Thermal and LiDAR

Why this matters

UAT 361 covered the basics of thermal imaging. This course extends it to energy systems, which need several sensors together. Thermal cameras find hot joints, LiDAR measures tree-to-conductor distances more precisely than photographs, and UV cameras see corona that eyes and ordinary cameras cannot. Each sensor has limits, and a wrong choice misses defects or raises false alarms.

Thermal cameras in energy work

Thermal cameras have far fewer pixels than colour cameras. A small hot spot such as a line joint must cover several pixels for a reliable temperature, because a pixel that covers both the joint and the sky averages to a value below the true one. The team must set the required pixel size and compute the distance as in UAT 361 module 4. NETD gives the sensitivity to temperature differences, while emissivity and reflected temperature must suit the material. Shiny conductor metal has low emissivity and can reflect the cold sky strongly, so it is easy to misread (Vollmer and Möllmann).

LiDAR

LiDAR fires laser pulses and measures their round-trip time. The range is $R=ct/2$ because the light travels both ways. A drone system sweeps the beam across a swath $w$ as it flies, so the pulse density on the ground is roughly the pulse repetition frequency (PRF) divided by the area swept per second.

The USGS Lidar Base Specification defines quality levels for LiDAR data, for example QL2 with at least 2 pulses/m² and QL1 with at least 8 pulses/m². These were written for wide-area surveys but are a useful reference when stating data quality.

import math

C = 299_792_458                     # m/s
PRF, FOV_DEG = 240_000, 70

for height, speed in ((60, 6.0), (100, 12.0)):
    t_ns = 2 * height / C * 1e9
    swath = 2 * height * math.tan(math.radians(FOV_DEG / 2))
    density = PRF / (speed * swath)
    level = "QL1" if density >= 8 else ("QL2" if density >= 2 else "below QL2")
    print(f"H {height} m, v {speed} m/s: echo {t_ns:.0f} ns, swath {swath:.1f} m, "
          f"{density:.0f} pulses/m2 ({level})")

H 60 m, v 6.0 m/s: echo 400 ns, swath 84.0 m, 476 pulses/m2 (QL1)
H 100 m, v 12.0 m/s: echo 667 ns, swath 140.0 m, 143 pulses/m2 (QL1)

Tree clearance from conductors

Trees growing close to a line cause faults and outages. The North American standard NERC FAC-003-5 sets a minimum vegetation clearance distance (MVCD) by voltage and altitude, an example of a clearly defined criterion. In Thailand the criteria of the utility operating the line apply. A LiDAR point cloud makes it possible to compute the distance from every tree top to the conductor.

import math

def wire_z(x):                      # conductor height (m) at distance x from the first pole
    return 20 - 4 * (1 - ((x - 100) / 100) ** 2)

wire = [(x / 10, 0.0, wire_z(x / 10)) for x in range(0, 2001)]
trees = {"T1": (40, 3, 14.5), "T2": (95, -2, 15.0), "T3": (150, 4, 12.0), "T4": (170, -6, 17.5)}
LIMIT = 3.0
for name, top in trees.items():
    d = min(math.dist(top, p) for p in wire)
    print(f"{name}: {d:.2f} m {'-> trim' if d < LIMIT else ''}")

T1: 4.20 m
T2: 2.24 m -> trim
T3: 6.40 m
T4: 6.02 m

UV cameras for corona

Corona discharge is partial discharge around damaged or dirty high-voltage conductors or insulators, and it emits ultraviolet light. Studies by Wang and Qian (2019) and Wang and colleagues (2014) used solar-blind UV cameras working below about 280 nm, where sunlight is filtered out by the atmosphere, so corona can be seen even in daylight. A UV camera shows where discharge occurs, but judging its severity needs specialists and other data.

Module lab

Common mistakes

Summary

• Hot spots must cover several pixels, and emissivity must suit the material
• LiDAR measures R = c·t/2; density is about PRF/(v·w) and can be compared with USGS quality levels
• Tree-to-conductor distance must be computed in 3D against the line owner’s criterion
• Solar-blind UV cameras see corona in daylight

Check your understanding

1. A LiDAR measures a round-trip time of 500 ns. What is the range?
2. At 80 m with a 70° field of view, how wide is the swath?
3. PRF 100,000 pulses/s at 10 m/s with a 100 m swath. What is the density, and which level does it meet?
4. Why is the tallest tree not necessarily closest to the conductor?
5. Why can solar-blind UV cameras be used in daylight?