I am doing a convex optimization problem related to reactive power management
on a network. When trying to compile my problem on CVXPY, I have a problem with two constraints
which do not comply with DCP rules. By doing some research, I saw that there is
a possibility of reformulating them in order to respect the DCP rules in particular by using
the cp.SOC function of CVXPY. The problem is that I don’t see how to do it.
The two constraints are as follows:

• V_n[idx_from_combined[l]] * V_n[idx_to_combined[l]] * (np.sin(theta_max))2 >= theta_l2: where V_n and theta_l are variables of the optimization problem. NB: This constraint is conical, therefore convex.
• Q_ol[l] * V_n[idx_from_combined[l]] >= (cp.sum(cp.square(P_sl[l]) + cp.square(Q_sl[l])))*X_l[l]:
  which is a rotated SOCP constraint with Q_ol, V_n, P_sl and Q_sl which are optimization variables of my problem

Can you please help me reformulate these constraints?

When I compile my problem, with those constraints written in their actual form, the problem
doesn’t follow the DCP rules.