Waterworks infrastructure presents a unique problem, in that it is often embedded underground and hard to diagnose problems when it is experiencing problems. Currently, waterwork companies such as Unityworks have a lengthy process in order to identify issues with pumps and other infrastructure, often sending multiple drivers to a site repeatedly in order to finally resolve the problem.
Streamline aims to enable preventative measures by using an API to pull data from the City of Moreton Bays road defect database and Unitywater's infrastructure database. By correlating this data we can give infrastructure around the road a risk factor enabling Unitywater to work directly with the City of Moreton Bay in order to preemptively solve the issue before it causes traffic delays and a negative impact, additionally, the API is able to help Unitywater decrease cost spent on vehicle activities and reduce overall driving.
Planned vs Reactive Jobs
Information from Unitywater showed that 70% of their jobs were planned maintenance (no time critical) and 30% were Reactive (Critical to operations). Diving into the data and using user centered design principles we contacted Ratna (an employee at Unitywater) who revealed through anecdotal evidence that field service staff would often have to make multiple trips to a site to first diagnose an issue then attempt a fix, often returning if the issue wasn’t resolved by resetting the pump. This was seen to be supported in the extrapolated data with two jobs requiring multiple visits in a month to resolve, from this we can draw the assumption that it is most likely happening with other vehicles in the Unitywater fleet. This raised the question of how do we reduce round trip distance for a given job.
Reducing Round Trip Distance
Our first idea to reduce round trip distance was to find a way to remotely diagnose pump issue, thereby reducing how many times a field service staff would have to be sent out, if a problem could be diagnosed without sending someone out to look at the pump, then we could provide the right parts and equipment to the right staff in order to effectively solve the problem. We investigated using visual data such as local CCTV camera, drones or satellite imagery to detect physical problems or utilising telemetry data in order to diagnose a problem pre-emptively. After lengthy consideration we discontinued this idea as most of the equipment is locked inside of a cabinet and the pump hardware/software didn’t support remote functions to reset it.
Back to the drawing board
As we couldn’t remotely diagnose problems we had to come up with a new solution, we spoke to Will about how we might be able to design an early detection system. From this we developed two initial ideas, the first being to scan and use grass hues to identify where leaks are and perform maintenance, the second idea was to combine road quality data with Unitywaters pipeline data and determine which were more at risk of breakage. Identifying grass hues proved to be quite difficult as there were many factors such as weather events, atmosphere and more that may lead to false positives wasting more time then it was saving.
By eliminating the other potential solution we then focussed on seeing if we could correlate data from the City of Moreton Bay and Unitywater. We began the process by pulling the Unitywater infrastructure data in order to verify if piping ran close to road networks, importing this data confirmed that pipelines did cross road networks and therefore may cause traffic be at risk of damage should the road degrade. Next, we pulled the AI Road Defect data from the City of Moreton Bay’s data hub and overlayed it with the Unitywater pipeline data this revealed a proof of concept where road defects were prevalent along a road with a pipeline running under it.