As the Woodsense sensors come out and cover major buildings in Denmark, Sweden and Norway, we get access to more data that we can use to improve our anomaly detection algorithms.
Our intelligent moisture meter is built around an algorithm that through weather data on precipitation, humidity, air pressure, dew point and wind direction and speed can assess the amount of moisture in the wood.
In order to catch moisture and irregularities, we combine our data with weather data, which enables us to catch critical situations.
The sensors have been able to capture many different fluctuations in the measurements and notify of fluctuations through an intelligent moisture alarm - here is a selection of the many varied fluctuations the sensors have captured.
The above example shows monitoring of proper drying of CLT elements after closure over a period of four months from December 2020 to April 2021.
During the period, some of the areas did not fall to the acceptable level of 10 % and therefore had to be re-checked.
Many wooden modules spend a long time in storage and it is therefore important for manufacturers and suppliers to ensure that the wooden modules are not exposed to poor conditions.
The above example shows monitoring of a wooden module that has been stored by a manufacturer. The measurements for the other monitored wooden modules did not change and therefore a red flag was raised for the above measurements.
In the measurements you can see how temperature and humidity suddenly changes around March 11, when there were extreme weather conditions with storms and rain.
Furthermore, since subsequent drying of the wooden module did not occur immediately afterwards, it is assumed that the plastic wrapping covering the wooden module has been broken.
Whether wood modules are stored at the manufacturer or on site, it is a good idea to carry out moisture monitoring to keep an eye on the amount of moisture in the wood elements.
As the above example shows, the sensors are able to track moisture and irregularities by collecting and applying weather data on the surrounding environment, including moisture content and precipitation, as well as humidity and temperatures.
As the following example shows, moisture monitoring can be used with great benefit after the execution phase; there are great advantages to be gained by performing moisture monitoring during the operating phase of the building in both basement rooms and risk areas.
The above example shows moisture monitoring, where the sensors are located in a basement. In the measurements, there is not a drastic increase, but since the sensor is placed in a basement, it was suspicious that the gradual increase occurred and did not dry out afterwards.
When the customer got the alarm and checked the basement, they found that road workers had drilled through the sidewalk and down to the basement. As a result, rainwater just as quietly began to fill up the basement.
This customer set up some of the sensors after they had a huge leak, so they were happy to avoid another water problem.
From day one, the goal of the Woodsense sensor solution has been to have smart algorithms that automatically detect anomalies and take into account the context of the sensor location.
Before we can get there, we put in effort to manually inspect data so that we can better train our algorithms. Therefore, it is more fitting to call it Human Assisted AI so far, rather than fully automated anomaly detection. We are developing intensely on our data detection and we look forward to being able to write more about it in the future. Follow us on this page or sign up for our newsletter.
.jpg)
