I only looked at the schematics for a minute or so, but there are immediately visible problems with it.
You seem to have an excessive number of pullups on the I2C bus to the point where it may not function anymore. A quick count suggests you're around 1k in total, which means you're right on the edge of the I2C spec which, if memory serves, outlines a max. of 3mA sink.
In the voltage detection circuit, there's no GND reference. I also doubt the effectiveness of zeners as input protection for the uC and would suggest using two Schottky's per channel, one to prevent against negative pulses and one protecting against pulses higher than the uC's Vcc.
Capacitors around the CH40T's (you mean CH340T?) crystal oscillator are missing; try 18 ~22pF.
The micro-SD interface is mysteriously powered through a 4k7 resistor; I'm not aware of any need or benefit in limiting the current to an SD interface to <1mA and I would expect to render it dysfunctional.
Decoupling capacitors around the 328P are missing.
I'd protect the BC548 by placing a freewheel diode across the unspecified buzzer in case a type is installed that causes back-EMF problems. It's a $0.01 insurance.
I'm surprised at the lack of any in-system programming provisions for the 328P, either through SPI or Serial. On the other hand, I'm also surprised at the presence of a CH(3?)40T linked to a MAX232, creating an onboard USB-Serial interface...that doesn't interface with anything else on the board??
So a lot is happening that raises questions. What's going on here?