Benchmarking in Action: Comparing the Costs of HVDC

by Steven J. Morris, UMS Group Inc.

The recent proliferation of utility-scale renewable energy projects, driven by legislative mandates, has created a need to transfer large amounts of electricity from distant rural areas, often in the middle of the continent, to urban load centers generally on the coasts. The long distances involved have led to a resurgence in interest in high-voltage direct current (HVDC) overhead transmission, a technology little used in North America in the last 30-40 years.

A handful of 400 kV or bigger HVDC overhead transmission lines were built in North America in the 1970s and ’80s, These included Nelson River, Quebec-New England, Pacific DC Intertie, Path 27, and the CU Line. While HVDC has been used in a number of undersea cable projects on the continent, since 1986 no new HVDC overhead line has been energized in North America.

That will soon change.

Over a dozen new projects are under development or construction across North America. Most are intended to transport hydro power or wind power from the North to the South or the middle of the continent to the coasts. The large amount of capacity needed and the long distance to be travelled make HVDC a viable option for these projects.

In addition to the myriad challenges involved in building any large-scale, high-voltage transmission project, HVDC brings additional challenges as there is little recent experience in building or operating these systems and few companies have HVDC lines. Therefore, there is a desire by companies building these systems to understand costs (and validate EPC estimates) of both constructing and operating these systems in order to ensure they are prudent and reasonable.

For one company developing a 500kV HVDC project in North America, the answer to better comprehension of what it would take to build and run the project was to use benchmarking to understand comparative costs. This is an approach which others would be advised to consider as it can serve to validate to the board that costs are reasonable, prove to regulators that costs are prudent, and assist asset management in understanding the likely O&M costs.

A limited number of 400 kV+ HVDC overhead line projects have been implemented in North America, although a number of projects are planned. Given the lack of recent North American systems and relatively small number of HVDC projects implemented, comparative data for a benchmarking exercise will have to be sourced globally, as most of the existing 400-600 kV+ HVDC overhead line projects have been built outside of North America.

Using data from other countries to benchmark brings a host of challenges which includes normalizing the data for currency fluctuations and wage rate differences, not to mention inflation for different time periods. In addition, there are significant differences in regulatory and environmental regimes around the globe which can impact costs.

Benchmarking Construction Costs

For lines specifically, there are also a number of factors which must be accounted for to ensure costs are comparable. The projects being compared will vary in terms of capacity and higher capacity lines translate into higher costs. Using linear regression, a formula can be developed by which capacity of all projects can be adjusted to an equivalent basis versus the average. By applying the formula to the difference in capacity from the average, an adjustment factor can be calculated. This factor can then applied to the project’s cost per line mile to determine an equivalent cost per mile (see Figure 1).

Figure 1: Example of Adjustments to Overhead Line Costs

Different tower types can also be used for projects. As guyed towers are less expensive than freestanding towers, the projects being compared must be adjusted to factor in the average cost difference between them to put the towers on an equivalent cost-per-mile basis. There are other adjustments which can be made to the comparative cost of overhead lines such as return type, foundation type, conductor type, etc. However, comparative data on these can be hard to gather and the value may not be worth the additional cost and effort.

Benchmarking lines is always an easier task than stations (regardless of AC or DC) because of the relative simplicity of transmission lines vs substations. As discussed previously, transmission lines mainly differ in just a few dimensions (tower type, foundation type, conductor type/size/capacity, etc). Substations are all unique with different numbers, types and voltage levels of equipment. Therefore, accurate benchmarking requires decomposing the substations to be compared into equivalent values (i.e., transformer capacity) and counts for major equipment.

For HVDC projects, this decomposition is impossible to achieve. There are only three vendors supplying converter stations. Price is heavily driven by the competitive pressure on the companies to procure projects at that specific time. Therefore, the same project procured in a different period might be more or less competitively bid. In addition, virtually all HVDC converter stations are procured through turn-key contracts with strict confidentiality clauses. This inhibits the ability to break station costs down into their constituent components hindering analysis of cost drivers. However, there are some normalization factors that can be used to assess stations. This includes assessing on a cost per MW basis, as well as adjusting to equivalent capacity (as with lines) (see figure 2).

Figure 2: Scatter Chart of Converter Station Cost vs. Capacity

From an overall project perspective, the benchmarking effort should look at location-specific drivers of cost differences. The most common of these is wage rates which can be used to normalize labor costs. However, a look at a functional cost breakdown (i.e., permitting, ROW acquisition) can also be used to identify not only where local differences are driving costs, but also where internal efficiency (e.g., project management, engineering) exists.

Benchmarking Operations and Maintenance Costs

The above discussion has dealt with benchmarking construction costs, but a company building a new HVDC project also needs to understand what it’s going to cost to operate and maintain the system. Typically, the converter station vendor will provide a recommended maintenance schedule for the DC yard, but there are a number of operating factors which impact maintenance that will not be known until operating experience occurs. These include power transfer levels, operating scheme, utilization rate, etc.

In addition, companies maintain their converter stations differently. Some maintain them remotely, while others maintain them locally. Some have dedicated DC staff, while others have shared staff with their AC stations. Some have 24/7 on-site personnel, while others only run one shift. The physical size of the facility and amount/type of the equipment in the station also impact the amount of maintenance required.

Another factor driving O&M costs is that not all projects face the same reliability requirements. Commercial projects typically don’t face strict system operator requirements for reliability, different regulators have different requirements for inspection and maintenance, and non-North American projects don’t have to meet NERC CIPS requirements. These factors all impact the amount of maintenance required and must all be taken into consideration when benchmarking staffing levels and O&M costs (see Figure 3).

Figure 3: Equivalent Staffing Levels for Converter Stations

Outside of North America, it is common for utilities to contract out maintenance of their HVDC converter stations. As many have only one or two converter stations, they do not see the point in staffing and training a small group just for HVDC. However, in North America, most major maintenance is performed in house. Regardless, availability of contract resources can impact staffing levels. Stations that are located in areas where the skills needed for HVDC are simply not available or in areas with heavy demand from other industries may not be able to take advantage of contractors, even if they wished to.

The number of maintenance outages taken also differs by company and can drive total O&M costs. Depending on the degree of redundancy in critical systems and the loading scheduling, some utilities may take outages biennially. However, most utilities with HVDC typically take scheduled maintenance outages once or twice a year. These outages can last from several days to several weeks, depending on the complexity of the tasks to be completed and are a key cost driver.

Newer stations have a high degree of remote, self-diagnostic capabilities, requiring less on-site monitoring. Remote operations by the control center allow for use of shared resources versus on-site operations which require dedicated staff. However, on-site operators are generally also able to perform minor maintenance, so there may be a cost trade-off which must be factored into to comparisons.

Finally, companies building HVDC projects, particularly those without existing HVDC assets, will face a learning curve on maintenance. During the first couple of years of operation there will likely be increased demand for O&M field personnel (e.g., support initial equipment troubles, capital loading for project deficiency corrections and warranty, etc.) that will decrease over time, supporting the release of dedicated technical resources and increased sharing of resources.

Benchmarking can serve a useful purpose for companies developing overhead HVDC projects. However, careful consideration must be given to those exogenous factors which drive cost differences to ensure that an apples-to-apples comparison is made.


Steven Morris is a Principal of UMS Group Inc. and its client sponsored benchmarking and best practice study leader. He has 20 years of utility industry experience and has assisted numerous utilities in benchmarking generation, transmission, distribution, and corporate services functions. Reach him at

Previous articleUtilities on the Front Lines of Environmental Stewardship
Next articleThe Future of Demand Response: The Practitioner’s View

No posts to display