EDITOR ’ S QUESTION
SANJAY KUMAR SAINANI , GLOBAL SVP OF BUSINESS DEVELOPMENT & GLOBAL CTO FOR THE DIGITAL POWER BUSINESS AT HUAWEI TECHNOLOGIES
refabrication , modularisation and
P standardisation have become the trend of the data centre industry . Prefabrication can greatly shorten the TTM of data centres . Modularisation enables data centre flexibility that is on-demand deployment and capacity expansion . Standardisation can ensure high-quality data centre delivery with fast construction . In recent years , with the development of digital and AI technologies , data centres are evolving towards automation in design , production , delivery , operation and O & M .
AI TECHNOLOGIES CAN EFFECTIVELY AVOID THE FAILURE RATE OF THE POWER SUPPLY SYSTEM IN THE DATA CENTRE .
The details are as follows :
1 . Achieve the effect of ‘ What you see is what you get ’. BIM ( Building Information Model ) is used in the design phase to automatically run the collision experiment , identify pipeline interference in advance and improve the design quality . The data shows that change workload can be reduced by 80 % using the BIM model 3D design .
2 . Achieve the effect of ‘ What we design is what you get ’. By means of digital production , i . e . the use of 3D twin models , fine production can be achieved to reach optimal air leakage rate , tightness and dimensional tolerance .
3 . Fast delivery . In the delivery phase , prefabrication and testing of the integrated power supply and distribution , cooling and fire extinguishing systems in the container are completed in the factory based on the prefabricated modular construction mode . Through accurate calculation and analysis , the lego-like construction of the container is completed onsite during the delivery process , achieving fast delivery . Take Asiacell , the largest mobile operator in Iraq , as an example , construction is quick and services can be rolled out within just 10 months .
4 . The DC status is optimal in real time . In data centre operations , the temperature control system of a data centre is used as an example . The temperature control system of a traditional large data centre is complex , including seven subsystems , such as a cooling tower and water chiller . More than 60 parameters need to be adjusted . However , it is difficult to achieve global optimisation based on human technology and experience . The AI algorithm of deep neural network is used to adjust the temperature control system . Global data centre optimisation is implemented , 7x24x365 online in real time , and PUE ( Power Usage Effectiveness ) is optimal in real time . In actual applications , China Unicom ’ s Zhongyuan data base in Henan province uses AI energy efficiency technologies to reduce the PUE from 1.54 to 1.35 , saving more than 8 million kWh annually .
5 . Enable automated driving in the data centre . According to Uptime ’ s latest user report , the failure rate caused by the data centre power supply system accounts for 43 % of the overall failure rate . AI technologies can effectively avoid the failure rate of the power supply system in the data centre . For example , intelligent algorithms can be used to predict the service life of vulnerable components , such as fans and capacitors , and abnormal noises can be collected to determine device running status . In addition , AI inspection robots are used to perform local inspection , collect images , temperature and sound , and use cloudbased analysis functions . www . intelligentdatacentres . com