Adaptive MPC Block Error

23 vues (au cours des 30 derniers jours)
LB_SD
LB_SD le 16 Déc 2020
Commenté : Majid Mazouchi le 17 Juil 2023
Question:
How can I correct the error message that shows when I run the model?
  • Error evaluating 'InitFcn' callback of Adaptive MPC block (mask) 'test22/Adaptive MPC Controller'. Callback string is 'ampcblock_InitFcn'
  • Caused by: For compatibility with the adaptive mode, the plant model specified in your controller object must be LTI state-space (OK), have the same sampling time as the controller (Violated), and be delay-free (OK). The "ss", "c2d", "d2d" and "absorbDelay" functions perform the necessary conversions.
What I Have Done:
1) I have tested the model update function independently and it works great.
2) I have tested the dynamic plant using the normal MPC block and it worked well.
Problem
Once I connect the Adaptive MPC block, the error shows up. Please see model and functions used below.
Model Update Function:
function [A,B,C,D,U,Y,X,DX] = fcn(M,x,u)
% Sample time
Ts = 0.02;
% Model parameters
kd=0.015;
R=0.05;
Tg=0.2;
% Continuous-time model
Ac= [0 1 0 0; -(kd/(M*Tg)+1/(R*M*Tg)) -(kd/M+1/Tg) 0 0; 0 0 0 0; 0 0 0 0];
Bc= [0 -1/(M*Tg) 0 0]';
Cc = [1 0 0 0; 0 1 0 0];
Dc = zeros(2,1);
% Generate discrete-time model
nx = size(Ac,1);
nu = size(Bc,2);
Md = expm([[Ac Bc]*Ts; zeros(nu,nx+nu)]);
A = Md(1:nx,1:nx);
B = Md(1:nx,nx+1:nx+nu);
C = Cc;
D = Dc;
% Nominal conditions for discrete-time plant
X = x;
U = u;
Y = C*x + D*u;
DX = A*x+B*u-x;
Dynamic Plant Function:
function xdot = fcn(M,x,u)
% Model parameters
kd=0.015;
R=0.05;
Tg=0.2;
% Continuous-time model
A= [0 1 0 0; -(kd/(M*Tg)+1/(R*M*Tg)) -(kd/M+1/Tg) 0 0; 0 0 0 0; 0 0 0 0];
B= [0 -1/(M*Tg) 0 0]';
C = [1 0 0 0; 0 1 0 0];
D = zeros(2,1);
xdot = A*x + B*u;
MPC Object
MPC object
---------------------------------------------
Sampling time: 0.02 (seconds)
Prediction Horizon: 10
Control Horizon: 1
Plant Model:
--------------
1 manipulated variable(s) -->| 4 states |
| |--> 2 measured output(s)
0 measured disturbance(s) -->| 1 inputs |
| |--> 0 unmeasured output(s)
0 unmeasured disturbance(s) -->| 2 outputs |
--------------
Disturbance and Noise Models:
Output disturbance model: default (type "getoutdist(mpcVSG4by4)" for details)
Measurement noise model: default (unity gain after scaling)
Weights:
ManipulatedVariables: 0
ManipulatedVariablesRate: 0.0595
OutputVariables: [5.9452 0.0595]
ECR: 100000
State Estimation: Default Kalman Filter (type "getEstimator(mpcVSG4by4)" for details)
Constraints:
-22 <= delP <= 22, -10 <= delP/rate <= 10, -311.02 <= State-Space(1) <= 345.58
-1 <= State-Space(2) <= 1
  1 commentaire
Mozhgan Sabz
Mozhgan Sabz le 19 Sep 2022
I have the same problem. did you solve your problem?
would you please let me know?

Connectez-vous pour commenter.

Réponses (1)

dekun wang
dekun wang le 24 Mai 2021
Hi there,
My guess is that you design the mpc controller using a continuous plant model but Adaptive MPC controller requires a discrete plant model.
You can easily transfer the plant model of your mpc controller into discrete in these commands:
cp=mpcobj.Model.Plant;
dp=c2d(cp,Ts);
mpcobj.Model.Plant=dp;
  1 commentaire
Majid Mazouchi
Majid Mazouchi le 17 Juil 2023
Thanks! It works!

Connectez-vous pour commenter.

Produits


Version

R2019b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by