YES

TRS:

  w(r(x)) -> r(w(x))
  b(r(x)) -> r(b(x))
  b(w(x)) -> w(b(x))

max/plus interpretations on N:

  w_A(x1) = max{0, x1}
  w#_A(x1) = max{0, x1}
  r_A(x1) = max{0, x1}
  r#_A(x1) = max{0, x1}
  b_A(x1) = max{0, x1}
  b#_A(x1) = max{0, x1}

precedence:

  b > w > r