Calculation of an Variable in Custom Simscape Component seems to be wrong

I'm testing a component of the Custom Simscape Library i currently develop.

There the following problem occurs: the value of the variable var_h_out does not match the value it should have when one take the values of T_e_out and T_e into account. However if I copy paste the corresponding equations and their parameters into a new component (see code below) and define T_e_out and T_e as inputs, var_h_out has the expected value. For T_e_out and T_e i used the values that the corresponding variables have in the original component. The value of var_h_s is correct in both cases.

Values:

T_e = 282.561 K

T_e_out = 282.582 K

var_h_s = 255.051 kJ/kg

right value of var_h_out = 404.118 kJ/kg

wrong value of var_h_out = 422 kJ/kg

What could be the reason for this error?

component Test
    
    parameters
      %saturated vapour specific enthalpy%%%%%%%%%%%%%%%%% h_s == bo * (b1 + b2*T_s - b3*T_s^2 - b4*T_s^3 + b0)
        bo = {1, 'J/kg'};
        b0 = {149048,'J/kg'};
        b1 = {249455,'1'};
        b2 = {606.163,'1/K'};
        b3 = {1.05644,'1/K^2'};
        b4 = {0.0182426,'1/K^3'};
        
        
        %Superheated vapour  specific enthalpy
        %del_T == T - T_s%%%%%%%%%%%%%%%%%%%%%%%%%%% h_sh == h_s*(c1 + c2*del_T + c3*del_T^2 + c4*delt_T*T_s - c5*del_T^2 * T_s + c6*del_T*T_s^2 - c7*delT^2*T_s^2)
        c1 = {1, '1'};    
        c2 = {0.00348186, '1/K'}; 
        c3 = {1.6886e-06, '1/K^2'}; 
        c4 = {9.2642e-6, '1/K^2'}; 
        c5 = {7.698e-8, '1/K^3'}; 
        c6 = {1.7070e-7, '1/K^3'}; 
        c7 = {1.2130e-9, '1/K^4'}; 
        parb0 = {149048, 'J/kg'};
        par_Kelvin = {273.15, 'K'};
    end
    
    inputs
    T_e = {0, 'K'};
    T_e_out = {0, 'K'};
    end
    
    outputs
    h_out   = {0, 'kJ/kg'};
    h_s     = {0, 'kJ/kg'};
    end
    
    variables
    var_h_out   = {0, 'kJ/kg'};
    var_h_s     = {0, 'kJ/kg'};
    var_T_e_out_K       = {0, 'K'};
    var_T_e_K           = {0, 'K'};
    end
    
    equations
     var_h_s == bo * (b1 + b2*(T_e - par_Kelvin) - b3*(T_e - par_Kelvin)^2 - b4*(T_e - par_Kelvin)^3);
     var_T_e_out_K == T_e_out - par_Kelvin;
     var_T_e_K == T_e - par_Kelvin;
     var_h_out == var_h_s*(c1 + c2*(var_T_e_out_K - var_T_e_K) + c3*((var_T_e_out_K - var_T_e_K)^2) + c4*(var_T_e_out_K - var_T_e_K)*var_T_e_K  - c5*((var_T_e_out_K - var_T_e_K)^2) * var_T_e_K + c6*(var_T_e_out_K - var_T_e_K)*(var_T_e_K^2) - c7*((var_T_e_out_K - var_T_e_K)^2)*var_T_e_K^2) + parb0;  
     h_out   == var_h_out;
     h_s     == var_h_s;
     
    end
    
    
end    

Answer 1

[Juan Sagarduy]

Hello Moritz

It looks like Simscape might initialize the variables Te and Te_out with a given value that then leads a different enthalpy as the differential algebraic equation system is solved in your model.

By defining Te and Te out as inputs, you decouple them from the state space problem as you actually prescribe them into your block. With a technical support request to MathWorks, this could be solved I believe.

Regards / Juan