A couple of points…

1. Unless your domestic water system is different to mine, every change in water flow is the direct result of a human hand turning a tap. Transistor-based electronics is different — it’s not just about switches controlling flows of current, its also about flows of current turning switches on and off. In short, it's about switches that switch switches... This is different from the point about complexity, since even a very simple transistor-based system can function in the way I’m describing. So I don't think the water system is a good analogy.

2. You argue from the premise that we don’t know how the brain produces consciousness, to your conclusion “computers cannot have feelings. They are not conscious.” In arguing this way, you’re using a lack of knowledge to supposedly prove a strong assertion. This sort of argument has been a recognised fallacy since the 17th century. John Locke called it “argumentum ad ignorantiam”.