Look at this doc it may be helpful for your topic .Good luck.

For example, talking about the non-linear term in the fluid mechanics equations, often the best choice is to use the conservative form. That means the divergence form for FD and the  integral form for FV methods.

 You can see pejan 's book

That the discretization preserves the  properties of the equation, such as conservation laws, to the precision expected.

Use of conservative form if possible, sometimes even makes a scheme stable when a non-conservative form does not.  Or try 'lagging the coefficients' where you evaluate one term from the previous iteration.