Innovate for Sustainability
Baipatra : Sustainable Ideas
Call Now

Editor's choice

Development of indicator based real time pen-stock condition monitoring system by the application of GIS and Cognitive tools

The GIS based condition monitoring systems can be developed to monitor the vital parameters of pen-stock networks which is used to carry water from the source to the power house for generation of energy. The plant efficiency is often compromised due to the reduction of conveyance efficiency of the pen-stock network. The reason for this decrease can be attributed to the diminution in the depth to width ration or increase in roughness coefficient due to sedimentation or increased infiltration due to erosion of the channel bed.

The problem is the decrease in efficiency will be different in different locations of the pen-stock network and this variation should be monitored so that the most vulnerable regions can be identified. As a result, if a GIS based map of the network is developed and connected to the tracking instruments used for monitoring the vital parameters which represent the conveyance efficiency, then a real-time system can be utilized to identify the most vulnerable regions and adequate measures can be taken to mitigate the problem. As all the vital parameters are not equally important in influencing the conveyance efficiency, an indicator based system will be more efficient in depicting the condition of the pen-stock network.

Selection of parameters in these cases can be done with cognitive tools like Neural Networks or Particle Swarm Optimization algorithms etc.

Development of a Spatial Predictive Model for Estimation of Flow Rate in Pen-stock Networks

A neural network model can be developed to predict flow rate at different points of the pen-stock network in a hydro power plant. The value of the parameters in the earlier junction can be used to predict the value of the same parameters in the present junction. A digitized map of the pen-stock network can be developed where each junction and pen-stock length between the junctions are digitized and the data of the vital parameters are stored in the attribute table linked to the shape files. The collected data stored and continuously updated in the attribute table can be used to train a model by the neural network methodology. Such models can be used to know the status of the pen-stocks at different points of the network.