There is a lot of scattered information about what a digital twin is, and many companies claim they can create one. But it's not always the case. Factory5 Product Director Artem Serebrov reveals the essence of this phenomenon and talks about the possibilities and history of the development of digital twins.
A digital twin is a virtual copy of a physical object or process. It was originally conceived a long time ago as a digital copy of equipment. But since then, a lot of time has passed and the digital twin has many variations. All of them are well worked out theoretically, but they are just beginning to develop in practice.
The present
The digital twin has several types, which are divided according to the scope:
The Digital Twin Prototype is the first digital twin in history. In the 1970s, it was used by Nasa to design the Apollo 13 mission. It was created to test how a future object would behave in the physical world. Later, NASA engineers found that the same twin could be used to monitor existing equipment and predict what would happen to it. However, since space conditions were poorly understood at the time, it was necessary to create constant real-time communication between the equipment and the twin in order to take real readings from the sensors. This led to modern digital twins and their next type.
Digital Twin Instance is a digital twin of an instance. In fact, it is collecting information from sensors installed on the equipment and obtaining a copy of a working unit in the system. This allows you to automatically monitor and predict the behavior of this equipment, predict emergency stops and create simulations of operating modes.
Digital Twin Aggregate is an equipment series model. This is a twin of existing equipment, which is primarily used for personnel training and optimization of maintenance and repair processes. With such a model, a specialist does not need to read manuals for operating the equipment, but can immediately practice and see the results of actions on a virtual model.
Future
Today, new types of doubles are emerging: some of them seem to be fantastic, and some are already being created and implemented.
1. Digital twin of the process
, . , . , .
2.
. .
3.
, , , . , .
, , . , , . , , : , . .
F5
Factory5 , . : , -.
:
,
( ) .
:
.
.
.
.
. . , : — . , , . , : , .
, . , .
. , . «» , , - .
MX- :
1. «-». :
.
, .
.
2. «». :
, .
.
.
, , .
ML-.
ML-.
3. «Data science ». :
.
, , .
.
, , , .
ML-.
ML-.
. , 12%, 30%, 50%. , , , , .
F5
1.
. . , . , , , .
2.
: , , . . - . 45 . 30%, 20% 100%.
3. - -
, . , - . , -. 70%.