Routine flaring is performed during maintenance – amongst other reasons - because the pressure is too low to inject it into the process. Equinor has set a specific climate ambition to reach zero routine flaring by 2030. Therefore, a way of minimizing flaring is to reinject the gas into the process with the aid of compressors. Two studies have already been conducted on reinjecting the gas into the process using existing flash gas compressors. The studies were conducted in 1997 and 1999. Based on the financial aspect, it was decided that the studies should not continue to the design phase before eventual stricter regulation or higher emissions fees to air were implemented. Since 1999 the regulations have become stricter, emissions fees higher, and gas prices have multiplied. A new flash gas compressor has also been installed with a larger and unused capacity. The flash gas compressor is referred to as flash gas compressor C.
This thesis investigates the possibility of reinjecting the gas into the process using flash gas compressor C. The objective of this thesis is the following. First, propose a design based on recycling time, gas recovery, and minimum modifications to the current system. Second, to create a case study on the proposed design in UniSim Design and evaluate it using dynamic modeling. The design shall be assessed on recycling time, amount of gas recovered (including emission and cost reduction), and feasibility of the proposed solution. Third, a final design will be presented based on the case study, including a simple control philosophy and safety evaluation.
The research was limited to only investigating two sections to be recycled for maintenance – the sections where the DPC (Dew point control) and Export compressor trains. A proposed solution was created and evaluated with case studies in UniSim Design by dynamic simulation. The case studies were performed in an already built simulation environment of the process plant at Kollsnes. By investigating the simulation environment, errors were detected in the volume implementation and in the compressor control. The volumes were corrected, and it was decided at first to run the compressor speed at 100 % performance with the PCVs and the recycling valves controlling the inlet pressures. After that, a proposed solution was created and evaluated with three case studies in UniSim Design. The first study found limitations in the inlet temperature. Therefore, a second case study was conducted with a heater before inlet step one. Furthermore, the compressor control was fixed, and a third case study was performed with a correctly implemented control. After that, a final – modified - design was presented based on the feasibility evaluation.
The results of this thesis indicate the following. First, the recycling time from shut-in pressure to seven barg is 39 min for the Export and 103 min for the DPC. To give it a reference, for the same amount of gas, the maintenance or HP flare would use approximately 6 and 15 min for the Export and DPC, respectively. Second, the amount of gas saved over one year with this recycling solution is 225 kSm^3. This results in a possible emissions savings of 608 tonnes of CO2e and 225 kg NOx. Thereby a combined cost-saving and earning of 2.4 MNOK per year. By comparison, the total emission of CO2e from Kollsnes is 76.3 ktonne, and the total NOx emissions from Kollsnes are 30.4 tonnes. That is approximately 0.8 % of the total CO2e and NOx emissions per year. Third, concerning the feasibility of the proposed solution. The solution needs MEG injection before the pressure letdown and an extra heat exchanger with a minimum capacity of 400 kW at the step one inlet. However, there is no need to upgrade the inlet separators for the recycling operation during normal conditions.
The simulation model and the changes in the control seem to give a valid and working model that can be useful for future simulations.
The primary uncertainties in this study are the size of the physical shut-in volumes and the liquid left in the separators for the DPC and Export trains. The physical volume is uncertain, and the liquid left in the separators will vary. Also, since the assumption of an empty separator for the Export trains was incorrect, the recycled gas amount will be too high. These uncertainties will affect the gas recycling amount, thereby the gas recycling time, saved emissions, and cost. These results should therefore be evaluated as uncertain.
A secondary uncertainty is the difference between the temperature of the sections at 26 barg and seven barg compared with the collected data from Baze. This mainly affects the predicted inlet temperature of the compressor. However, based on the arguments of the discussion, implementing MEG injection before the PCVs and a heater for step one should be considered necessary.
Therefore, a more comprehensive investigation should be conducted to find precise shut-in volumes of all the sections at Kollsnes. The simulation model in UniSim Design should be implemented according to the P&ID specifications and with heat loss from the equipment.
Since the saved emission from the measure is low – only between 0.6 ktonne - compared with the total emission of 14.2 ktonne from flaring, other solutions should be investigated. A new flare gas recovery unit should be implemented in the LP flare gas system. This could result in 3.5 ktonne of CO2 savings per year. The LP compressor could be used together with the proposed design of this thesis. Then low-pressure gas and high-pressure gas could be recycled. This would allow Kollsnes to recycle most of the gas which now is routinely flared - if the pipe network were extended to all equipment - thereby helping Equinor reach zero routine flaring by 2030. | |