• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 83 subscribers
  • Views 684 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMDSCNCD28P55X: About arc detection setup at Arc Fault Detection using Embedded AI model

Seungwoo Nam
Prodigy 60 points

Part Number: TMDSCNCD28P55X
Other Parts Discussed in Thread: TIDA-010955,

Tool/software:

Hi, I was looking through this device's design guide and application in arc fault detection using embedded AI models.

This device is composed with TIDA-010955 and control card C2000.

I was wondering if I can get information about arc detection test conditions which is shown in 'Arc Fault Detection using Embedded AI models' at 7.4.

Such as, 

  1. Type of experiment done to simulate arc fault (ex. twist, short, mechanically separating electrodes like in UL1699B ...etc)
  2. Type of line (ex. copper, tungsten, ... etc)
  3. DC link Voltage
  4. Amount of Current
  5. Used Inverter type and specifications (ex. switching frequency, input voltage, output voltage, rated power, ... etc)

Thank you again for your kindness.

  • 0
    Prodigy 60 points

    Part Number: TMDSCNCD28P55X

    Tool/software:

    Hi, I was looking through this device's design guide and application in arc fault detection using embedded AI models.

    This device is composed with TIDA-010955 and control card C2000.

    I've noticed that the example data given for the MATLAB pre-processing & for uploading the data for model training in model composer are different. They can be found on 'Application Note' at 5.2 and 6.2.

    As long as i know, data for MATLAB pre-processing should be used for model training in model composer. If not, please tell me which part I'm getting things wrong.

    Therefore, by utilizing MATLAB pre-processed data won't work in model composer for model training but data given for only model training works well.

    So please tell me the right procedure from data capture to model training at model composer.

    Thank you so much again.

  • 0 in reply to Seungwoo Nam
    TI__Genius 11985 points

    There is another thread on this and a response is posted there. It shows this has been resolved. So I am closing this one

  • +1 in reply to Seungwoo Nam
    TI__Expert 5145 points

    Hello Seungawoo

    Section 6.2 gives an example of data set that is ready to be used for model training. You can use this data to train the model and see model performance on the same data, just for the learning purpose. However, the trained model using the given dataset cannot be used to detect the arc fault on another setup. The model needs to be trained for each setup. So the first steps are

    1. Build the setup. Please refer to the UL1699B standard to build the setup

    2. collect the data by generating arc. The data may contain both arc and non-arc current 

    3. use the MATLAB scrip to separate the arc and non-arc data

    4. train the model

    5. integrate the model to CCS project and deploy.

    Thank you

    Amir Hussain