Infrastructure Analytics is the term we use at Zensor to refer to the products we provide for the monitoring of Civil Infrastructure. This could be bridges, tunnels, lock doors, or even entire airports. The aim of this follow-up is to ensure that the structure can serve us as long as possible, in a safe way, and at the lowest possible maintenance cost. In order to achieve this we need to ‘listen’ and ‘feel’ what the asset is doing. The goal: detect problems in the earliest possible stage. We of course do not do that manually, by visually inspecting it, or putting a stethoscope on it. The solution to achieve this, the basis of our products, is a fully digital and automated one, always consisting of the same building blocks:
- Data collection through sensors installed on the asset
- Data management and storage
- Data analysis to detect flaws and make predictions
- Interfacing with the clients by sending warnings and alarms and by providing context through on-line dashboards.
The 3 last stages take place on our cloud-based platform. Here the real heavy lifting is done by the Analytics engine. This component relies on a broad set of various algorithms that continuously treat the data coming from the different sensors. Included are (for example):
- Classical threshold watching: are components too hot, did they sink below a permitted value, or did a crack grow outside the range allowed?
- Trend detection: Do we see the distance between 2 objects increasing, while it has been stationary for 2 years? Do we see the moisture content of concrete increasing after 5 years of stability? Does the vibration amplitude on the bridge joint increase with time?
- Correlation analysis: Is there no more correlation between the water level in the reservoir and the amount of rainfall? Do we see that the inclination measured on the wall is no longer correlated to the ambient temperature or ground water level? Is the deformation under load changing as a function of time?
- Usage insights: Do we see a drastic change in number of passages, or in the weight of traffic passing by? Was there a collision on the pillar last night?
The algorithms are parametrized for each specific piece of infrastructure, to ensure a super tight follow-up. Parameters are set in the beginning, but of course get optimized once the asset ages and its ‘normal’ behavior is known better. As such we can safely say that the follow-up is getting smarter with time.
The heart of all the follow-up is thus the analytics used to track the health of the infrastructure: Infrastructure Analytics, or in digital terms: #InfraLytics . It’s nice to have your own Hashtag once in a while, isn’t it? 😉