Here are some concerns I have with your design, and some ways to solve them. I'll leave it up to you to build and combine these modules, and test them together.
First, a clean supply of +15V to provide up to 100mA without needing a heatsink on U1, easily sufficient for a modest load like the comparator and its periphery:
simulate this circuit – Schematic created using CircuitLab
Your L1 hints at a dirty supply, but an inductor in that position would likely cause more problems than it solves. I achieve a similar filtering function with R1 and C1 here, which will mitigate momentary spikes or dips on the 24V supply, and the +15V at P should be clean and stable. Install C2 and C3 very close to regulator U1's terminals.
I've chosen 15V because it's well under the \$V_{GS(MAX)}\$ limit of ±20V of most common MOSFETs, so that the comparator output can drive an N-channel MOSFET directly, but is large enough to switch the transistor fully on. By the way, your transistor symbol is a darlington pair, which could also work, but it is an inappropriate symbol for a MOSFET, and very misleading.
Your arrangement of C4 and R6 is a high-pass filter, when you should be passing a DC potential; swap them, to form a low-pass filter instead. Also, 500Ω is a heavy load on the sensor output, you can increase R6 significantly, and lower C4 commensurately. D5 looks like it's supposed to prevent excessive voltage at the comparator input, but it's the wrong way around. This is what I think you meant, for illustration, but will need further modification still so don't use this as shown:
simulate this circuit
Instead of low-pass filtering the sensor signal, with a capacitor, a better solution to mitigate noise and oscillation is hysteresis, which you implement with a small amount of positive feedback:
simulate this circuit
C5 isn't optional, it's a supply decoupling capacitor to keep the comparator's supply voltage steady while currents change during switching of states. Install it close to the IC's power pins. Positive feedback is via R7 and D2. The values shown will produce about 0.5V of hysteresis (the "gap" between switching thresholds, which can be seen here, as \$V_{IN}\$ (blue) crosses up and down through those thresholds. Orange is the comparator output \$V_Z\$, which requires pull-up resistor R8 due to the open-collector outputs of the LM393:
The 0.5V threshold gap (between the green markers) is probably too much for your application, but you can reduce it by increasing R7. Doubling R7 will roughly halve the hysteresis gap, you should experiment with the simulation to find the exact gap you require.
R8, the pull-up, needs to be as small as possible, to source as much current as possible to charge the MOSFET's gate capacitance quickly, but not so small as to overload the comparator output. The value 2.7kΩ should produce about 6mA, which the comparator is able to sink, according to the LM393 datasheet.
It will be difficult to interface an indicator LED to this open-collector output comparator, without an additional transistor, because that output cannot source current to illuminate an LED - it can only sink current. Therefore I suggest placing the LED across the relay coil instead. Usually I would recommend a 1N4001 fly-wheel diode to protect the transistor when the relay switches off, but with a fragile LED there too, I think you'll need something faster, like a UF5401, which is D4 below, replacing your own D2.
The MOSFET you choose, the 4N60C is not recommended, since while it is robust, it has significant on-resistance, around 2Ω. If the relay coil requires, say, 2A of current, this transistor will dissipate a lot of power, requiring heat-sinking:
$$ P = I^2R = (2{\rm A})^2 \times 2{\rm \Omega} = 8{\rm W} $$
There are dozens of better choices, with on-resistance \$R_{DS(ON)}\$ under 100mΩ, like the IRFZ44, which claims \$R_{DS(ON)}\$ under 30mΩ:
simulate this circuit
Strictly speaking, D5 isn't necessary, but without knowing more about the 24V supply, and the relay, I feel it prudent to protect both MOSFET Q1's gate and the comparator, from any shenanigans going on at Q1's drain.
