Dinorwig is only part of the answer as you said, because the 'run around efficiency' is only about 55% when you consider the transmission, motor, pipe, turbine and transmission losses back to the network. It is far better to give customers a discount on their bills if they keep their electrical load very constant by installing heat-storage cookers and large hot water storage tanks.
I built the first tidal stream turbine over 20 years ago for £20,000 and made no profit, but those who jumped on the 'band-wagon' have made a fortune out of UK and EU funding. Tidal power is perfectly able to make a reasonable commercial profit, provided there aren't hundreds of bureaucrats and consultants wanting to climb on the backs of those who want to do something.
There shouldn't be any particular problem with silt and boulders on a hydro project, if the intake is correctly designed. I have worked on projects in Nepal, where the Monsoon brings everything down the river. You can't fight it, so you have to have a 'cunning plan' as they say!
Any reasonable hydro project should pay for itself in under 10 years without subsidy and last at least 50 years. The last lot of investors I talked to about a hydro plant wanted a three year payback or about 40% return on their investment, so I am not surprised that there aren't many projects being built.