Ramsey optimal policy yields zero approximated value of welfare

I am studying Ramsey optimal policy problem using mod.file from professor Pfeifer’s GitHub. It generates IRF well but consistently yields zero approximated welfare(for both conditional and unconditional). I experienced similar results from Willi Mutschler sample code file.
example.mod (7.2 KB)

It seems the problem comes from the below warning but I cannot exactly figure out why welfare estimates are zero.
Warning: Some of the parameters have no value (Omega, lambda, kappa, vartheta) when using stoch_simul. If these parameters are not initialized
in a steadystate file or a steady_state_model-block, Dynare may not be able to solve the model. Note that simul, perfect_foresight_setup, and
perfect_foresight_solver do not automatically call the steady state file.

How can I get exact welfare estimate by modifying above mod file?

Welfare is a stochastic object that depends on higher order properties of the model. In linear models, average welfare deviations from steady state are trivially zero.

In this youtube video, however, approximated planner objective value is not zero.(see 40:19).

Since the planner objective is a quadratic function of inflation and the output gap, 0 approximated value of unconditional welfare means that ramsey policy can achieve divine coincidence(i.e inflation=ouput gap = 0). This is also somewhat weird since under the existence of cost-push shock there exists trade-off between stabilizing inflation and stabilizing output gap.

My mistake. You want a second order approximation (order=2). At first order, you end up in steady state.