Anyone heard of using Digital Twins to predict the future?

CJ_London

Super Moderator
Staff member
Local time
Today, 16:30
Joined
Feb 19, 2013
Messages
17,841
I went to a very interesting presentation yesterday on the subject.

The focus of the talk was on creating a digital twin of Blenheim Palace. The concept is to digitise every aspect of the palace, gardens and other properties, the estate (circa 12500 acres), power consumption, type and quality of building materials, number of visitors - on foot, by car and by coach, all aspects of the weather, the local fauna, insects, wildlife etc. Literally everything.

They have gone back through some 300 years of maintenance records, using AI to convert hand written notes into data, they have sound and motion detection devices around the whole estate tuned to pick up the sights and sounds of insects, bats, rodents etc (not big brother!)

The attempt is to get the level of detail down to what they described as 'the level of a lego brick) - where it came from, when it was used, what treatments it has had (such a lichen removal). Same applies to windows, doors, roofing materials and the like.

With the digital twin, they can now predict how many visitors they will get on any given day - which varies on time of year, public holidays, the weather to name a few. From this they can then determine how many car parking attendants then need, what the cafes needs in the way of supplies, how many guides, etc. This has resulted in labour saving, reduced wastage in the cafe for example.

On the palace, they can now predict when certain areas need maintenance, moving from a reactive state to a proactive state - an example provided was the roofing materials, not really designed to cope with the heavier downpours we now experience - AI identified the most at risk areas based to be addresses first. It also identifies sections of wall which are in imminent danger of losing a stone

On energy consumption, they have reduced consumption by 20-30% by (for example) having AI monitor rooms - if they are not required, the heating is turned down, if the room is empty, the lights are turned off, etc

On the estate, they can identify which practices minimize reduction in local fauna and which increases it - what is right now, may not be right with different weather conditions. They have identified a species of bee, unique to the Blenheim estate. They live around 30m up in ancient (400+years) oak trees in holes created by woodpeckers. They have a solar farm and AI suggested a different way of laying out the panels in a way to reduce ground compaction and increase warmth and light beneath the panels.

For those interested - see this link https://cesium.com/blog/2026/01/15/blenheim-palace-smart-conservation-with-cesium/
 
I wonder about such efforts, because for some house of historical significance it makes sense, but how much energy is required to do this AND how much will the AI-suggested changes cost to implement? If the investment is greater than the return, then the more significant savings would be to never implement the digital twin. Therefore, for special cases this is OK but for the rest of us poor schmucks, this is a two-pass algorithm. But real life doesn't allow two-pass algorithms (a.k.a. do-overs).
 
That question was asked - energy saving are after cost of running AI which is hosted on in-house servers. The cost of their energy management app was £10k to install.

With regards the initial development cost of the various algorithms, yes in the millions, but they are in the process of selling to other historical sites. In terms of their own savings on maintenance, wastage, energy etc, they reckon it will have repaid itself in 6-7 years
 

Users who are viewing this thread

Back
Top Bottom