I see no self-contradiction in a consequentialist approach which just declares certain consequences (e.g. "murder and children crying") be be unacceptable.

Pardon me. I left off the technical qualifier for the sake of terseness. I have previously observed that all deontologial value systems can be emulated by (suitably contrived) consequentialist value systems and vice-versa so I certainly don't intend to imply that it is impossible to construct a consequentialist morality implementing this particular injunction. Edited to fix.

It is also contrary to virtually all consequentialist value systems of the kind actually held by people here or extrapolatable from humans. All consequentialist systems that match this criteria are abhorrent.