How to fix NLMPC validateFcn() error in input arguments

Hello all,

I am using MPC to design Energy Management System (EMS) of an energy storage, charging and discharging (MV) at differnt time of the day, looking for arbitrage based on load and electricity price (MD of my MPC) variation.

The state x and output y are the State of Charge (SOC). (OutputFcn as in y = x)

The objective function is customized cost (forecasted for the next 24 hours).

I ran into validateFcn error.

Error using nlmpc/validateFcns (line 179)

Expecting 6 input arguments but "Model.StateFcn" appears to take 4 inputs.

Error in MPC_Toolbox (line 68)

validateFcns(nlobj, x0, mv, md, params);

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Based on what I have read here https://www.mathworks.com/help/mpc/ug/specify-cost-function-for-nonlinear-mpc.html , I assume the inputs of State Fuction are required to have 3 inputs maximum.

Both my nlmpc code and stateFun are attached.

Thank you for your help. It's my 1st question here so please let me know if I need to provide any additional files.

function Xnext = myStateFunction(X, U, params)
    % params = [Ts; eta_dis; eta_cha; E_rated]
    Ts       = params{1}; 
    eta_dis  = params{2};
    eta_cha  = params{3};
    E_rated  = params{4}; % rated energy of BESS
    U_mv = U(1); 
    if U_mv >= 0 % discharging 
        deltaE = -U_mv * Ts / (eta_dis * E_rated);
    else % charging 
        deltaE = -U_mv * Ts *  eta_cha  / E_rated;
    end
    % next SOC
    Xnext = X + deltaE;

end

clear;clc;
%% ————————parameters————————
Ts = 5/60; % 5min 
eta_dis = 0.95;
eta_cha = 0.95;
eta_trans = 0.98;
E_rated = 1000/1000; % 1 MWh
u_min = -0.4; % 400 kW (charge) 
u_max = 0.4; % 400 kW (discharge)
% du_min = -0.1;
% du_max = + 0.1;
SOC_max = 0.9;
SOC_min = 0.1;

Np = 288;
Nc = 288;

%% ———————— nlmpc object ————————
nx = 1;   % state：SOC
ny = 1;   % ov：SOC

% creation of nlmpc
nlobj = nlmpc(nx, ny, 'MV', 1, 'MD', 2: 577);
% specify time
nlobj.Ts                  = Ts;
nlobj.PredictionHorizon   = Np;
nlobj.ControlHorizon      = Nc;

%% ———————— Specify Model Function ————————
nlobj.Model.StateFcn = "myStateFunction";
nlobj.Model.OutputFcn = "myOutputFunction";
nlobj.Model.NumberOfParameters = 4;
nlobj.Model.IsContinuousTime = false;

%% ———————— Constraints ————————
% MV（P_Bess） limits
nlobj.ManipulatedVariables.Max = u_max;
nlobj.ManipulatedVariables.Min = u_min;
% MV rate limits 
% nlobj.ManipulatedVariables.RateMax = u_max;
% nlobj.ManipulatedVariables.RateMin = u_min;
% 
nlobj.ManipulatedVariables.Name = "P_Bess";
% SOC limits （states）
nlobj.States.Max = SOC_max;
nlobj.States.Min = SOC_min;
% SOC limits（output variables）
nlobj.OutputVariables.Max = SOC_max;
nlobj.OutputVariables.Min = SOC_min;
 
nlobj.Optimization.CustomIneqConFcn = "myIneqConFunction";
%% ———————— Customized Cost ————————
nlobj.Optimization.ReplaceStandardCost = true;
nlobj.Optimization.CustomCostFcn = @(x,u, e, data)econCost(X, U, e, data);

%% ———————— validation  ————————
x0    = 0.5;    % initial SOC
params = {Ts, eta_dis, eta_cha, E_rated};
mv = -u_min +(u_max - u_min) * rand(1);

md_load = getLoad_v2(1, Ts, 1, "weekday") ;
md_price = getPrice(1);
md = [md_load; md_price]';

validateFcns(nlobj, x0, mv, md, params);

[Bay Jay]

You stated 4 parameters.

Matlab expect you to pass the parameters as comma seperated at the last entry to the validatfcn {param1, param2, param3,param4}

In the statefunction, you can then assgn them as below

Ts = param1;

eta_dis = param2;

eta_cha = param3;

E_rated = param4;

Answer 1

[Bay Jay]

You stated 4 parameters.

Matlab expect you to pass the parameters as comma seperated at the last entry to the validatfcn {param1, param2, param3,param4}

In the statefunction, you can then assgn them as below

Ts = param1;

eta_dis = param2;

eta_cha = param3;

E_rated = param4;