I have participated in a number of detailed HAZOP analyses. In that method, you consider possible hazardous situations that could arise from things like: High Flow, Low Flow, High Pressure, Low Pressure, etc. If you find a hazardous situation, you consider the consequences and likelihood. Based on that, you determine how many layers of protection you need to protect against that outcome. In that analysis, a check valve cannot be considered for credit as a layer of protection. This is because check valves frequently stick open. This failure cannot be predicted or prevented. And, even if the check valve does close, there is a very good chance it will leak over time.
This does not mean that check valves do not serve a useful purpose. It does not mean that they are not needed. We install check valves on the discharge of all centrifugal pumps (expect molten sulfur) and most rotary positive displacement pumps.
I took the question to ask if you could consider a check valve as a tight shut-off, leak tight valve that would prevent the discharge pressure of the running pump from pressuring up the standby pump if he suction valve to the standby pump were closed. If that is the question, the answer would be no.
I am assuming that each pump has its own suction block valve and those are the valves the OP was asking about. If the suction block valve was accidentally closed on the running pump, I would expect that pump to fail. But I would not expect any effect on the standby pump.
If I had a pump that was not rated for the discharge pressure of the other pump running in parallel, I would not consider the check valve as protection. If the likelihood and consequences of an operator accidentally closing the suction block valve were high, I would insist on another protection such as a PSV inside the block valves. If the overpressure potential was less than the hydro-test pressure of the pump, but still above its pressure rating, I might consider a lesser protection such as a high pressure alarm. Since the block valve can only be closed if an operator is present, I might be able to take credit for operator intervention. But, if the check valve were to leak slowly over many hours, the operator could be long gone before the overpressure event occurred. All of this would be taken into account in the standard HAZOP method we use.
Johnny Pellin
