Is That Tree Too Close?

No one at Bonneville Power Administration (BPA) remembers a prior situation where so many employees were mobilized to inspect and document the condition of the entire 15,000 miles (24,000 km) of high-voltage transmission circuits. Previous vegetation-caused outages set the stage for this massive undertaking. During fiscal year 2008, thousands of man-hours and millions of dollars were spent on inspection, documentation and mitigation to ensure the transmission system was clear of subsequent potential vegetation-caused outages.

During this effort, five different inspection techniques were used to assess vegetative conditions within the right-of-way (ROW) corridors of BPA:

Transmission line maintenance (TLM) working patrol field inspection
Helicopter aerial survey with TLM observers
BPA's Vegetation Clearance Categories

Helicopter aerial survey with natural resource specialist (NRS) observers
Contract field ground inspection conducted by a private firm
Light detection and ranging (LiDAR) remote sensing.

Accuracy of Inspection Techniques

In assessing the efficacy of these inspection techniques, two questions were asked: Which vegetative sampling technique was most accurate in determining clearance violations? Which vegetative sampling technique was most cost effective in determining clearance issues?

The emphasis of the field and aerial inspections and the LiDAR remote sensing was to determine if any vegetation fell within three categories of vegetation-to-conductor clearance criteria that defined potential threats to circuits: danger brush, danger tree grow-in and high brush.

The term “brush” in BPA vernacular refers to any type of growth form: tree, shrub, side limbs and so forth. Danger brush is any vegetation located on the transmission line ROW extending into the minimum clearance distance at maximum sag from the conductor: less than 15 ft (4.6 m) for 287 kV to 500-kV lines and less than 10 ft (3 m) for 69-kV to 230-kV lines. A danger tree grow-in is a tree just off the ROW that may present a clearance hazard as its lateral branches or horizontal leaders elongate. High brush is located on the ROW extending into the minimum clearance distance at maximum sag from the conductor: 16 ft to 25 ft (4.9 m to 7.6 m) for 287 kV to 500 kV and 11 ft to 20 ft (3.4 m to 6.1 m) for 69 kV to 230 kV.

Danger brush is the most critical clearance distance because it has the highest potential for flashover. Clearance distance conditions in the high brush and danger tree grow-in categories do not pose an immediate threat to the transmission system. Vegetation at this distance may remain for another growing season followed by subsequent removal.

LiDAR outperformed all other field inspection techniques; it located the most clearance violations. This held true even when accounting for an approximate 12% rate of false positives found in verifying LiDAR reports in the field. The false positives were objects other than vegetation, typically transmission line hardware such as jumpers or other foreign objects such as light poles, abandoned wood poles or birds. What makes the LiDAR method stand out is the number of false negatives found in more subjective, or human-based, inspections. While the LiDAR approach locates false clearance issues, other methods miss real clearance issues.

In the danger brush comparison, the TLM working patrols found the least number of unique clearance issues (data points). The private firm found more unique data points than LiDAR in both the danger brush and high brush categories. BPA speculates the reason for this is because the private firm misclassified the specific data points. LiDAR would accurately report the data point in either category while human error, whether in measurements or data recording, would contribute to misclassifying these sites.

The same scenario is the reason why the TLM working patrol had more unique field data points than LiDAR in the high brush category. It is equally plausible that the TLM working patrols misclassified danger brush as high brush, therefore increasing the unique field data points for LiDAR in the danger brush category while increasing the unique field data points in the high brush category.

Another accuracy-related observation of the comparisons is that the private firm's ground inspection found almost twice as many field data points for danger tree grow-in as the next highest field survey with TLM working patrols. TLM ground patrols found the next highest number of field data points. Both helicopter survey techniques found the least number of observations.

However, an interesting error occurred when 2800 reports of vegetation were incorrectly identified as danger tree grow-in by TLM crews. Clearance distances were actually greater than the specification for this category. Though not a critical mistake, unnecessary time and resources were diverted from working on real critical problems in order to sort through data to differentiate the mistaken reports or, worse, to respond to them in the field. These diversions were costly, unnecessary and increase the risk of missing more critical work.

In the high brush category, LiDAR outperformed the other inspection techniques. The private firm's ground inspection and TLM working patrols found the next highest number of field data points. Helicopter surveys performed relatively poorly in identifying any high brush.

In terms of detecting vegetation-to-conductor clearance issues, there is only a slight gain in accuracy when using NRS versus TLM observers during helicopter aerial inspections. However, other program benefits are arguably important enough for BPA to consider instituting a vegetation-only helicopter aerial inspection with NRS observers as a new work practice into BPA's transmission vegetation management program. NRS observers gain a better understanding of the character of the ROW they manage by participating in an annual helicopter tour of their lines. The aerial perspective they gain imprints spatial relationships in their mind and enhances their memory of the corridors, ROW and circuits entrusted to them in ways that cannot be obtained from the ground-only perspective.

As slightly less than half of the NRS staffs are new hires — with less than one year of experience to become familiar with their districts — the annual aerial surveys can significantly accelerate their learning curve and skills. For this reason alone, it appears this work practice should be incorporated as a permanent part of BPA's transmission vegetation management program, until at least the new staff becomes more familiar with their districts.

Another benefit is that data gained through aerial surveys with NRS observers does not require subsequent ground verification by the NRS observers, thus eliminating one additional step in the process and saving on relative costs.

Relative Cost

In terms of comparative costs, helicopter aerial vegetation surveys proved least expensive at just over US$26/circuit mile, whether an NRS or TLM observer was in the observer's seat. There was a negligible difference in cost between the two variations of the same sampling technique. One should note that from an accuracy standpoint, this was the least effective method of survey. The TLM field surveys cost the next lowest at approximately $141/circuit mile, followed by the private firm's ground patrol at $330/mile and topping out with LiDAR remote sensing, averaging $948/circuit mile for initial establishment and going as low as $257/circuit mile for subsequent fly-overs.

An interesting caveat to the cost of LiDAR should be explained. LiDAR has been trending downward in cost since its initial usage in 2007. The cost effectiveness of LiDAR is presently tied to the number of circuits per corridor. The more circuits in a given corridor, the lower the per-circuit-mile cost. And as corridors are reflown, amortized cost per circuit mile drops incrementally, even when taking into account inflation adjustments.

Furthermore, BPA Aircraft Services has been investigating the potential for performing LiDAR fly-overs with leased hardware that may further reduce the overall cost. It is hypothesized that LiDAR generated from BPA Aircraft Services could be collected for approximately $129/circuit mile and that processing this data would cost approximately $200/circuit mile, for an estimated total cost of approximately $329/circuit mile.

Going Forward

Five different techniques were used to inspect vegetation-related conditions within BPA's ROW. Some techniques were trials that BPA committed to in settlement with its regional regulatory organization, the Western Electric Coordinating Council, for violations of reliability standards from vegetation grow-in-related outages. LiDAR proved most accurate in identifying vegetation-related clearance issues. The most cost-effective but least accurate method was using helicopters with either NRS or TLM personnel serving as aerial observers.

In the future, BPA will acquire new LiDAR data annually so all North American Electric Reliability Corporation sanctionable and BPA critical circuits are eventually modeled in five years. Corridors will be re-flown to update the vegetation-condition LiDAR data on a scheduled cycle. Also, an annual vegetation-only NRS ground inspection of all circuits west of the Cascades will be started.

BPA will continue to train and provide hardware and software to TLM linemen to facilitate reporting of vegetation in the in-house enterprise geospatial information software (eGIS), rather than in TLM applications (another in-house product used exclusively by working patrols), until the new eGIS/TAS corporate database is deployed. Also, iso-clearance line contour maps will be created from LiDAR data for use by ground and aerial inspectors. The reliance of annual aerial helicopter and TLM working patrols of all circuits will trend toward exclusively identifying imminent risk.

Steve Narolski ([email protected]) is program manager for Bonneville Power Administration's vegetation management and access maintenance program. He earned his bachelor's degree from Penn State University and his graduate Continuing Education in Forest Ecology & Silviculture (master of forestry equivalent) from Washington State University. His professional licenses and certificates include a California-registered professional forester, a certified forester of the Society of American Foresters, an International Society of Arboriculture arborist and a Tree Farms of America inspector

Analysis and Assumptions

In comparing and contrasting each field inspection technique, several assumptions were made. For instance, only single circuits were analyzed not double circuits, as the results from LiDAR cannot be compared to ground surveys for double-circuit lines. Clearance distance findings were attributed to the associated circuits in different ways for each of these inspection techniques, making comparisons invalid. Additionally, those circuits in which vegetation maintenance occurred between field review were discounted and removed from the sample population.

When the field data was filtered, a total of 14,321 data points were identified for 325 circuits. There are more than 700 circuits in the system. From this data pool, one-to-one comparisons were run between each inspection technique in all possible iterations to generate unique data points per clearance category and per comparison.

Danger Tree Grow-In Inspection Method Comparison

MethodsShared pointsUnique points for first methodAerial TLM vs. Ground contractor 17 61 Aerial TLM vs. Ground TLM 123 81 Aerial TLM vs. aerial NRS 191 14 Ground contractor vs. Aerial TLM 17 603 Ground contractor vs. Ground TLM 63 372 Ground contractor vs. Aerial NRS 28 602 Ground TLM vs. Aerial TLM 123 321 Ground TLM vs. Ground contractor 63 117 Ground TLM vs. Aerial NRS 93 273 Aerial NRS vs. Aerial TLM 191 84 Aerial NRS vs. Ground contractor 28 114 Aerial NRS vs. Ground TLM 93 138