The dynamic, bustling city of Fukuoka, Japan, relies on a complex network of pipes, ducts and cables to carry its power, water, communications and gas--the city's lifeblood. Increasingly, many of these utilities are buried underground.

Accurate mapping of this network is essential for planning and maintenance. Much of this work falls to NTT Infrastructure Network Corp. (NTT Infranet), part of the NTT Group, which manages the communications infrastructure in Japan. In fact, NTT Infranet has become so specialized in accurately surveying underground facilities that it not only serves NTT Group but also subcontracts its services to other utilities and municipalities within Japan. The company is viewed as the leader in utility infrastructure management.

Retaining this leadership position requires a constant reevaluation of data collection processes and a willingness to challenge conventional practices. Several years ago, NTT Infranet began investigating the use of model-based methods for underground 3D utility mapping. Could a system comparing data from multiple sensors against a physical model of the magnetic field expected from a utility line provide the confidence and error data the company was seeking? It was an intriguing proposition, and the team was enthusiastic about the prospects.

Standard pipe and cable locating equipment determines the position of the pipeline by detecting the emitted magnetic field from a current-carrying conductor. By finding the horizontal position of the signal peak at ground level, field workers can invoke a depth function to compute the vertical elevation from ground level to the underground utility, and a GNSS system can be used to determine the geospatial horizontal position. However, “ghost” or “bleedover” currents frequently contaminate these measurements because of inductive or capacitive coupling between the targeted utility and other cables or pipelines lying in the same utility right-of-way.

NTT Infranet had developed correction tables to map depth to truth in areas of field distortion, but this solution offered varying levels of success and site-dependent results. Mapping congested duct banks proved particularly challenging. Although the company tried several different mapping systems, all of them had poor 3D accuracy, and none offered an associated statement of RMS error or confidence bounds on the location information. Without these details, the quality of the results was sometimes ambiguous. Yet confidence bounds can only be derived when there are a larger number of measurements than unknowns to be estimated in the system. The two measurements gathered by traditional equipment (usually a top and bottom magnetic field sensor) to estimate two unknowns, such as the depth of a cable and amount of AC current carried by the conductor, were insufficient.

“There were three major problems for locating by traditional technologies,” says Toyokazu Fukui, director of development at NTT Infranet. “First, utility mapping work was generally limited to single-point location only, and it missed the complexity of underground utility direction changes (horizontal or vertical), which are essential for infrastructure management. Capturing points in sufficient detail to detect the 3D aspect of the infrastructure was also costly due to the increase of location points. Second, we could not locate accurately with conventional electromagnetic pipe and cable locating tools because of signal bleedover to nearby utilities. And third, it took much time and effort to create the map from data. With conventional locating equipment, we measured each located point, recorded the GPS data, and moved the data to the office. Map creation then required plenty of manpower. We continually sought a better method.”

In August 2009, the company heard a presentation on a model-based technology called the Spar from Optimal Ranging Inc. Using two 3D magnetic field sensors, a triaxial accelerometer and digital compass, the system could identify the offset, depth, current and yaw angle to the underground utility regardless of its position in the radiated field. The system only needed to be in the approximate vicinity to calculate a position with corresponding confidence bounds. The technology’s ability to jointly estimate the key physical parameters of any co-linear bleedover sources in the same underground area allowed it to decouple these bleedover currents from the corrected target cable position. By combining the system with GNSS, automatic geopositioning could occur even from the side of the actual utility.

According to the technology developer, the system would work for any underground utility that had metallic content or, in the case of a plastic or concrete pipe, a tracer wire. Sonde transmitters (point sources) could also be traced in 3D, allowing the targeted utility line to be nonconducting tunnels and conduits. The system operated at a low frequency (491 Hz was standard, although frequencies between 22 Hz and 10 kHz were also supported) and was therefore unaffected by soil types or conductivity.

Beyond the convenience offered by the continuous measurement of line offset and depth from an arbitrary point in the field, the method promised to resolve field distortion. “We were intrigued by the ability of the FieldSens technology to decouple the effects of bleedover from the measured signal by analyzing the magnetic field and removing interfering sources,” Fukui says.

Seven months later, after a significant amount of research and testing, NTT Infranet purchased five prototype units, which it integrated with its existing dual-frequency GNSS technology and a tablet computer into an innovative cart-based system that it calls the Utility Exploror. The company almost immediately began seeing a return on its investment. “We discovered that the technology achieves an excellent and accurate measurement even in bleed-over conditions and provides an improved target utility location,” Fukui explains.

Two modes of operation provide flexibility in how the system is used. In the “precise” mode, field workers perform a series of transects, or walkovers, across a cable or pipeline at specified intervals. At each of these points, the system measures a cross-section of the magnetic field to determine a single offset and depth of the cable or pipeline at the location of the walkover. In the “vector” mode, which more closely corresponds to traditional locating methods, workers follow the cable and thus gather significantly more measurement points than with the precise mode. However, the vector mode is unable to decouple currents from bleedover sources. For this reason, the precise mode is used to occasionally confirm the results obtained when moving down the cable or pipeline in vector mode.

In all applications, the system’s software (known commercially as FieldSens View) merges the measured 3D offset to the utility with the 3D geospatial position, which allows an automatic, real-time position of the utility to be recorded on the map in the local coordinate system, the Japanese Geodetic Datam 2000. Confidence bounds are 95 percent. “By our calculation, the Exploror is 11 times more efficient in processing time from locating to mapping and saves 80 percent in cost compared to conventional technology,” Fukui says.

The new technology is generating acclaim for NTT Infranet. The team was recently recognized for their pioneering work on the collection of underground utility information at an NTT Group product innovation fair. The technology has also given the team the opportunity to present at international conferences hosted by the Geospatial Information and Technology Association (GITA) and other organizations.

More importantly, the technology has enabled NTT Infranet to expand its business beyond communications infrastructure to managing other underground utilities. For example, when the municipality of Fukuoka needed to determine the precise position of a 2-meter-diamater steel water main, NTT Infranet was awarded the project because of its ability to satisfy the requirements. “A very effective and highly reliable locating method was needed to meet the limited contractual terms for 20-kilometer water main pipe,” Fukui explains. “The client’s map was very old, and they wanted to generate a new map with an accurate pipeline location. All the data had to be linked with GNSS to manage the client’s GIS. With the FieldSens software and the Utility Exploror, we were able to provide that solution.” The company is also expanding into the power and gas industries.

“The complete integration of the FieldSens software with survey-grade GNSS, including all metrics used to qualify data accuracy and confidence, makes a big difference in our operations,” Fukui says. “The technology computes the utility’s 3D location in local coordinates, supporting easy transfer into the drawing and GIS system without losing the link to key measurement quality metrics. This system is opening more business opportunities for NTT.”