In that case sensors could simply be assumed to operate by sending out small probes via Alcubierre drive to the area of effect. That makes the most sense in light of how they work.

The issue of assuming that sensor range represents a number of probes being sent out is twofold. One, it requires that you have the ability to launch large numbers of probes in short periods of time. Two, it requires that the probes move very, very rapidly. If you assume that your ship has a sensor range of R tiles and a speed of S tiles per turn, and that each probe has a scanning range of r tiles, then the probes will need to move at a minimum speed of (R - r)*S/(1 tile) to allow the ship to scan out to its maximum sensor range in the time take by 1 move action, and you will need to launch (6*R)/r of these probes in the time taken by a single move action to fully cover the area within your ship's sensor range. If you increase the sensor range of the ship by N tiles, then you increase the minimum speed required by N*S/(1 tile) and you increase the number of probes required by 6*N/r. If a sensor module represents a probe-launching device, then it likely has the ability to launch X probes per time unit, with each probe moving at some speed s which is independent of the number of sensor modules. If each sensor module represents a part of a kind of probe catapult, then you're likely to see each sensor module increase the speed at which each probe moves by some amount but adding additional sensor modules is unlikely to significantly affect the launch rate of the probes. Either way, you're only halfway there - you need both the launch rate and the probe speed to increase linearly with the number of sensor modules installed, but you can only reasonably expect to see at most one of these terms increase linearly with the number of sensor modules installed.

Assuming that the probes have their own warp drives does not help this (and there's a reasonable case to be made that the probes could not be equipped thusly, at least not while being capable of the speed required), as then you run into the issue of requiring the probe launch rate to be proportional to the square of the number of sensor modules installed, which is unreasonable.

Diminishing returns on sensor range would sort this out.

Diminishing returns has its own problems. Namely, it can be quite a bit of a pain when you get bonuses and discover that your ships are now wasting space because that 10% bonus to sensor range means you now need 5 modules rather than the 6 you previously needed for your 8 sensor range, and that those 6 modules are not currently enough to push your sensor range out to 9 with the 10% bonus, so you ought to redesign the fleet to take advantage of the extra space you've managed to obtain from that bonus. It can also be a bit of a pain to figure out how much of a bonus you'll get for something with diminishing returns, which could become rather annoying with the upcoming mutually-exclusive technologies. +10% sensor power per sensor module isn't so useful when sensor power is converted to sensor range as R = 0.5*sqrt(B + S), where R is the sensor range, B is the base sensor power, S is the total sensor power of the sensor modules equipped, and sqrt(y) is the square root function. It also makes comparisons a bit more of a pain; maybe instead of an inverse square law we have R = B + sum(s/n, n = 1, N) where R is sensor range, B is the base sensor range, s is the bonus per module, and N is the total number of modules added. Is it better to increase s by 10%, or reduce the size of each module by 20% (=> up to 25% more modules can be installed for a given amount of space)?

Be careful what you ask for, or you might find yourself playing Spreadsheet Civilizations or Galactic Spreadsheets rather than Galactic Civilizations.