wrong initial conditions obtained in initial equation section

[Michael Wetter]

The attached plot shows results from the model
Buildings.HeatTransfer.Examples.ConductorSteadyStateTransient
with the assertEquality block configured so that its parameter startTime is 864000 and hence the block does not stop the simulation.

The figure shows on top the results from Dymola and on the bottom the result from the nightly built of OpenModelica and the trunk of ​https://github.com/lbl-srg/modelica-buildings (sha=42288ed91ac7ec992846335505403fe808fcfa86)

The section

initial equation 
  // The initialization is only done for materials that store energy.
    if not material.steadyState then
      if steadyStateInitial then
        if material.phasechange then
          der(u) = zeros(nSta);
        else
          der(T) = zeros(nSta);
        end if;
      else
        for i in 1:nSta loop
          T[i] = T_a_start+(T_b_start-T_a_start) * UA *
            sum(1/(if (k==1 or k==nSta+1) then UAnSta2 else UAnSta) for k in 1:i);
        end for;
      end if;
    end if;

seems to be processed incorrectly in OpenModelica. OpenModelica obtains the correct values T_a_start=T_b_start=293.15, but then computes wrong initial values for T. The expected result is that T[:]=293.15 at t=0.

[Michael Wetter]

Results of Dymola (top) and OpenModelica (bottom)

[Martin Sjölund]

Milestone renamed

[Martin Sjölund]

Milestone renamed