The plot thickens. The flow switch works properly, which means RL7 should be deenergized. But even with the flow switch disconnected from the circuit, connection point 115 in the diagram has 12v, which is inconsistent with the logic diagram. So the problem is starting to look more like a problem...