How To Implement Artificial Neural Network (ANN) in MATLAB/SIMULINK ? (Part-3) | Dr. J. A. Laghari

[@saeidreza6736]

Excellent presentation, covers the main topics in the subject.

[@DrJALAGHARI]

Thank you so much for the compliments.

[@ArifHussain-fs5jz]

THANKS FOR THE WONDERFUL VIDEO AND WAITING FOR YOUR NEXT VIDEO SERIES SPECIFIC FOR POWER SYSTEM IN SIMULINK WITH VARIOUS TEST SYSTEM. IN PROTECTION AND FOR ISLANDING DETECTION

[@DrJALAGHARI]

Thanks for the compliments. Yes sure

[@ridharehouma9087]

sir, what we use if the simulink model work with discrete sampling time when we generate neural network

[@DrJALAGHARI]

Thanks for the comment. You may use the same block it will not have any impact.

[@DrSoniM]

Very nicely explained. Great sir. Thank you so much.

[@DrJALAGHARI]

Thanks for the compliments. I would suggest don't miss the last part. It will be more interesting and helpful in ANN implementation in simulink.

[@navid.abd98]

The best. Thank you very much Mr. Doctor. please show us how to tuning a NN controller in matlab and simulink....

[@DrJALAGHARI]

Thanks for the compliments. Kindly watch all four parts. I have discussed it.

[@Jamesiah08]

Good tutorial. Thanks. Jazakallah Khayran.

[@DrJALAGHARI]

Thanks for the compliments.

[@stephenpeters4393]

Thank you Doctor, Your videos really help students after a lot of theoretical knowledge; these videos give us something to practice with #still_helpful_till_date

[@DrJALAGHARI]

Thank you so much for the compliments. Kindly share it within your circle too so that they may also get benefit from it.

[@ايةقاسم-ي7ب]

Thank you so much dr. Plz dr. I want to know how can i have a data (I-V) from the matlab/ simulink of a pv array under PSC from (g =100 to 1000w/m2) to training it by ANN???

[@DrJALAGHARI]

Thanks for the compliments and query. In this case, you have to simulate all cases at different irradiances and measure the voltage and current. Save each value in the excel sheet or matlab workspace. Then do the training.

[@similoluwaokunowo5127]

Excellent video! Please how can I get training data samples from a SIMULINK model for a nonlinear control system?

[@DrJALAGHARI]

Thanks for the compliments. To get training data from simulimk, kindly watch it's fourth part. Similarly do visit my channel homepage and playlists, there are several videos in which i have implemented the ANN in simulimk model of power system.

[@mohsinalibhutto5489]

Dear sir, which type of ANN you are using here? MLP or SLP?

[@DrJALAGHARI]

Thanks for the comment. I don't remember exactly. But broadly speaking ANN is commonly considered as MLP.

[@panashecollins6388]

thank you Doc if i want to implement ANN for a hydraulic system what is my input and target data. I have pressure ,vibration, f low and temperature values of the system. Do you have videos on how to then use the ANN network to estimate RUL.

[@DrJALAGHARI]

Thanks for the compliments and comment. I am not sure. But first you try to find any research paper related to your work on this topic. Then it will be easy for you to implement it.

[@abroma4512]

Thank you Sir for such a nice conceptual teaching. I would request you please tell how to obtain the regression equation after performing ANN modeling and Matlab, as we see the regression results come satisfactory. At the end, the equation is also needed. Your answer would be appreciated

[@DrJALAGHARI]

Thanks for the compliments. You don't have any need of equations. You just need input and output data for your model only. You will use that data for training and the regression equation curve will automatically be developed for your model by the ANN.

[@sshh287]

Great Videos Dr, Great works indeed. But I want to inform you, nntool (from 5:22 onwards) has been removed from MATLAB, now it's replaced with nnstart

[@DrJALAGHARI]

Thanks for the comment and update. Yes. You are right.

[@erdemsaglam3926]

thank you doctor, appreciated !

[@DrJALAGHARI]

Thanks for the compliments.

[@giancarlomoreiragomes159]

Ótima explicação, muito obrigado

[@DrJALAGHARI]

Thanks for the compliments.

[@sajids7146]

how do you partition the dataset in nntool? For example 70% for training, 15 for validation and 15 for testing? There seems to be no option to change this in nntool? Also how do you use radial basis function? Everytime I choose this option, the training option can't be used? Would appreciate your advice. Thank you.

[@DrJALAGHARI]

Thanks for the comment. May I ask for what topic you are using the ANN. in the nntool, you have to separate the data for training, testing and validation before giving it to nntool.

[@atiyahwaael1955]

Dear Doctor thank you for your lectures , my question ,I do my model and when test training in Train network appear following message (inputs and targets have different number of samples) what your advice to solve that

[@DrJALAGHARI]

Thanks for the comment. This error means your input data or output data has more samples than input or output. Kindly check it in the workspace section whether both are equal or not.

[@abhisekdas7747]

sir in case of closed loop boost converter ,how can we write the syntax for nn instead of pi . plz help me sir.

[@DrJALAGHARI]

Thanks for the comment. Watch the fourth part for this purpose. It is easy and already explained in fourth part.

[@ayushvarun776]

The video was helpful sir. Please tell me how to prepare landslide susceptibility maps using ANN in MATLAB ?

[@DrJALAGHARI]

Thanks for the compliments. Read some papers on this topic to find the input and output data for the landslide susceptibility maps.

[@emanahmed9489]

Hi, when train network, message appears that say inputs and target have different number sampels

[@DrJALAGHARI]

Thanks for the comment. It means your inout and output data is not equal. Try to take equal number of samples.

[@tahirusman1129]

Thanx . Jazak Allah Sir.

[@DrJALAGHARI]

Thanks for the compliments.

[@ankitsinghal9288]

sir thank you so much sir, in my neural network training regression is not coming in popup window during implementation in GUI

[@DrJALAGHARI]

It should come. Is training successfully completed or not? Then you click on regression, it will show the regression curve.

[@saswatipattnaik7681]

Thankyou Sir..it was really helpful.very well explained..kindly upload the video for data generation from simulink

[@DrJALAGHARI]

Thanks for the compliments. I have already uploaded the video. Have a look.

[@ashtonGM2550]

Thank you, i had a question, when the ANN is being trained, it will be finished by happening one of these statuses: maximum epoch reached or maximum gradient reached, which one means our ANN has better performance? In addition, when we open the regression diagram we see value R for 4 diagrams (Training, validation, test and all) which one should be close to 1? the training or all of them? thank you

[@DrJALAGHARI]

Gradient reached is more important. Training curve regression is important.

[@sravankumar1263]

Excellent video sir.thankyou for sharing valuable knowledge sir. how ANN will be used for state estimation sir.please suggest me sir.it will helps me alot for my research.

[@DrJALAGHARI]

Thanks for the compliments. Ann can also be used State estimation. To do this, first simulate the system with mathematical state estimation in order to get data for Ann. Then train the ANN for that data. In the end, incorporate the trained ANN in test system and perform simulations and analyse its performance.

[@sanammirza2425]

Thank you You really helped me in abundance :)

[@DrJALAGHARI]

Thanks for the compliments.

[@alrawdasuliman6649]

Awesome !!

[@DrJALAGHARI]

Thanks for the compliments.

[@mahalakshmig4456]

in GUI tool, while training the ANN i am getting warning message like this "Inputs and targets have different number of samples". can you pls help me in this regard

[@DrJALAGHARI]

Thanks for the comment. It means that number of samples are different. Both input and output must have same number of samples. Go to workspace and double click on the input data and output data and check their total numbers.

[@wondimagegnbekele8915]

please give us tutorial on fault location and classification using ANN...doctor

[@DrJALAGHARI]

Thanks for the comment. Yes Sure. I will upload the video on this topic very soon.

[@krishnayk9]

Sir in GUI method you typed as nntool. what if we typed as nnstart?

[@DrJALAGHARI]

Is there any command nnstart?

[@krishnayk9]

@@DrJALAGHARI I got it sir.nntool and nnstart are same.In matalb updated or future versions nntool will be removed and instead nnstart is used.

[@DrJALAGHARI]

@@krishnayk9 Thanks

[@kanjani1]

Thanks with love

[@DrJALAGHARI]

Most welcome.

[@techzila5379]

sir can you do matlab simulation of neural network for non linear system with taking some research paper?

[@DrJALAGHARI]

Thanks for the comment. Kindly fourth part of this ANN series to understand how ANN can be used in simulink.

[@shraddhagajbhiye4513]

sir, can the data in constant block(x1) shows only one data. How we can use it in simulink for changing data.

[@DrJALAGHARI]

You may use variable function block or slider to change the data.

[@shraddhagajbhiye4513]

sir it is showing error, may i know ur mailid please so that I can send u the error files.

[@deepanjalinaik4750]

Could you explain how to use ANN to forecast wind speed

[@DrJALAGHARI]

First yiu must have the wind speed data of atleast 5 years. Then yiu can use this data in ANN to forecast.

[@engg.w1983]

Thanks professor you are great

[@DrJALAGHARI]

My pleasure.

[@sanammirza2425]

Could you please explain how to use ANN in a solar-based MPPT system?

[@DrJALAGHARI]

Very simple. First watch the fourth part of this series to learn how to get the data for Ann training from the simulink. Now, just follow the steps in your solar model to get the data for Ann training. Rest of the procedure is explained in the video. Kindly go through it. Do subscribe, like and share this channel and its content in your circle. Thanks

[@sanammirza2425]

@@DrJALAGHARI sure, I have already shared videos with my friends

[@do_not_reply4784]

pranaam sir , in my case , the result from PID is much better that ANN. There is an overshoot in ANN at the start. The output values , i kept under 1 , while the input values were even less than 1. can you suggest any research paper on this particular topic so that i can refer it to write a research paper. very good videos though , from 2 days , i am watching your channel full day. Thanks sir

[@DrJALAGHARI]

Thanks for the comment. Basically in this particular case we have maximum value or margin to 1 therefore it takes a lot of effort to improve the response. But in case there is a big difference between the. Output results, then ANN will be surely best. To publish paper kindly watch my videos on LFC and AGC and try to implement the latest techniques.

[@nazerinnazeer5941]

Sir, the class was really nice. I have a doubt that how the healthy data & faulty data of a induction machine is trained in the ANN & also how the selection of output of this two?

[@DrJALAGHARI]

Thanks for the compliments. Watch the fourth part, i hope you will get the idea .

[@nazerinnazeer5941]

@@DrJALAGHARI thank youu sir

[@krishnayk9]

Sir if we given epoch value as 1000 in code means.while training it should reach 1000 epoches compulsory?if it is not reached 1000 and stopped at 670 means.is our ANN training failed?

[@DrJALAGHARI]

1000 epochs means iterations. It is not compulsory . The stop criteria are two. One is goal or target which is by default 10-12. You can reduce it if you want. Second is maximum iterations. Therefore, program or ANN will stop whichever criteria is achieved first. But then we have to check the regression curve if it is close to 1. Then ANN training is successful otherwise we may have to increase the number of epochs or some other steps as discussed in the video.

[@krishnayk9]

@@DrJALAGHARI thank you sir

[@vaishnavivaishnavi6118]

thankyou so much sirrrrr

[@DrJALAGHARI]

Most welcome

[@shahidamin2640]

youre great

[@DrJALAGHARI]

Thanks for the compliments. Do share it within your circle too.

[@furqan1234100]

Once i run the command nntool. The MATLAB gives an error ''nntool' has been removed. How to resolve this ?

[@DrJALAGHARI]

Thanks for the comment. May be you are using the latest version of matlab. I have used matlab version ,2017a.

[@furqan1234100]

@@DrJALAGHARI Thanks.

[@DrJALAGHARI]

@@furqan1234100 you welcome

[@LoyelRosePoovathingal]

Matlab 2022 version shows nntool has been removed. How to start Neural Network Data manager in 2022 version?

[@DrJALAGHARI]

Thanks for the comment. Yes. It is removed in matlab 2022 version. You may now use nnstart command. Try it.

[@hassanafak6540]

hi.prof. how I cantact with you please ?

[@DrJALAGHARI]

You may contact me at my email [e-mail] If I know any thing I will try to help you.

[@boualemsassia8120]

How implement extreme learning machine in simulink

[@DrJALAGHARI]

Extreme Learning machine in your case means ?

[@boualemsassia8120]

@@DrJALAGHARI sir I want implement extreme learning machine algorithm in simulink can you help me

[@NaveenKumar-hy4js]

Tq sir

[@DrJALAGHARI]

Most welcome

[@fahadamyal620]

What is your research field ?

[@DrJALAGHARI]

Thanks for the comment. My research field is electrical power system, islanding detection, operation of DG, load shedding schemes, application of computational intelligence based techniques in power system.

[@fahadamyal620]

@@DrJALAGHARI are you interested in publishing papers regarding control of switched reluctance motors via ANN

[@DrJALAGHARI]

@@fahadamyal620 Thanks for your invitation. At the moment, i am busy in other activities. Therefore, i am unable to accept your invitation.

[@niguseassefa1141]

Hello. would you please send me the reference? I'm eager to learn, the video is exceptional one!

[@DrJALAGHARI]

Thanks for the compliments. I couldn't understand what do you mean by reference?

[@engg.w1983]

🙆‍♂️💯🌷

[@DrJALAGHARI]

Thanks

[@techzila5379]

prof sir can u implement this paper DOI: 10.1109/TII.2019.2959658

[@DrJALAGHARI]

Thanks for the comment. I am busy therefore, it will not be possible for me to do it.

[@techzila5379]

@@DrJALAGHARI 😓