Heat map displaying pipe section burst risk profile
Twitter Marker Cluster showing where the public have tweeted possible leakage
Twitter Marker Cluster (expanded) showing highlighted leakage ranger
Augmented Reality View of the pipe network highlighting various monitored attributes
The main themes that the application is focusing on are:
- Identification of emerging leaks and locations
- Crowdsourcing customer data
Risk of Burst Failure
There are three main categories that were taken into account, in order to calculate the risk of a burst:
- Historic Time Series on burst data
- Pipe Material
- Pipe Age
However, there are several ways that this model can be improved. This can be done by taking into account other factors, such as:
- Soil Data
- Geographical Altitude
- Pot holes
The risk of burst failure across all DMA’s is displayed using a heat map based on pipe data. This allows us to show not just the DMA that is affected but also the actual pipe section that is at risk.
Using Twitter to Crowdsource Leakage Data
The system also explores crowdsourcing leakage data via twitter. Tweets to a specific account are displayed on a map based on the geographic location of the tweet. We then use clusters (meaning that all tweets in a similar location are grouped together) to identify the priority of the leakage. The more leaks reported in the same location the higher the priority. This is reflected by the cluster changing colour, going from green to red dependent on the number of tweets. We also use the hashtag '#leakageranger', this identified that a tweet has been added by one of NWG’s Leakage Rangers. These tweets will always take precedence, therefore any clusters that have a tweet with this hashtag will automatically turn red and be classified as a priority. This enable us to ensure that we don't miss major leakages that are within rural areas for example.
Augmented Reality Engineer Tool Kit
We can leverage upcoming technologies to both log extra data to be analysed using image processing and cloud point data analysis. Including the potential to:
- Automatically detect corrosion on exposed pipes.
- Detect ground level or environment shifting.
- Create rich ground composition data.
Combining this Lab in a box data with GIS pipe data, and our predictive models engineers could have a Augmented Reality view of the pipe network and where possible leaks will occur.