The artificial intelligence that scientists had been working on much earlier was released from laboratories and became a part of life after a computer called Deep Blue was defeated by the world-famous chess champion Garry Kasparov in 1996 and his developed model Deeper Blue defeated Kasparov in 1997. Deeper Blue had an algorithm for calculating all possibilities in a certain game with rules, and a processor that made that calculation much faster than a human could. In this way, he was making the best move out of millions of possibilities. Artificial intelligence, which was faster than humans in numerical processes, used to fail when comprehension is concerned.
The human brain has a comparative decision-making mechanism by using its learning ability and past experiences, and while it can take decisions even with incomplete data, artificial intelligence without learning ability could not take decisions when even any of the data was incomplete. We have become capable of using the software that can learn with forecasting algorithms and data mining applications developed in the last 20 years, add new algorithms by itself and modify existing algorithms. Whether we are aware of it or not, artificial intelligence, which can learn from the personal assistants loaded on our smart phones to internet browsers that predict the products we want to purchase by archiving and analyzing all kinds of data about us and bringing up appropriate advertisements, is everywhere.
We made introduction to the subject of smart building in our article titled Smart Buildings and Automation . The purpose of making buildings intelligent is to ensure that all electromechanical systems in the building can perform their functions independently of humans.
We can compare smart building applications to the Deeper Blue computer. As Deeper Blue did, it manages electromechanical systems according to the results that it finds by calculating with the algorithms loaded on it. Since there is no system that can learn, it continues to use the same parameters and algorithms, even if the conditions change. The use of fixed parameters and algorithms will not be a problem for routine applications, but when it is necessary to decide according to the changing conditions, the operator will have to take this decision and change the system parameters. This applies to all security and automation systems used in smart buildings.
We can make the issue more understandable through energy efficiency. Energy Saving? Energy Efficiency? We stated in our article titled above that energy efficiency is not a static application that is completed by applying it once, but a process, and the process continues in the measurement, evaluation and improvement cycle as long as the building survives. Although the system performs the measurement process in smart buildings, it is carried out by the operator using the evaluation system and the improvement decisions to be taken are also taken by the operator. If the operator does not perform the evaluation and improvement work at all or performs incompletely, the system will continue to use the existing fixed parameters and algorithms. This means increasing energy consumption and the worsening energy efficiency process.
In smart buildings, artificial intelligence technology is used on a system basis such as cameras and software with video analysis, fire detectors that decide the alarm level according to the environment, thermostats using fuzzy-logic and control devices. Also, if we put through the example of energy efficiency, a control device with fuzzy-logic feature measures and evaluates values such as ambient temperature, humidity and air quality, makes a demand forecast at the end of this evaluation, unlike other control devices, and makes improvements in the control parameters of the equipment such as the valve and damper motor connected to it. Since it constantly updates the demand forecast by using the parameters it records, it will make more accurate estimates as retrospective records increase. In this way, it will provide the same comfort by consuming less energy than a conventional control device at the end of a few days of the learning process.
When an artificial intelligence building is concerned, it is mentioned that systems or system components manage the entire building as well as the use of artificial intelligence. This artificial intelligence will, of course, be Integrated Building Management Systems abbreviated as EBYS or EBKS. We evaluated the integrated building management system applications system applications in our article titled You Decide How To Manage Your Facility. The use of fixed parameters and algorithms in classical smart building EBYS applications can be evaluated as in sub-systems. While classical EBYS software is sufficient for routine applications, the parameters will need to be changed by the operator when it is necessary to decide according to the changing conditions. An artificial intelligence supported EBYS application will make predictions by using the data collected and recorded from all sub-systems and will change the control algorithms and parameters of the sub-systems according to these predictions.
As we mentioned at the beginning of our article, learning artificial intelligence is based on data and the more data it obtains from the devices and systems, the more accurate algorithms and parameters it will create. An EBYS with artificial intelligence will take its place in the buildings of the future, along with the Internet of Things (IoT), smart mobile device applications and building information modeling (BIM).
