Brian Sheppard's thesis "Towards Perfect Play of Scrabble" [0] has a lot of good info on Scrabble strategy which can be applied to human play as well as to AI (though obviously things like B* and Monte Carlo search aren't realistic for human players)

    [0] https://project.dke.maastrichtuniversity.nl/games/files/phd/Shepard_thesis.pdf