Can anyone help me find a data set for a non-linear system such as cascades tank or Model of DTS200 Three Tank

I need these datasets for my thesis and I can't find them anywhere.

Answer 1

[Shivansh]

Hi David,

I understand that you are not able to find a dataset which meets your requirement.

One dataset that you can use for modelling a non-linear system, such as cascades tank, is the "TwoTankSystem" dataset available in MATLAB's Control System Toolbox. This dataset represents a non-linear cascaded tank system and can be used for system identification and control purposes.

To load the "TwoTankSystem" dataset in MATLAB, follow these steps:

1. Make sure you have the Control System Toolbox installed.
2. Open MATLAB and navigate to the MATLAB Command Window.
3. Load the "TwoTankSystem" dataset using the following command:
4. load('twotankdata.mat');

This command will load the dataset into your workspace which provides the following variables:

u (input signal): A vector representing the input signal applied to the system.

y (output signal): A vector representing the output signal of the system.

You can also try and simulate a dataset using a mathematical model or by simulation using Matlab or Simulink which represent the behaviour of the cascaded tanks or DTS200 Three Tank system. Following steps can be used for this purpose:

1. Set the desired number of data samples N.

2. Initialize empty arrays to store the input and output data: inputs = [], outputs = [].

3. Define the initial conditions for the system variables (e.g., tank levels, valve positions).

4. For i = 1 to N:

a. Randomly generate an input signal or control action for the system (e.g., valve opening/closing).

b. Apply the input signal to the system and simulate its response for a given time interval.

c. Record the input signal and the corresponding output response.

d. Update the system state variables based on the output response.

e. Store the input and output data in the arrays: inputs[i] = input signal, outputs[i] = output response.

5. Save the input and output data arrays as a dataset.

You can simulate your own dataset using the above steps if the “twotankdata” doesn’t fit your requirement.

[David Vatavu]

Hi Shivansh,

I want to thank you first of all for your advice on my dataset, but I would like to ask you a few more questions about a feedforward neural network that contains B-splines.

I would like you to give me some steps I should follow to create this neural network.

[Shivansh]

Hi David,

If you want a general introduction about the use of B-splines in MATLAB, you can refer to the following documentation https://www.mathworks.com/help/curvefit/constructing-and-working-with-b-form-splines.html.

For the neural network part, you can refer to documentation for the Deep Learning Toolbox here https://www.mathworks.com/help/deeplearning/index.html. You can select from a range of existing functions from the toolbox which suits your usecase.

Hope it helps!