Predicting the Resistance of Power Cables to Flame Propagation by Neural Networks (PSAM-0069)


In this work the prediction of the resistance of power cables to flame propagation is addressed. This is a very important safety issue in several critical installments, such as the nuclear power plants. In this respect, two issues are of fundamental safety importance: i) cables that supply electric power to safety systems must work also in case of fire; ii) cables must not provide a means of propagation of the fire throughout the nuclear power plant. To verify this second issue, large scale tests are performed according to different international standards. These tests, done in suitable laboratories with controlled atmosphere, are destructive of the cable and very expensive.

For these reasons, there is a growing interest in investigating the feasibility of developing models for predicting the fire resistance of cables. The models for the prediction of the fire resistance properties are important not only in the design phase of new cables but also for the condition monitoring of the installed wire systems.

In this work, the results of a study on the feasibility of developing a model for predicting the test outcome of the international standard test DEC 60332-1 are reported. Two independent neural network models have been developed for the prediction of the test outcome (pass/fail) and of the charring height reached at the end of the test The results of the two models have been appropriately combined to obtain a reliable prediction of the outcome of the test The aggregate model has been verified using real data, with no misclassifications of the test outcome. The model also gives a “don't know” indication for cables whose characteristics fall outside the range of operation of the trained neural networks.

The proposed modeling approach is suitable to be used for assisting the design of new types of cables, allowing a reduction of the number of tests to be performed in the preliminary design phases.

  • Abstract
  • 1. Introduction
  • 2. Flame Propagation Resistance Tests
  • 3. Multi-Layered, Feed-Forward Neural Networks
  • 4. Definition of the Prediction Model
  • 5. Conclusions
  • References

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In