The problem in such cases is that the waste water contains x mM of an offending compound, which have to be brought down to y mM. The amount of enzyme required depends on the volume of waste water, the concentration difference x-y and the incubation time allowed.
1 U of enzyme breaks down 1 µmol of substrate per min under substrate saturation (this is an oldfashioned way to express activity, nowadays we use katal = mol/s). As the reaction proceeds, [S] and hence v decrease, if you plot [S] over time, you get a hyperbola. Eventually, you'll reach the equilibrium concentration of S and P, then the overall reaction velocity becomes 0 and [S] no longer changes.
I'd determine these hyperbolae for different [E] and then select the lowest [E] that gives the required degree of purification in the time allowed. Note that under the conditions of waste water treatment, you may not reach the enzymatic activity stated for the enzyme, because of temperature, pH, substrate concentration, presence of inhibitors... So there is no way to predict the optimal [E], you need to do the experiments.