Hello All,

I am working on setting up a fuel delivery system for large capacity storage tanks. The fuel delivery system will deliver at the rate of 1500 l/min (higher is also possible) with/without using a pump (using gravity when no pump is used). This system is integrated with a measurement system which can detect the amount of fuel that has been discharged at a given time. This measurement system cannot differentiate between liquid and gas therefore, there is a gas sensor placed in the downstream (but well before the measurement system) which detects for the presence of bubble. When a bubble is detected the sensor/controller family shut's the discharge valve which is placed further downstream (therefore the chances of bubble escaping in the discharge is 0%). But the trouble here is to get this bubbles out of the system through a bypass line which takes nearly a minute (a loss of 1500 l of fuel). The discharge valve cannot be opened unless this bubble is out of the system which is confirmed by another sensor placed in the bypass line. Everytime a bubble is detected, almost a minute is lost. Therefore a total discharge volume of 20,000 l takes more than 30 mins when it should just take little more than 13 minutes. The trouble starts when the tank is almost empty. That means everything upto say 17,000 to 18,000 liter goes as per wish. An experiment carried out with a transparent glass system (a prototype of 1:4) shows that at this stage a whirlpool is formed above the outlet of the tank (the outlet is in the bottom face of the tank) which sucks the gas into the discharge line and guides it all the way upto the delivery valve (no gas sensor included in this experiment). CFD is being used in the development (since the problem is general, I am posting it in all the forums for expert opinions) together with experimentalist. In any case, both the experimental work (due to construction of large systems) as well as numerical (due to size of the geometry, mesh) is progressing slowly. Anyone have similar experience or any idea which can guide me to the solution. I have already one alternative and work is going on in another one. What I have learned till now is, I cannot get rid of this bubbles fully but can speed up the discharge process by reducing the time from 30 mins to lower. I am open for all kinds of ideas (even if it is going to lead to some fancy accessories at the end).

Thanks a lot. Best Regards.

Jan

its great food to thought! I dont think CFD will give you a economic(solution expenses and time) solution, but if the problem is seen from physics point of vies I xould see the solution is lying with an idea of reducing he vertex by some baffles. The baffles I think are exactly above the outlet in the tanlk and paralle to the tank axis. I will elaborate it in graphical form i you can provode me a prototype geometry details if this is not hampering confidentiality of the project, but again It given me a great food for thought, you are working on some interesting physics

please let me kno if it works sachcool80@gmail.com