WEBVTT

1
00:00:01.000 --> 00:00:05.004
- Without electricity, almost every important activity

2
00:00:05.004 --> 00:00:08.001
would come to a grinding halt.

3
00:00:08.001 --> 00:00:10.008
The goal of the entire power generation,

4
00:00:10.008 --> 00:00:13.006
transmission and distribution system is

5
00:00:13.006 --> 00:00:18.003
to provide more reliable power at lower prices.

6
00:00:18.003 --> 00:00:22.005
The challenge in doing so comes from the massive complexity

7
00:00:22.005 --> 00:00:28.001
and dynamism at every step of the giant power networks.

8
00:00:28.001 --> 00:00:31.006
Given climate concerns, increasing demand

9
00:00:31.006 --> 00:00:34.000
for renewable energy has become

10
00:00:34.000 --> 00:00:37.007
one more disruptor in this industry.

11
00:00:37.007 --> 00:00:42.003
AI technology is now becoming an important tool

12
00:00:42.003 --> 00:00:45.003
in these sectors' arsenal.

13
00:00:45.003 --> 00:00:47.003
Take power generation.

14
00:00:47.003 --> 00:00:50.004
Renewables, such as solar and wind,

15
00:00:50.004 --> 00:00:53.009
are becoming important sources of power.

16
00:00:53.009 --> 00:00:57.000
In order to ensure reliable supply,

17
00:00:57.000 --> 00:01:02.001
utilities need to predict how much output they can expect

18
00:01:02.001 --> 00:01:05.004
from the ease renewable sources.

19
00:01:05.004 --> 00:01:09.004
AI algorithms can combine enormous amounts

20
00:01:09.004 --> 00:01:12.005
of meteorological data with the capacity

21
00:01:12.005 --> 00:01:15.008
of the solar and wind farms to predict

22
00:01:15.008 --> 00:01:21.003
the likely output several hours in advance.

23
00:01:21.003 --> 00:01:26.009
AI is also helping utilities to optimize power generation,

24
00:01:26.009 --> 00:01:30.005
even in seemingly stable contexts.

25
00:01:30.005 --> 00:01:35.004
Take Norway, which relies heavily on hydro energy.

26
00:01:35.004 --> 00:01:39.005
Despite the abundance of water in Norway,

27
00:01:39.005 --> 00:01:42.007
only a small portion of it is available

28
00:01:42.007 --> 00:01:45.001
for electricity generation.

29
00:01:45.001 --> 00:01:50.002
Algorithms help utilities, such as outdoor energy,

30
00:01:50.002 --> 00:01:54.006
optimize the use of water.

31
00:01:54.006 --> 00:01:59.009
AI is also being deployed for fall prediction in equipment.

32
00:01:59.009 --> 00:02:03.005
IoT sensors embedded in every piece

33
00:02:03.005 --> 00:02:08.009
of equipment generate a constant stream of data.

34
00:02:08.009 --> 00:02:12.000
Algorithms analyze this data to predict

35
00:02:12.000 --> 00:02:15.001
the risk of failure in advance

36
00:02:15.001 --> 00:02:20.006
and how any failure could cascade through the entire system.

37
00:02:20.006 --> 00:02:24.000
For example, geothermal power plants

38
00:02:24.000 --> 00:02:28.001
in Japan utilize predictive diagnostics

39
00:02:28.001 --> 00:02:31.005
to automatically spray chemical agents

40
00:02:31.005 --> 00:02:35.002
to avoid turbine shutdowns.

41
00:02:35.002 --> 00:02:41.001
In addition to IoT sensors, drones are also starting to play

42
00:02:41.001 --> 00:02:45.007
a role in monitoring electricity transmission systems.

43
00:02:45.007 --> 00:02:50.001
UK's national grid utilizes drones equipped

44
00:02:50.001 --> 00:02:54.001
with high-resolution still and infrared cameras

45
00:02:54.001 --> 00:02:58.002
to take images of wires and pylons.

46
00:02:58.002 --> 00:03:01.006
Computer vision algorithms analyze these images

47
00:03:01.006 --> 00:03:05.005
to predict the need for preventive maintenance ahead

48
00:03:05.005 --> 00:03:08.002
of potential breakdowns.

49
00:03:08.002 --> 00:03:10.003
The advantage of drones lies

50
00:03:10.003 --> 00:03:13.006
in their ability to cover vast areas

51
00:03:13.006 --> 00:03:18.007
and difficult terrain quickly and inexpensively.

52
00:03:18.007 --> 00:03:21.008
Drone imagery also plays a crucial role

53
00:03:21.008 --> 00:03:27.004
in faster recovery from disasters, such as hurricanes.

54
00:03:27.004 --> 00:03:30.004
Transmission and distribution networks need

55
00:03:30.004 --> 00:03:33.009
to be optimized on a real-time basis.

56
00:03:33.009 --> 00:03:38.008
There's huge variation in customer demand minute by minute.

57
00:03:38.008 --> 00:03:41.007
Smart meters track these variations

58
00:03:41.007 --> 00:03:44.000
right down to each customer.

59
00:03:44.000 --> 00:03:47.009
There's also huge variation in electricity generation

60
00:03:47.009 --> 00:03:52.004
from different sources, including solar and wind farms.

61
00:03:52.004 --> 00:03:56.009
AI algorithms help by optimizing the routing

62
00:03:56.009 --> 00:04:00.008
of electricity from varying supply sources

63
00:04:00.008 --> 00:04:03.009
to varying customer demand.

64
00:04:03.009 --> 00:04:07.003
On the customer side too, changes are afoot.

65
00:04:07.003 --> 00:04:11.008
Rooftop solar decentralizes electricity production.

66
00:04:11.008 --> 00:04:15.004
Customers with rooftop solar panels

67
00:04:15.004 --> 00:04:18.006
can also sell unused electricity back

68
00:04:18.006 --> 00:04:22.001
to the grid, then there's a growing need

69
00:04:22.001 --> 00:04:25.002
to charge electric cars and buses.

70
00:04:25.002 --> 00:04:28.001
The distribution grid must predict all

71
00:04:28.001 --> 00:04:31.009
of these variables on an ongoing basis.

72
00:04:31.009 --> 00:04:34.005
Customers don't care about the variability.

73
00:04:34.005 --> 00:04:36.009
All they want is stable supply

74
00:04:36.009 --> 00:04:40.000
of power when they need it, where they need it

75
00:04:40.000 --> 00:04:43.004
and at as low a price as possible.

76
00:04:43.004 --> 00:04:48.001
AI algorithms help distribution grids analyze this mass

77
00:04:48.001 --> 00:04:52.001
of data and ensure that the network continues

78
00:04:52.001 --> 00:04:56.007
to match supply and demand in an optimal manner.

79
00:04:56.007 --> 00:05:01.005
Summing up, the role of AI in the power sector is

80
00:05:01.005 --> 00:05:06.002
to help optimize an extremely complex network

81
00:05:06.002 --> 00:05:11.006
with ongoing variability at every node and every link.

82
00:05:11.006 --> 00:05:15.004
This technology is becoming a major contributor

83
00:05:15.004 --> 00:05:19.006
to the stability and sustainability of power supply.