Error: Complex values ​​are not supported

Hi. I'm trying to perform a multiobjective optimization together with aspen plus. I'm using the optimization toolbox (gamultiobj). But after hours, I found the following error: "Complex values ​​are not supported". How should I proceed?

function [TAC] = Objective(x)  
global Aspen 
 
Aspen = actxserver('Apwn.Document.36.0'); 
    [~,mess]=fileattrib; 
    Aspen.invoke('InitFromArchive2',[mess.Name '\SimulationACRYL2.bkp']);
    Aspen.Visible = 1;
    Aspen.SuppressDialogs = 1;
    Aspen.Engine.Run2(1);
    while Aspen.Engine.IsRunning == 1
        pause(0.5);
    end
    
Stages = round(x(1) + 2);
Pressure = x(2);
Feed_Stage = round((Stages-2)*x(3) + 1);
Aspen.Tree.FindNode("\Data\Blocks\B2\Input\NSTAGE").Value = Stages;
Aspen.Tree.FindNode("\Data\Blocks\B2\Subobjects\Column Internals\INT-1\Input\CA_STAGE2\INT-1\CS-1").Value = Stages - 1;
Aspen.Tree.FindNode("\Data\Blocks\B2\Input\PRES1").Value = Pressure;
Aspen.Tree.FindNode("\Data\Blocks\B2\Input\FEED_STAGE\S50").Value = Feed_Stage;
    Aspen.Reinit; % Reinit simulation
    Aspen.Engine.Run2(1); %Run the simulation. (1) ---> Matlab isnt busy; (0) Matlab is Busy;
    time = 1;
    Error = 0;
    while Aspen.Engine.IsRunning == 1 % 1 --> If Aspen is running; 0 ---> If Aspen stop.
        pause(0.5);
        time = time+1;
        if time==40 % Control of simulation time.
            Aspen.Engine.Stop;
            Error = 1;
        end
    end
         
    
Conv = Aspen.Tree.FindNode("\Data\Results Summary\Run-Status\Output\PER_ERROR").Value; 
    if Error == 0 && Conv == 1
       % column cost
        Lang_F = 4.74;
        Diameter = Aspen.Tree.FindNode("\Data\Blocks\B2\Subobjects\Column Internals\INT-1\Input\CA_DIAM\INT-1\CS-1").Value;
        Number_trays = double(Aspen.Tree.FindNode("\Data\Blocks\B2\Input\NSTAGE").Value - 2);
        height_Column = 0.5*(Number_trays)+2;
        heat_reboiler = (Aspen.Tree.FindNode("\Data\Blocks\B2\Output\REB_DUTY").Value)*4.1868;
        Presure_project = 441.3;
        Tension = 108000;
        Column_thickness = ((Presure_project*Diameter)/((2*Tension*0.85)-(1.2*Presure_project))) + 0.003;
        Density_carbon_steel = 7861.1;
        Column_side_volume = 3.14*Column_thickness*height_Column*Diameter;
        Column_side_mass = (Column_side_volume*Density_carbon_steel)*2.205;
        External_radius = (Diameter + 2*Column_thickness)/2;
        Cover_volume = 4*3.14*(External_radius^2)*Column_thickness*2;
        Cover_mass = (Cover_volume*Density_carbon_steel)*2.205;
        total_mass =  Column_side_mass + Cover_mass;
        Cv = exp((7.2756 + 0.18255*(log(total_mass))) + (0.02297*((log(total_mass))^2)));
        Diameterft = Diameter*3.281;
        height_Columnft = height_Column*3.281;
        Cpl = 300.9*((Diameterft)^0.63316)*(height_Columnft^0.80161);
        Factor_material_column = 1;
        FNT = 1;
        FTT = 1;
        FTM = 1; 
        CBT = 468*exp(0.1739*Diameterft);
        CT = CBT*FTM*FTT*FNT*Number_trays;
        CP = Factor_material_column*Cv + Cpl;
        Cost_column = CP + CT;
        
        % condenser cost
        Temp_out_fluid_cold = 20;
        Temp_in_fluid_cold = 5;
        Temp_in_fluid_hot = Aspen.Tree.FindNode("\Data\Streams\S4\Output\TEMP_OUT\MIXED").Value; 
        Temp_out_fluid_hot = Aspen.Tree.FindNode("\Data\Streams\S4\Output\TEMP_OUT\MIXED").Value;
        Temperature_exchange_termic = ((Temp_in_fluid_hot - Temp_out_fluid_cold) - (Temp_out_fluid_hot - Temp_in_fluid_cold))/log((Temp_in_fluid_hot - Temp_out_fluid_cold)/(Temp_out_fluid_hot - Temp_in_fluid_cold)); 
        Factor_exchange_termic = 0.85;
        coefficient_exchange_termic = 850;
        Heat_condenser = (Aspen.Tree.FindNode("\Data\Blocks\B2\Output\COND_DUTY").Value)*4.1868;
        Area_condenser = ((Heat_condenser)/(Temperature_exchange_termic*coefficient_exchange_termic*Factor_exchange_termic))*10.76;
        Cost_condenser = exp(11.0545 - 0.9228*(log(Area_condenser)) + 0.09861*(log(Area_condenser)^2)); 
        Pipe_length = 1;
        Type_Material_Condenser = 1;
        pressure_condenser = 1;
        Condenser_total_cost = Cost_condenser*Pipe_length*Type_Material_Condenser*pressure_condenser;
        
        %Reboiler cost
        Temp_in_fluid_hot_reboiler = 127.5;
        Temp_out_fluid_hot_reboiler = 127.5;    
        Temp_out_fluid_cold_reboiler = Aspen.Tree.FindNode("\Data\Streams\12\Output\TEMP_OUT\MIXED").Value;
        Temp_in_fluid_cold_reboiler = Aspen.Tree.FindNode("\Data\Streams\12\Output\TEMP_OUT\MIXED").Value;
        Temperature_exchange_termic_reboiler = ((Temp_in_fluid_hot_reboiler - Temp_out_fluid_cold_reboiler) - (Temp_out_fluid_hot_reboiler - Temp_in_fluid_cold_reboiler))/log((Temp_in_fluid_hot_reboiler - Temp_out_fluid_cold_reboiler)/(Temp_out_fluid_hot_reboiler - Temp_in_fluid_cold_reboiler)); 
        Factor_exchange_termic_reboiler = 0.85;
        coefficient_exchange_termic_reboiler = 1560;
        Area_reboiler = ((heat_reboiler)/(Temperature_exchange_termic_reboiler*coefficient_exchange_termic_reboiler*Factor_exchange_termic_reboiler))*10.76;
        Cost_reboiler = exp(11.967 - 0.8709*(log(Area_reboiler)) + 0.09005*(log(Area_reboiler)^2)); 
        Pipe_length_reboiler = 1;
        Type_Material_reboiler = 1;
        pressure_reboiler = 1;
        Reboiler_total_cost = Cost_reboiler*Pipe_length_reboiler*Type_Material_reboiler*pressure_reboiler;
        % cost of utilities
        Chilled_water_cost = ((Aspen.Tree.FindNode("\Data\Blocks\B2\Output\COND_DUTY").Value)*4.1868)*31536000*0.0000000009258;
        Steam_Cost = ((Aspen.Tree.FindNode("\Data\Blocks\B2\Output\REB_DUTY").Value)*4.1868)*0.000000004607*31536000;
        
        TAC(1) = (Cost_column + Condenser_total_cost + Reboiler_total_cost)*Lang_F*1.14*1.1906/10;
        TAC(2) = Steam_Cost + Chilled_water_cost;
          
    else
        TAC(1) = inf;
        TAC(2) = inf;
        
    end
    
Aspen.Quit();
end

Answer 1

[Alan Weiss]

I don't know where your objective function creates complex values, though I suspect that one or more of your logarithm calls has a negative number as an argument. You can fix this problem by making your code more robust. Test all inputs to logarithms (or square roots or similar) to ensure that they are not negative before applying the functions. Or, if you can, impose bounds or linear inequality constraints to ensure that negative arguments do not get passed. Unfortunately, gamultiobj does not satisfy nonlinear inequality constraints at intermediate iterations, so you cannot fix the problem by imposing nonlinear constraints.

Alan Weiss

MATLAB mathematical toolbox documentation

[Eduardo Braga]

Thank you very much! Now the following error is happening: Server Creation Failed: Server execution failed. Do you know how I should proceed?

[Alan Weiss]

Sorry, I don't know where that error arises. Is that from Aspen? If so, it is not a MATLAB issue, and I don't know anything about it.

Alan Weiss

MATLAB mathematical toolbox documentation