Error when running simulation

i need help in solving the error i got when runing the simulation;

Error using sqpInterface

Nonlinear constraint function is undefined at initial point. Fmincon cannot continue.

Error in fmincon (line 900)

[X,FVAL,EXITFLAG,OUTPUT,LAMBDA,GRAD,HESSIAN] = sqpInterface(funfcn,X,full(A),full(B),full(Aeq),full(Beq), ...

Error in nlmpc/nlmpcmove (line 174)

[z, cost, ExitFlag, Out] = fmincon(CostFcn, z0, A, B, [], [], zLB, zUB, ConFcn, fminconOpt);

Error in nmpcblock_interface (line 163)

[mv, ~, Info] = nlmpcmove(nlobj, x, lastmv, ref, md, Options);

------------------------------------------------------

End of Error Report

------------------------------------------------------

the code bellow is the NMPC code

clc, clear all
%% Simulink Model
% To run this example, Simulink(R) is required.
if ~mpcchecktoolboxinstalled('simulink')
    disp('Simulink is required to run this example.')
    return
end
%%initialize NMPC object and set properties
nx = 12;
ny = 6;
nlobj = nlmpc(nx, ny, 'MV', [1,2,3,4], 'UD', [5,6,7,8,9,10]);
%The prediction model sample time is the same as the controller sample time
Ts = 2;
nlobj.Ts = Ts;
%Prediction and control horizon
nlobj.PredictionHorizon = 12;
nlobj.ControlHorizon = 2;
%weights on output and manipulated variables
nlobj.Weights.OutputVariables = [100,100,100,100,100,100];
nlobj.Weights.ManipulatedVariables = [1,1,1,1];
nlobj.Weights.ManipulatedVariablesRate = [0.1,0.1,0.1,0.1];
%constraints on manipulated variables
nlobj.MV(1).RateMin = -200;
nlobj.MV(2).RateMin = -150;
nlobj.MV(3).RateMin = -10;
nlobj.MV(4).RateMin = -150;
nlobj.MV(1).RateMax = 200;
nlobj.MV(2).RateMax = 150;
nlobj.MV(3).RateMax = 10;
nlobj.MV(4).RateMax = 150;
nlobj.MV(1).Min = 1500;
nlobj.MV(2).Min = 1300;
nlobj.MV(3).Min = 400;
nlobj.MV(4).Min = 12400;
nlobj.MV(1).Max = 2600;
nlobj.MV(2).Max = 2000;
nlobj.MV(3).Max = 600;
nlobj.MV(4).Max = 13200;
%constraint on output variables to keep them within reasonable bounds
nlobj.OV(1).Min = 0;
nlobj.OV(2).Min = 920;
nlobj.OV(3).Min = 0.85;
nlobj.OV(4).Min = 0;
nlobj.OV(5).Min = 0;
nlobj.OV(6).Min = 770;
nlobj.OV(1).Max = 0.02;
nlobj.OV(2).Max = 935;
nlobj.OV(3).Max = 1.25;
nlobj.OV(4).Max = 0.02;
nlobj.OV(5).Max = 0.02;
nlobj.OV(6).Max = 800;
%% specify the nonlinear state and output functions
nlobj.Model.StateFcn = 'fccuStateFcnCT';
nlobj.Model.OutputFcn = 'fccuOutputFcn';
%% validate the nonlinear MPC object with initial conditions
y0 = [0.002156;427.2; 0.2136; 0.03598; 0.002156; 449.3; 0; 0; 0; 0; 0; 0];
u0 = [2080; 1720; 503; 13028];
validateFcns(nlobj, y0, u0);

[Torsten]

Nonlinear constraint function is undefined at initial point. Fmincon cannot continue.

Before starting the simulation, call "ConFcn" with your initial vector z0. Most probably, something undefined will be returned. At least that's what the error message says.

[Swastik Sarkar]

Hi @Hammed,

I attempted to run the code you provided, but encountered a different error than the one you specified

Error using nlmpc/validateFcns (line 179)
Function fccuStateFcnCT does not exist.

Is there any additional information or configuration I need to specify before executing this script?

[Sandeep Mishra]

Hi @Hammed

Can you please share the 'fccuStateFcnCT'and 'fccuOutputFcn' functions for better analysis?

[Hammed]

@Sandeep Mishra @Swastik Sarkar thanks for your response. I try to modify the simulation to control 4 output instead of 6 that i was trying to do earlier, i will upload the code below, though i am still experencing the same issue

NMPC code

clc, clear all
%% Simulink Model
% To run this example, Simulink(R) is required.
if ~mpcchecktoolboxinstalled('simulink')
    disp('Simulink is required to run this example.')
    return
end
%%initialize NMPC object and set properties
nx = 6;
ny = 4;
%'UD', [5,6,7,8,9,10]
nlobj = nlmpc(nx, ny, 'MV', [1,2,3,4]);
%The prediction model sample time is the same as the controller sample time
Ts = 2;
nlobj.Ts = Ts;
%Prediction and control horizon
nlobj.PredictionHorizon = 10;
nlobj.ControlHorizon = 2;
%weights on output and manipulated variables
nlobj.Weights.OutputVariables = [1000,1000,1000,1000];
nlobj.Weights.ManipulatedVariables = [0.01,0.01,0.01,0.01];
nlobj.Weights.ManipulatedVariablesRate = [100,100,100,100];
%constraints on manipulated variables
nlobj.MV(1).RateMin = -200;
nlobj.MV(2).RateMin = -150;
nlobj.MV(3).RateMin = -10;
nlobj.MV(4).RateMin = -150;
nlobj.MV(1).RateMax = 200;
nlobj.MV(2).RateMax = 150;
nlobj.MV(3).RateMax = 10;
nlobj.MV(4).RateMax = 150;
nlobj.MV(1).Min = 1500;
nlobj.MV(2).Min = 1300;
nlobj.MV(3).Min = 400;
nlobj.MV(4).Min = 12400;
nlobj.MV(1).Max = 2600;
nlobj.MV(2).Max = 2000;
nlobj.MV(3).Max = 600;
nlobj.MV(4).Max = 13200;
%constraint on output variables to keep them within reasonable bounds
nlobj.OV(1).Min = 0;
nlobj.OV(2).Min = 920;
nlobj.OV(3).Min = 0.85;
nlobj.OV(4).Min = 770;
nlobj.OV(1).Max = 50;
nlobj.OV(2).Max = 935;
nlobj.OV(3).Max = 1.25;
nlobj.OV(4).Max = 800;
%% specify the nonlinear state and output functions
nlobj.Model.StateFcn = 'fccuStateFcnCT';
nlobj.Model.OutputFcn = 'fccuOutputFcn';
%% validate the nonlinear MPC object with initial conditions
y0 = [0.002156;427.2; 0.2136; 0.03598; 0.002156; 449.3];
u0 = [2080; 1720; 503; 13028];
validateFcns(nlobj, y0, u0);
fccuStateFcnCT
function dydt = fccuStateFcnCT(y, u)
%definig the parameters
deltaHfv = 498;
deltaHcr = 506;
deltaHrg = -394200;
cpo = 3.1335;
cpa = 1.074;
cps = 1.005;
sf = 14.57;
sc = 787.9601;
Ecc = 41.79;
Ecr = 101.5;
Ecb = 158.6;
Din = 915.1;
R = 0.0083143;
%kfo = 0.8;
kcc = 0.0189;
kcr = 19620;
Oin = 0.2136;
kcb = 187940;
ksig = 0.1197;
alpha = 0.12;
Vc = 175738;
Va = 20;
Pris = 100000;
Vris = 2500;
%Frc = 13028;
%Fin = 1720;
%Fa = 2080;
%Toil = 503;
Ta = 298;
%yf0 = 0.8;
%tc = 1.453;
%phi = 0.724;
%%% state parameters
Crc = y(1);
Treg = y(2);
Ofg = y(3);
Ccat = y(4);
Csc = y(5);
Tris = y(6);
%%% input parameters
Fa = u(1);
Fin = u(2);
Toil = u(3);
Frc = u(4);
dydt = zeros(6,1);
dydt(1) = ((Frc*(Csc - Crc))/Vc) - kcb*Ofg*Crc*exp(-Ecb/(R*(Treg+510)));
dydt(2) = ((Frc*(Tris - Treg))/Vc) + (((cpa/cps)*Fa*(Ta - Treg)))/(Vc) +  ((-deltaHrg*kcb*Ofg*Crc*(exp(-Ecb/(R*Treg)))))/(sf);
dydt(3) = ((0.0313*Fa*(Oin - Ofg)))/(Va) - (ksig*kcb*Ofg*Crc*(Vc/Va)*exp(-Ecb/(R*(Treg+510))));
dydt(4) = (-Frc*Ccat/Vris) + (kcc*Pris*exp(-Ecc/(R*(Tris- 60)))/((Ccat*Crc^0.06)));
dydt(5) = Frc*(Crc - Csc)/Vris + kcc*Pris*exp(-Ecc/(R*(Tris- 60))/(Ccat*Crc^0.06));
Kr = kcr*exp(-Ecr/R*(Tris - 60))/(Ccat*Crc^0.15);
dydt(6) = (Frc*(Treg - Tris))/(Vris) + (cpo/(cps*Vris))*Fin*(Toil - Tris) + (0.875*(-deltaHfv*Fin))/(sc*Vris) + (-deltaHcr*Pris*Din*Fin*Kr)/(2*sc*(Fin + 0.0313*Vris*Pris*Din*Kr));

fccuOutputfcn
function k = fccuOutputFcn(y, u)
kcr = 19620;
Ecr = 101.5;
R = 0.0083143;
Frc = u(4);
Fin = u(2);
Tris = y(6);
Ccat = y(4);
Crc = y(1);
alpha = 0.12;
yf0 = 0.8;
phi = 0.724;
tc = 1.453;
Kr = kcr*exp(-Ecr/R*(Tris - 60))/(Ccat*Crc^0.15);
COR = Frc/Fin;
yf = (alpha*yf0)/(alpha + Kr*phi*(1 - exp(-alpha*tc*COR)));
k = [yf;y(2);y(3);y(6)];

fccuMeasFcn code
function k = fccuMeasFcn(y, u)
kcr = 19620;
Ecr = 101.5;
R = 0.0083143;
Frc = u(4);
Fin = u(2);
Tris = y(6);
Ccat = y(4);
Crc = y(1);
alpha = 0.12;
yf0 = 0.8;
phi = 0.724;
tc = 1.453;
Kr = kcr*exp(-Ecr/R*(Tris - 60))/(Ccat*Crc^0.15);
COR = Frc/Fin;
yf = (alpha*yf0)/(alpha + Kr*phi*(1 - exp(-alpha*tc*COR)));
k = [yf;y(2);y(3);y(6)];

[Sandeep Mishra]

Hi @Hammed

I tried running your simulink file and found there exist one more function 'fccuStateFcnDT' required to run the simulation.

Are you treating the 'fccuStateFcnCT' function same as 'fccuStateFcnDT' function? If i do so, then i am getting error of signal dimension mismatch at Mux6.

[Hammed]

@Sandeep Mishra @Swastik Sarkar thanks for the response, i forgot to upload the fccuStateFcnDT, i am not treating it as fccuStateFcnDT,

fccuStateFcnDT code is uploaded below;

function yk1 = fccuStateFcnDT(yk,uk)
hstep = 1;
yk1 = yk(:);
Nsteps = 2;
uk1 = [uk(:)];
for i = 1:Nsteps
    yk1 = yk1 + hstep*fccuStateFcnCT(yk1,uk1);
end

i will be awaiting your response

[Sandeep Mishra]

@Hammed

When I run the Simulink file after executing the script and adding all the functions, I get the following error:

Error:Invalid setting for output port dimensions of 'fccu_2023b_4_output/Mux6'. The dimensions are being set to 1. This is not valid because the total number of input and output elements are not the same

Error:Error in port widths or dimensions. 'Input Port 1' of 'fccu_2023b_4_output/Zero-Order Hold' is a one dimensional vector with 1 elements.

Am I missing something or doing something wrong?

[Hammed]

@Sandeep Mishra

Thanks for the response, what version of matlab are you using?? besides, did you run the MNPC code in matlab environment before running the simulink code?

[Hammed]

@Sandeep Mishra

i was having same issue with matlab2024a but i was unable to resolve it, but it worked on 2023b. So i had to install 2023b

[Sandeep Mishra]

@Hammed

Yes, I ran the MNPC code in matlab environment before running the simulink code and I am using MATLAB 2024a.

[Sandeep Mishra]

@Hammed

In which version of MATLAB are you encountering the 'Error using sqpInterface' issue? What error message did you receive in MATLAB 2024a?

[Hammed]

@Sandeep Mishra

Thanks for your response.

i am receiving 'Error using sqpInterface' issue in 2023b but i am reciving this particular error in 'Error:Invalid setting for output port dimensions of 'fccu_2023b_4_output/Mux6'. The dimensions are being set to 1. This is not valid because the total number of input and output elements are not the same

Error:Error in port widths or dimensions. 'Input Port 1' of 'fccu_2023b_4_output/Zero-Order Hold' is a one dimensional vector with 1 elements.' in 2024a that is why i switched to 2023b