Unfortunately I didn't saw such a thing there and I have seen 4-5 of this sensors in auto parts shop, and all of them look the same as mine and donot have any isolations.
I think my easiest way would be to lower the current, because I think isolating the fuel sensor would be too hard.

Dear Farshid,

there are two ways to make your tank possibly explode:

1. Your sensor heavily overheats and fires the fuel-air mix touching the sensor element.
2. Occuring of a spark inside the tank, which also fires the fuel-mix.

I think the following protection circuit will stop both mechanisms:




You can see, that two transzorbs are connected anti-parallelish. Here, the extremely low turn-on time of forward biased pn-junction is used to achieve an ultra fast clamping of any overvoltage hitting the sensor cable. Transzorbs are able to turn-on in less than 1nsec, while a standard 1N400X diode can need up to 10µsec!
You can also see, that a fuse is connected in series to the sensor cable.

The overheating can only occur, if too much current is flowing through tank sensor. But if you choose a rather small sensoring current, as you mentioned, then overheating is impossible.
Nevertheless, there can be circumstances, where the sensor gets too much voltage, though. Assume for instance a break of cable in combination with a direct connection to battery voltage. Rare, but possible. Then, transzorbs clamp the voltage to about 1V, or so. As consequence an enormous current is flowing through the transzorbs resulting in a fast smelting of fuse.

Can the transzorb be damaged before the fuse is acting?
No, specified damage of a transzorb is shortcircuiting! Only, when so much current is flowing, that the whole diode is smelting, there will be no shortcircuit. But in combination with a fuse, this is impossible.
Be careful, you must use a fuse, that can break very high currents. So, use a fuse with sand filling. Also, the fuse must not be located to close too the tank, because a fuse can become warm when acting. But better a warm fuse at save distance to tank, than a burning tank sensor, right?

Now about the sparking:
Sparking occurs, when there's enough electrical field strength that air becomes conducting, which is at about 1kV per millimeter. But because sensor is capsulated and both terrminals are centimeters away from grounded tank enclosure, only a relevant voltage BETWEEN both terminals, namely at contacts of sensor could cause a spark. And only, if the wiper eroneously loses connection to substrate. But by limiting the voltage between both terminals by the help of the two transzorbs, it's very unprobable, that a spark can be formed having enough energy to fire the fuel-air mix. Means, the spark energy is limited to very little values.
The transzorbs are so fast, that even ESD events coupled to the wires will be clamped!


Finally, a word about saftey:
Please do not make any experiments with fuel tank! The protection circuit above is no replacement for very very careful handling of this tank stuff! Do not feel more secure only because of this protection circuit! It can only help to make certain things to happen less probable, but not to convert a dangerous situation into a total safe one!!!!!! Please be very careful!!!!

Kai