OGC Underground Infrastructure Concept Study Engineering Report

For any hope of having an impact on how infrastructure information is captured, structured, stored, shared and used, it is essential to have an understanding of the various organizations that play important roles in owning, managing, and regulating underground assets.

Who is responsible for the data ?

The parties responsible for collecting and curating data about the underground environment can be grouped into some general categories. There is inevitably some overlap between the categories.

The status of the parties is variable they can be central or local government bodies, quasi-government bodies or commercial companies. Their drivers are inevitably different and generally reflect their role or status.

Asset Operator

Asset operators include bodies responsible for the supply of services such as water supply and disposal, electricity, gas, heating (for example steam) and telecommunications. Additionally, there are parties who manage transport networks that will have data about their infrastructure such as tunnels, stations, ventilation shafts, access points and so on as well as the roads and rail lines themselves. There are also organizations responsible for environmental management or protection, such as surface water drainage, flood prevention, public open space management and so on who will have data about underground assets to assist them in their activities.

Examples from the RFI include US Highway Authorities with 3D inventories and the Dubai government department responsible for electricity. Of note in the RFI, with the exception of Dubai, individual asset operators were not represented directly as participants.

Data supplier

Data suppliers in the underground environment as a purely commercial activity are not a common element. The costs of data capture and maintenance mean that the data entry costs are high if no specific users have been identified. Bodies such as the British Geological Survey (BGS) and BRGM (Bureau de Recherches Géologiques et Minières – French Geological Survey) collect and freely distribute relatively low resolution sub-surface geological data, though they also provide some higher-resolution data on a fee basis. Organizations such as Ordnance Survey collect surface topographic or land-base data which is used to register and depict underground assets.

These organizations are also commonly data collators and will also work on project-based activities using their expertise and knowledge to deliver more precise content where demand exists (see below).

Data collator

Data collators combine data from different sources for conflation as a potential commodity or as a ‘public’ service.

Parties may combine data, often from disparate sources, to create content data that has more value, for example assembling borehole data, contaminated land, mining records, and surface data to create a model that can be owned. In these cases the data may be used to provide a service such as liability to subsidence for land and property. Responsibility for this combined data is likely to lie with the collator. As part of the RFI Columbia University, BRGM, and BGS outlined this type of approach.

In other cases the collator will not own the data but combine it to offer a service. For example the KLIP service in Flanders and KLIC service in the Netherlands combine and supply data to users to reduce conflicts when excavations are planned. In these collation services responsibility for the data still tends to sit with the asset owners. Such services are necessarily provided to the collator with caveats related to quality. Other services such as Dig Safe in New England and Call Before you Dig in Connecticut provide in essence a reverse collation service, by collating and distributing individual excavation notices to the utilities so they can respond without contributing their data.

Data collation services were quite prominent in the RFI responses, for example KLIP, CityGML – Rotterdam, ASK (BGS) and NUAG (Les Guest).

Another type of data collator would be public bodies charged with emergency responses such as those highlighted by University of Munich and New York City. In these examples the conflated data would need to be tested against the identified use cases but is not likely to be shared outside the government body, though findings and recommendations from analysis may be.

Project based

Project based actors are bodies that have a requirement to collect information for a specific project. They tend to capture data to a very high level of detail and quality in accordance with the CI/ASCE 38-02 Standard Guidelines for the Collection and Depiction of Existing Subsurface Utility Data, an engineering standard that raises the utility investigation activity to a professional effort akin to a geotechnical investigation or property boundary land survey, in which a licensed professional certifies the submittal. This costs of capture are typically small compared to the budget for a construction project, and the return on investment is typically on the order of 2 to 10 times the cost. For example UMS Berenice highlights recent light rail projects at the Los Angeles International Airport and in Honolulu.

At a different level, the City of Boston require conduits for broadband fiber to be installed and design / as-built data to be submitted whenever new street work projects take place. Another type of project based capture was highlighted in the submission from New York where commercial companies gather borehole data as part of actual and potential development activities.

It is likely that the ownership of the detailed data will lie with the construction project. However, in the case of the New York geo-technical boring data, it was reported that the companies who capture the data will tend to view it as valuable for use in future projects as it provides information not available to rival companies.

In comparison to mapping the above ground an obvious difference arises. In most cases the asset owner is responsible for the data about their asset and it many cases it is the only record. For above ground there are commonly bodies that collect at least a framework of shareable content, for example mapping or city authorities. If this approach was continued above ground it would be akin expecting the individual property owners, the highway authority, the river authority and so on to all capture data about their area of interest and then share it hoping to make a seamless dataset.

Land Administration

An additional type of responsibility is for registration of ownership or other rights, highlighted by the Singapore Land Authority and Erik Stubkjær. In this case a body, likely to be public, may be recording the location of the ownership or other rights below ground level. There are related uses where some local authorities require knowledge of basements and similar structures for planning, taxation and risk assessment. The correlation with other data may be variable. Land Administration systems also record easements, the right to cross or otherwise use someone’s land for underground utility purposes.

Singapore Land Authority for example report that landowners hold title to the land surface and to 30 m of depth from the Singapore Height Datum below that; everything deeper than 30 m from the SHD belongs to the State. In addition the government can acquire additional underground strata when developing public projects.

What makes a party own the data – running their business, legislation, best practice, benefits

Asset operators

Asset operators originally captured data on paper records, as described by Les Guest for UK and by Munich University for the City of Rotterdam. The driver was for the operator to be able to maintain and extend their services by knowing what assets they had and where they were located. Paper maps do not readily lend themselves to three dimensional representation and anyone combining the data from different operators in crowded areas would struggle with simply overlaying the maps physically. As data became increasingly digital then paper records were digitized to benefit from reduced costs of storage and longevity issues with paper records. This allowed data to be readily overlain but the source data quality limitations generally remain.

From RFI and experience in the UK it would seem that the main drivers for ‘business as usual’ data ownership is to capture only enough information to allow the asset owner to conduct their normal activity. Evidence for data improvements driven by the asset owners was not widespread from the RFI though the benefits were recognized. EPRI reported https://www.epri.com/#/pages/product/1024303/ that GIS data for underground assets is commonly not updated as changes occur in the real world as the team undertaking the work rarely consider themselves to be a data owner. EPRI report only 46% of respondents report significant benefits from high quality data, similarly only 15% reporting repercussions of poor data suggesting a perceived lack of benefit. Additional requirements may be created by those with a more holistic view where the main aim of is to facilitate data sharing and reduce accidental strikes; for example KLIP in Flanders, KLIC in Netherland, ASK in Glasgow, and so on.

Local standards do exist, for example in Switzerland for water (Robin Dainton) and Streetworks legislation in UK. However is most cases the operator is not required to enhance existing data due to the costs of capture/improvement. From the RFI responses there is little evidence of utility providers actively seeking to improve their data for internal operational reasons such as efficiency or best practice beyond that required by statute or contract.

Data supplier

Data suppliers in the underground environment, as evidenced by the RFI responses, tend to capture, collate, and distribute sub-surface geological data at a relatively low resolution. A proposal to periodically capture and monitor the NYC sub-surface was mentioned at the workshop; however, the costs of capture, if not driven by project, particular risk or asset owner, are likely to discourage enhanced or repeated capture. The national geological bodies are likely to have a mandate to maintain a national level of coverage, however the rates of change at this resolution are almost nil. The costs of capture at locally detailed levels are likely to be too high for wholesale capture to take place. This contrasts with the increasing availability of high-resolution surface topographic reference data.

Data collator

Data collators will conflate data from different sources to offer a service. For example Columbia University, BRGM, and BGS combine disparate subsurface sources into data that has value. This can be on a wholly commercial basis as a product, including services, that is sold or as part of some type of ‘public task’. For example data such as contaminated land or mining records may have to be made available.

For strike avoidance initiatives the collation can be voluntary and funded by the utility community; leading to distributed notification systems such as Dig Safe in New England and Call Before you Dig in Connecticut. Alternatively they can be mandated in legislation like the KLIP and KLIC programs in The Netherland and Flanders or the permit system implemented in Dubai and proposed for Chicago. They can also be implemented through contracts whereby a requirement is for an as-built survey to be provided once work in complete.

Collation for emergency responses is likely to be led by the Government bodies responsible for identifying risks, planning and responding. How the data is sourced did not emerge from the RFI.

Project based

Project based actors are bodies that have a requirement to collect information for a specific project. The data is required to plan and monitor a project. It is expected that the data belongs to the commissioning party and that is usually handed over to the operating body when a project is complete. Informal discussions in the UK would suggest that initial the high quality of data is not always maintained as changes occur after the initial completion of the project.

In other cases, such as those identified in New York, core borehole data typically gathered for one project is regarded as a valuable asset. This data is held by the companies that have collected it as it is costly to source and is expected to be reusable on future projects in such a dense urban area.

Land Registration

Land registration bodies are there to "Protect property ownership rights" as reported by Singapore Land Authority in addition to supporting decision making and planning. Singapore Land Authority described how two-dimensional plans have been converted to three-dimensional data.

Data is likely to created and shared as ownership is claimed or as projects are delivered that require ownership or other stakeholders to be established. This could be new construction where private ownership wishes to obtain security of tenure or driven by an infrastructure project in public ownership.

How do the stakeholders relate- licensee/licensor, competitor, govt, supplier?

Asset operators

The asset operators may be in direct competition with one another, for example rival telecommunications operators. Asset operators may also be in competition for space in the real world to place their infrastructure. This may create some reluctance to share data, this was hard to identify from the RFI due to the lack of responses from utility operators. Government can have a role to play is it is likely to license operators to allow them to use public thoroughfares to route assets and in return demand they make their data available to other stakeholders. The KLIP and KLIC initiatives are good examples of this where the aim is to avoid utility strikes by mandating data sharing. There are similar but different versions of this approach around the world, for example in the UK the focus is on minimizing traffic disruption by mandating permits to excavate the pubic highway. Typically permits to dig are volunteered or mandated that alert utilities and traffic management bodies to possible activity and allow them to either supply data, to visit the ground to mark up assets or (ideally) to combine collocated works projects.

The asset owners are typically private companies and the costs of data capture and improvement are significant. In the UK regulation is fragmented across sectors and there is little appetite to impose additional costs that would be passed onto consumers either by sector or across all the asset operators.

The asset operators will supply data to the data collators either through self-interest or to meet statutory requirements. They may also consume data from these collators as part of their day to day activity to plan activity better and to reduce the chance of strikes.

If ownership of the sub-surface becomes an issue then bodies will engage with Land Administration bodies. To a certain extent the identification of who 'owns' a street is a light version of this.

Data supplier

Data suppliers in the underground environment, as evidenced by the RFI responses, tend to supply data at a level they have either been mandated to do by government and/or at a level where they can realize some return. They can also respond to specific projects, government or private sector led. There appears to be little demand from asset operators for soil/geology data; however the soil/geology community appear to be more pro-active in seeking out the asset owners for collaborations, for example the ASK project in Glasgow, Project Iceberg in the UK, and Columbia University in New York.

Data collator

Data collators will necessarily interact with asset owners, data suppliers to source the data required. The supply of data may be required in a contract, a statutory requirement or supplied out of self-interest (to reduce strikes or help with emergency response). In many cases the asset owners will also be customers of the data collator, for example strike avoidance services.

Project based

Project based actors will have a more ad hoc relationship with the other types of party. They may source data from them with a view to assessing and if necessary improving it or they may choose to recapture it entirely to a new standard, for example the Los Angeles airport World Airport Automated People Mover project described by UMS Berenice.

Land Registration

Land registration bodies are likely to interact with Asset owners when they want to create initial records of ownership and stakeholders and will offer a service to all of the other parties when their activities

Data is likely to created and shared as ownership is claimed or as projects are delivered that require ownership or other stakeholders to be established. This could be new construction where private ownership wishes to obtain security of tenure or driven by an infrastructure project in public ownership.

In the vast majority of jurisdictions, utilities are completely buried underground and must be excavated in order to be serviced; only in rare instances have large, easily accessible underground vaults been built to contain infrastructure networks. In almost any urban or suburban street in every developed country there may be five, six, seven or even more different types of utility networks providing services including water, sewer, gas, district heating, electric power, telecommunications and transit.

There are many work processes that depend upon prompt access to accurate and current underground infrastructure information. Sub-optimal access often leads to significant inefficiencies and outright risks to health and safety. The following is a review of major use cases that interact with and depend upon infrastructure information. In each accompanying case study it is demonstrated that major benefits can result when utilities adopt data and operating practices that facilitate effective data integration and rapid access. A key observation is that data organized according to standardized, geospatially-enabled data models enables one version of information from multiple utilities and utility networks to be interoperable, i.e. to be exchanged and combined readily to address a variety of both expected and unexpected uses and applications.

The above use cases and case studies illustrate how improved and interoperable underground infrastructure data can have significant impacts in six major facets of urban life. Representatives from Flanders, Belgium to the OGC Underground Infrastructure Workshop, held in New York City in April, presented a particularly compelling case study of how Flanders is meeting its underground utility data challenges with an established and successful standards-based data integration program.

Flanders, Belgium
In 2004 a large gas transmission line in Ghislenghien was damaged due to excavation activities. The pipeline began to leak gas which led to an explosion that killed 24 people and injured more than 120, many of them badly burned.

Following this accident, Flanders initiated a single-point-of-access system that required the 300 utilities operating in the region to share their data whenever a request for an excavation was submitted. From 2011, Flanders took it a step further and started the “KLIP Digital” project to mandate that all utilities create accurate digital drawings and associated attribute data in conformance with a common model (IMKL, an information model on cables and pipes), which is an extension of the INSPIRE utility network data standard. INSPIRE data models were designed so that all utilities could register their utility information to a common set of data layers. Data from the different utilities could then interoperate and be exchanged and combined seamlessly. The IMKL extension of INSPIRE made interoperability practical, by refining the model to meet the regional requirements of the Flemish utility sector and excavation businesses.

Five years later, in January 2016, Flanders implemented the digital phase of the KLIP system that largely automates the process of bringing utility data together to support excavation operations. A central utility information clearinghouse receives excavation requests, identifies the utilities active in the area and sends them the outline of an area excerpted from their basemap. Utilities excerpt the appropriate portion of their infrastructure map and send it back to the clearinghouse which brings the information together and sends it to the requestor of the excavation. Access to the KLIP system and its Web API is tightly secured and authenticated to protect both information privacy and integrity. As noted above, significant time and financial savings have been realized so far, and the number of insurance claims has been substantially reduced.

The path followed by Flanders represents an effective response to the need to support infrastructure related operations with interoperable information that is as accurate and complete as possible from the outset and that will have a key long-term positive effect on data quality. Jurisdictions around the world are likely to need to go down a similar path, in order to reduce excavation damages and inefficiencies. The most significant way to support these initiatives is to develop interoperable UGI data models that support a range of infrastructure work processes. OGC is perhaps in the best position as an organization to promote development of common data model standards that can save individual jurisdictions the expense and pitfalls of developing their own specific / proprietary models. Such an effort would start similarly to KLIP with the INSPIRE models and harmonize with other models currently in use or under consideration. OGC can also begin to address security and legal issues that have in the past resulted in a reluctance on the part of utilities to share information. Furthermore, OGC can gather cost benefit information that will help jurisdictions build compelling business cases that justify data improvement efforts through significant positive returns on investment.

This section has shown that the implementation of a number of critical use cases hinge on the availability of high-quality underground infrastructure data that can be shared and integrated across stakeholders, organizations, and technologies. When such data is coupled with methods for rapid access and integration it allows data about the numerous utility lines sharing the same streets and districts to be brought together for a variety of operations support and analytic purposes. When such data is not available, jurisdictions pay a high price in construction delays, costly repairs due to accidental utility strikes, expensive change orders, and poorly informed responses to emergencies and disasters.

The Flanders KLIP program offers a particularly compelling case study where standardized digital infrastructure data is made fully available from all utilities for rapid access and integration, with a focus on reducing the time and costs associated with street excavations. The Flanders KLIP program was motivated by a disastrous explosion caused by an excavation error that damaged a gas transmission pipeline, but the benefits of the resulting improvements go far beyond accident avoidance. Results from Flanders so far suggest significant time and cost savings. While in-depth cost benefit analyses are needed to fully confirm improvements, taking the evidence we now have from Flanders and from other jurisdictions and service providers, we can surmise that there is a strong likelihood that underground utility data development will result in significant and even profound benefits.

Following sections of this report explore some of the options for standardized and interoperable data models of underground infrastructure and underground environments, as well as discuss how data integration standards can make it possible to model how entire cities and regions operate, thereby expanding the tool set available to planners, engineers, architects and developers. They also cover various emerging methods of data capture and remote sensing that are likely to aid underground infrastructure data development in ways that can improve quality and reduce costs.

Perfecting interoperable data models will not by itself solve the problems associated with managing underground assets. Additional report sections identify some of the challenges to a supportive policy and legal environment that could motivate utilities, both private and public, to transform their current data into standardized data, and to create new data where there are currently gaps. Security methods and procedures will also need to be developed and put into place to assure utilities that their data is protected from theft and misuse which has so constrained data sharing in the past. Means must be found to exploit opportunities that existing practices provide to collect data from excavations and to assemble information locked away in isolated silos, such as data from previously collected core samples. To ensure that these steps are taken, strong, comprehensive, and financially reasonable business cases will be needed to convince key decision makers that, in the long run, benefits will significantly exceed costs.

Underground utilities form the backbone to support city operations in the form of networks of water, electricity, gas, sewage, telecommunication, etc. Unfortunately, there exists no widely accepted international standard for an underground utility data model. This section summarizes 5 standards for underground utilities as shared in the workshop. Some of these standards have been defined and applied in practices for more than a decade, and some are at the finalization stage. These standards also focus on different aspects in utilities, be it data quality or network hierarchy, etc. The summarization aims to identify the coverage of existing standards, consolidate common datasets/features and attributes, as well as differentiate the application areas among them.

Data models represent, record, and share information about underground structures and utility networks. Inadequate and uncertain information about location and depth of underground utilities and underground structures is a major cause of damage during excavation, construction, and emergency operation. As mentioned by Philip J. Meis during the workshop, on January 10, 1996, a routine capital improvement project caused damage to an electrical cable at Newark International Airport, resulting in more than $1 billion of impacts, including hundreds of canceled and re-routed flights, disruption of travel to tens of thousands of people, and complete closure of the airport for more than 24 hours. National and local authorities and organizations worldwide are developing data models for management of underground information; however, approaches differ significantly. The RFI response from Sisi Zlatanova and Ben Gorte (Delft University of Technology, Netherlands) listed important matters to be considered when designing data models.

The following models have been developed for and/or applied to representing data about underground infrastructure for utilities and other subsurface features. Note that these models apply directly to data about UGI entities being represented. Other models and standards cover the metadata describing data provenance and quality, or the surveys and observations from which the entity data are derived.

Tatjana Kutzner and Thomas H. Klobe (from Technical University of Munich) presented their comparison on existence of characteristics relevant to network modeling in various data models. They concluded that CityGML Utility Network ADE best meets the requirements for modeling utility network characteristics. They also pointed out that the aim of the CityGML Utility Network ADE is not to replace other models or systems, but to provide a common basis for the integration of the diverse models in order to facilitate joint analyses and visualization tasks, e.g., integration of data from IFC or ArcGIS models by means of mapping to/from the ADE.

Underground utilities are often classified into different types of networks based on the service the utility provides but still share common attributes, both geospatial and non-geospatial. Common networks represented in the CityGML Utility Network ADE and INSPIRE Utility Networks models include:

Attributes that are common across datasets modeling both utility networks and associated protection elements include:

The accurate detection, identification, verification, and location of utility assets have always been difficult tasks, and also subject to interpretation and inaccuracies. Not having accurate or sufficient information will increase the risk of: safety of workers and public, abortive and unnecessary work, damage to third party assets, and inefficient design solutions, which in the end increase the social cost.

The present ASCE 38-02 standard defines utility data Quality Levels (QLs) and the corresponding means and methods to be used by engineers in order to investigate and depict utility information in design plans.

Based on ASCE 38-02, PAS 128 also proposes 4 survey category types for different accuracy grades

The INSPIRE Utility Networks model specifies its own 13 data quality elements, such as completeness in commission, positional accuracy, etc. The purpose of the data quality information is to:

There is another category of underground information that is necessary to complete the picture of what is going on beneath the street surface. Insights into this category of information were provided by Columbia University in NYC and by the Geological Survey Divisions of France and Great Britain. Termed the Underground Environment (UGE) it includes everything beneath the surface of the sidewalk or roadway that is not an active utility line. The underground environment is the medium through which utility networks pass and upon which all infrastructure above and below ground sits. The nature of that medium can have profound effects on the utilities it hosts. Components of the underground environment include:

There are also physical characteristics of the underground environment that interact in subtle ways with utility networks and other infrastructure components. These include electromagnetic field strength, temperature, fluid pressure, oxidation state, pH, stress, and vibration.

The underground environment is also a challenge to represent in data because it is a continuous and continuously varying medium in contrast to the discrete nature of typical infrastructure components. Geology, hydrology, and engineering disciplines confront this challenge in a variety of application-specific ways. Decisions about how best to identify significant features and characteristics of the underground environment, as well as represent them geometrically in space and time, will be important steps in developing any comprehensive data-driven applications for UGI.

Models for the underground environment should accommodate information about the location and characteristics of the following elements: .. Earth materials . Soils . Sediments . Fill . Bedrock . Groundwater . Other materials

It is clear that the collection, representation, analysis, and depiction of underground environment data can vary widely depending on perspective and application. While it is unlikely that one model will be able to accommodate every need, there is a clear case for the capability of translating information between models and applications by defining clear mappings and transformation processes based on common concepts.

A large concern about underground data standards is the cost of creating the data. Various new and existing technologies and techniques can help lower data development costs as well as improve the coverage, accuracy, and currency of the data.

A summary of several technologies was provided in the RFI response by Accenture.

A UK initiative, Mapping the Underworld, and the US SHRP2 program were highlighted by Columbia University. Similarly, Stream EM was described by the Technics group. These efforts demonstrated the benefits of combined sensor technologies to capture the best possible representation of the underground. These programs have demonstrated cost effective capture along roads, which could also be applied to other transport corridors. Philip Meis described using these technologies in versions ranging from handheld, to small wheeled push carts, to multi-frequency/multi-channel arrays pulled behind powered vehicles in different terrains.

Some of the RFI respondents already conflate data to create additional value. For example BRGM, Columbia, and BGS are combining data from boreholes to create geological maps. Conceptually this was also described and extended in presentations from ASCE, BGS, BRGM, Delft University, and Dassault where the concepts of combining all the data held by an organization to create more value was outlined.

These approaches could combine data from multiple sensors with existing data to offer a ‘most likely’ value for any particular location. Understanding the reliability of results would likely be problematic at a very precise resolution but may well be acceptable for certain levels of enquiry. For example answering questions such as “How confident can I be that no Fuel pipelines run through my development site?” or “What is likely to be at a particular underground location?” could be answered without an absolute certainty but with appropriate confidence metadata. The confidence metadata could inform the next steps, for example undertake more detailed survey, excavate carefully, or reject the potential purchase of a site.

Advances in capture, modeling, and representation anticipated by some of the RFI respondents are expected to reduce the costs and/or increase the usefulness of captured data. For example Geo.works demonstrated a comprehensive interface for rapid capture of survey information; Columbia University described modeling 3-D soil using “random fields;” Accenture using overlapping high definition photos or scans to create models of exposed assets; and Dassault described using multiple sources of data from sensors, scans, surveys and existing data to create models with appropriate confidence levels.

Developing comprehensive, accurate, and interoperable utility maps and data may appear to be an overwhelming task. However, Flanders and other business cases provide us with excellent examples of how underground infrastructure information can be developed and used to support utility-related work processes that require excavation of the underground infrastructure. A number of speakers at the OGC NYC Workshop also made presentations and provided RFI responses about common business process approaches for building underground data that limit costs and deliver quality.

This section summarizes information from the RFI responses and workshop presentations relevant to deploying an Under Ground Infrastructure Information System UGIIS. The summary in this section leads to a recommended architecture based upon 1) Functional requirements from previous sections, e.g,. Business scenarios, 2) Typical characteristics of a UGIIS Deployment environment; 3) Functional architectures for different core UGIIS capabilities, 4) Tier-architecture approach to reduce silos; and 5) design tradeoffs to be resolved in particular deployments of the UGIIS platform architecture.

Earlier sections have detailed the user scenarios and business processes that define requirements for a UGIIS. For convenience, the earlier material of sections 8.2. "Governance" and 8.3 "Use cases and applications" is summarized here as UGIIS functional requirements.

Existing user scenarios requiring underground information:

Common business processes of integrating data:

The components of a federated UGIIS are deployed in multiple locations using open interfaces to access information across the Internet. Most deployment environments for UGIIS will share these characteristics:

An example of the distribute communications required for UGIIS is the KLIP system as described in the RFI Response from Informatie Vlaanderen. The figure below shows how a Map Request that is presented to the KLIP platform components triggers messages to network distributed components at the utility network authorities (UNA’s).

Requirements derived from this functional architecture include:

Key decisions discussed more below include:

Most or all of the UGIIS implementations in the RFI responses and the Workshop presentations required ingesting data to a platform database in advance of operations, e.g., requirement #4 in the Functional architecture for user requests in the previous section. The functional architecture for data ingest is suggested in this figure from XXX (TUM?).

The figure above shows the transformation using FME tools. The FME tools were also used to create the above-ground NYC CityGML model, but other tools exist as well. The HL Consulting RFI response discusses data ingest and the need for tools such as HALE (Humboldt Alignment Editor) to perform the mapping from the relational (staging) database to the IMKL GML encoding.

Requirements derived from this functional architecture include:

Integration of underground information in a UGIIS architecture allows for analysis and predictions that are not possible without UGIIS functionality. Based on ingesting and hosting data in a UGIIS platform enables: 1) Conflict Recognition Service associated with routine utility operations (e.g. RFI Response from the CityGML Chair); 2) Simulation of cascading effects due to natural disasters or man-made disasters (TUM Workshop presentation); and 3) Modeling of underground environment effects on built infrastructure (e.g., BGS and BRGM workshop presentations). Having the different network systems and the city objects linked in one data model will facilitate vulnerability assessment of the city infrastructure in order to develop mitigation plans. Development of Smart City information technology infrastructures will enhance and expand the types of underground modeling.

Requirements derived from this functional architecture include:

The preceding sections on functional architectures motivates a discussion of "moving beyond silos" that was in several RFI responses and Workshop presentations. For example, the Bentley RFI Response described ‘Silo’ as the situation where different disciplines do not communicate effectively, which can lead to a lack of coordination, and conflicts occurring – for example where street lights or safety barriers are positioned on the same alignment as drainage pipes. The ability to see models created by other disciplines and consume them – for example, by deriving elevations of utility chambers from a proposed road surface model - is an important factor in being able to produce a coordinated design.

The figure below from the Dassault workshop presentation shows the architectural approach of moving from silos to a tier architecture. Tiers contain a platform of components that perform business functions across the diverse datasets.

The RFI response and Workshop presentation by Accenture about the Chicago UGIIS shows a layered architecture as in the figure below. Note that the figure shows four tiers: Users, Visualization/Applications, Data Management, Data Inputs.

.

Based on the functional architectures described above, a consolidated list of components is presented below organized in a tiered architecture:

The components listed above are motivated by earlier sections with the exception of the "Augmented reality","Vector-Raster conversion", and "sensing and observation" components. Augmented reality clients were presented in the workshop by Esri and Bentley. The TU Delft RFI response and workshop presentation addressed the value of vector-raster conversion. Sensing and observations were presented in the workshop by Columbia/CCLS and Dassault. Columbia/CCLS proposed regular, and where appropriate, continuous geophysical mapping and monitoring of underground infrastructure to enhance situational awareness, forecasting, and predictions.

The tiered architecture presented above is a generalization of the multiple UGIIS architectures that were reviewed in the RFI Responses and the Workshop presentations. The architecture can be expected to be generally useful as a validated starting point for an organization-specific deployment. Each UGIIS deployment will need to be tuned to requirements specific to their environment. Several of these topics to be refined are presented below as design tradeoff studies. There is no single answer, but the discussion below provides guidance for applying the generalized tiered architecture.

The RFI Responses and Workshop presentations provided various approaches on what types and how much data is to be held in a UGIIS Database. Some approaches minimized the UGIIS database contents to mostly brief metadata and an identifier. Other approaches centralized all UGI data into the UGIM platform database. The minimal UGIM database approach depends upon real time access to UNA remote servers and provides the most update information by getting it from the source. The maximal UGIM database approach provides highest performance and availability because it relies little or not at all or remote servers. In addition to these system performance considerations, the contents of the UGIIS database will be affected by the data sharing policies to be agreed between the UNA organizations and the organization hosting the UGIIS database. (Engineering reports prepared for phases of the GEOSS Architecture Implementation Pilot provide an extended discussion of this tradeoff).

Related to the question of what is in the UGIIS database is the question of where and where do heterogenous data models get transformed? If all of the data was collected into the UGIM database it would probably be transformed upon ingest and before insertion into the database. But some data will be accessed remotely at the time of user requests and therefore transformed on-the-fly to meet user needs including conversion to visualization products.

Data must move between the distributed components of the UGIIS. Different exchange protocols were discussed in the RFI Responses and Workshop presentations. The protocols could be ground into three main areas: 1) file transfer, 2) Web service protocols, and 3) Application Programming Interfaces (APIs). The later two categories tend to overlap.

File transfer is chosen when a bulk transfer of a dataset as a file is appropriate. This could be done by the venerable and still useful ftp or by cloud-oriented file storage. Bulk transfers would be in advance of user requests. This may be particularly useful for the initial population of a UGIIS database.

Web service protocols enable access to portions of datasets. Several RFI responses and presentations described use of the OGC WMS and the OGC Web Feature Service (WFS). WFS is used for access to CityGML and other feature oriented data. The OGC Web Processing Service (WPS) was discussed earlier as a method for access to models. OpenMI also provides for the coordination of integrated analytical models.

APIs for the web have become very popular and very diverse. APIs can be very effective in rapid client-server developments in particular on the web using Javascript. The proliferation of diverse APIs has degraded interoperability. OGC has recently issued a white paper on geospatial APIs and is currently conducting several initiatives related to geospatial APIs. Results from those initiatives should be available for the Underground Pilot initiative.

Linked data techniques are applicable to the UGIIS environment. The EPRI RFI Response identified that data stored outside of GIS should be linked back to the GIS via a unique identifier to provide a seamless user experience. To achieve this requirement to coordinate information about individual equipment across systems, the Business Objects Registry Standard (IEC 61970 part 454) has been created. The standard provides a common system to identify and track business objects, i.e. poles, breakers, or transformers, across systems, despite a variety of names and representations in use between different applications and user groups. Creation of a Master Resource ID (MRID) for each object instance allows the coordination and translation of names between non-standardized systems and mitigates the impacts of version changes in the common information model. Without the MRID, systems cannot communicate seamlessly.

As this study demonstrates, the development of interoperable underground infrastructure data standards is extremely important to support a range of business processes that are essential to our economy and society, now and into the future. The geospatial community and the wider engineering and construction world have long understood the value of standardized and interoperable infrastructure data, but there have been formidable barriers to creation of these data. One of the key obstacles has been the absence of standardized data models to guide data development. The central focus of this effort by OGC is to finally provide such standards. At this point in our discussion we would also like to identify some of the other barriers to achieving this goal and begin the process of devising solutions, so that there is a real chance that efforts will be made to create data in conformance with the standards under development.

While not formally part of the RFI, it is a commonly held understanding that underground infrastructure data is highly sensitive and that data owners are strongly inclined to impose strict access restrictions. Large transmission lines and major generation and storage facilities, especially when they represent single points of failure, trigger points for cascading effects, or major supply points for other utility networks are potential targets for terrorism and sabotage, because damage to them can cause outages that affect many people and businesses. On the utility distribution side, knowing the location and characteristics of local shut-off valves and service lines can enable contractors to bypass normal permitting processes and take matters into their own hands. Additionally, some utilities with overlapping service areas – such as telecommunications companies - are reluctant to share location information about their networks because of competitive concerns. In other cases private utilities do not want government regulators to know details about their infrastructure if they believe it might lead to expensive mandates for upgrades, and to possible liability.

While the above factors have inhibited infrastructure data sharing, modern methods of data security can virtually guarantee that information can be shared safely, while inappropriate data access can be effectively blocked. Flanders, Belgium allows utilities to maintain their own data locally while sharing only those small portions with Informatie Vlaanderen – the government excavation coordinator - needed to support digging operations. In the Flanders case, data is exchanged by way of a Web API over an OAUTH-secured HTTPS connection, optionally as digitally signed data packages. GEO.works, working with UDOT, has also recommended sharing with utility engineers, upon request, just small data subsets relevant to the project area of interest.

Other potential security measures might include the use of end-to-end data encryption and the utilization of centralized secure fusion and analysis facilities. It is likely that some combination of these methods will enable security levels superior to the ones currently being deployed, while at the same time allowing critical sharing of data between trusted parties when necessary. Security and intelligence agencies of several counties might be willing to support the identification of infrastructure information security options.

Another factor that inhibits the development of standardized underground infrastructure data is the perception on the part of both government and private utilities that the costs of development are too high and are far in excess of any benefits that might be realized. However, based on information provided through this RFI, it became clear that there are strong indications that significant benefits can be derived from achieving infrastructure data standardization and interoperability. These benefits include fewer utility strikes, reduced construction delays, fewer and less expensive change orders on large scale projects, and minimizing damage and potentially, the loss of life, during emergencies and disasters. Being able to definitively document and quantify benefits would provide strong justification for the investments required to improve infrastructure data creations and sharing. It will be desirable for a UGIIS to demonstrate direct, tangible benefits to individual operators and other data owners by cleaning or value adding to their own underground data as part in return for submitting data to the UGIIS.

The development of a rigorous cost benefit analysis is possible if a sponsor can be found to finance the work, and a highly qualified financial analyst team identified and hired. It is likely that RFI respondees who provided benefit numbers (including Chicago/Accenture, Flanders, Ordnance Survey of Great Britain, GEO.works) would be willing to dig a little deeper and come up with more detailed information. Having good ROI information – which also factors into the cost effectiveness of options for data development - would then provide support and guidance to utility planners and decision makers to move forward in a fiscally responsible way. A GIS ROI calculator developed by the NYS GIS Association can be found at https://www.nysgis.net/featured/emerging-gis-resources/ and serves as an example of what might be done. A more detailed ROI study will be an important adjunct to the technical implementations of planned underground pilots.

To a significant extent, the systematic development of standardized utility data, depends upon the ability of local and state government to require its development. However, there remains uncertainty about government authority in this area. Many national, state and local governments already mandate OneCall and SafeDig functions but not utility data standardization. In the case of Flanders, Belgium authorities were able to authorize the development of interoperable digital utility data in response to the deadly Ghislenghien gas pipeline explosion. To promote more widespread development of standardized utility data it will be useful to understand all the legal mechanisms that can be used to require it.

Legal research, as part of Phase 2 of this project, could lead to a greater understanding of how government can encourage private utilities to improve their as-built data and make it available under secure conditions. Among the things we know already are: franchise agreements between utilities and government jurisdictions allowing utilities access to publicly owned streets and rights of way can include data development and reporting requirements. Similarly, street excavation permitting processes can be used to require utilities and construction companies to document utility location and characteristics whenever the pavement or sidewalk is opened. Researchers could look into these and other mechanism in common use to inform jurisdictions about their options and to provide examples of where and how they are being applied.

Financial resources are always scarce and even if security, ROI, and legal roadblocks are overcome, the money necessary to finance the development of interoperable underground utility data may still not be forthcoming. Utilities have limited budgets and many priorities, and may still put spending on brick and mortar, shovel and pipe ahead of improved information.

Phase 2 of this project could include the identification of options for financing the development of standardized, interoperable underground utility data. One such solution could be the establishment of a data development bank offering loans that are paid back by the stream of benefits coming from the use of improved data. Another option could be a government surcharge on utility bills that goes towards paying for data development. This method is successfully used to fund 911 emergency response operations in many jurisdictions but may not be universally desirable.

A goal of this study has always been to prepare the way for standards development by assembling knowledge and thinking about underground infrastructure, surveying data models and datasets that are proposed or in use today, and then making the case for any new data interoperability standards that can bring benefits worth the cost and difficulty of adoption. Prior sections of this report lay out evidence and present findings that information about underground infrastructure is often incomplete, out of date, of poor quality, and inaccessible to those who have need of it. Documented consequences of this situation include increased construction and maintenance costs, delays and breakdowns in essential utility services, barriers to improvements in urban livability, and significant threats to public safety. Initiatives detailed by RFI respondents and workshop participants to address UGI data issues clearly point towards positive ROI in cost savings, improved health and safety, and higher-functioning urban environments. From these findings, the report makes three recommendations for steps that can lead to an improved UGI data situation.

This section of the report takes up the first recommendation and connects it with steps and activities for future standards consultation, harmonization, specification, implementation, and adoption. Prior sections (Data Models, Underground Environment) review a number of existing utility and underground environmental data models that cover many of the features already identified as both common and strategic. For example, the INSPIRE utility data models provide an invaluable head start since they serve as the foundation for KLIP, Flanders’ successful utility information integration application – one of the few public systems in the world that utilizes standardized digital data from all utility companies operating in a broad region. The existing data models will be the starting point, in the context of the Use Cases also detailed in this report, for creating a new generation of 3D-4D geo-enabled data models ready for prototyping activities and consideration as new standards.

One particular aspect which is currently not embedded in these data sharing data models is the ability to bring together and fuse feature data and observations from different sources, for example existing utility asset records or scanned underground data and data from open pits, such as ones dug for emergency repair or trial holes.

Standards are essentially agreements to cooperate on the basis of a common understanding and shared (or at least complementary) goals. Development of a common standard data model is no different and involves similar stages of (typically iterative) activity and progress.

The creation of underground infrastructure interoperable information (UGII) and information systems (UGIIS) will be a difficult task, but just as challenging will be to convince government organizations and utilities to create the data that conforms to the systems. Inhibiting data development are concerns about data security, finding the right legal mechanisms to support compliance, and the expense of building standards-based data which may not yield benefits worthy of the investment. These critical policy issues require further research to establish frameworks within which efforts towards integrated UGII and functional UGIIS implementations can succeed. Given the resource limitations of this project, it is likely that research in these areas will need to be limited to surveys of OGC members and summaries of current methods, potentially augmented by partnerships with industry organizations such as EPRI.

The OGC project team will reach out to all project participants and to organizations and subject matter experts with knowledge about state-of-the-art security issues and protocols. Options to be explored include secure, central clearinghouses for utility data, encrypted API access, and role-based authentication. The team will also look at current security concerns and implemented measures for existing utility information systems.

Research will be conducted to better understand legal agreements between government entities and utilities regarding the provision of infrastructure data. OGC UI project participants will be contacted as will the UI Community of Interest made up of about forty US local governments and organizations who are tracking progress on this project. Legal arrangements that can support compliance with data standards can include franchise agreements, excavation and construction permitting requirements, and OneCall/SafeDig requirements.

In their responses to the OGC UI RFI and in presentations made at the UI workshop, a number of participants including Accenture/Chicago, Flanders, Ordnance Survey of Great Britain and GEO.works provided intriguing information about the losses and costs associated with not having standards based infrastructure data; and the benefits of having such data available. The OGC project team will interview these project participants and others seeking to obtain more detailed information. An attempt will be made to put this information into a financial model that will allow users to select from a range of costs and benefits, to determine potential ROI customized for each particular jurisdiction.

A major factor inhibiting the development of comprehensive underground infrastructure data is the perception that the cost of data development is very high. The OGC project team will work to identify methods that have the potential to make the funding process less onerous. One option to be explored will be the imposition of a utility surcharge such as the one on telecom bills that fund 911 operations. Other options include surcharges on excavation and construction permits and loans paid back from the monetized stream of benefits achieved through the use of standardized, digital utility data.

Once data model prototypes have been developed and reviewed, they need to be tested in a variety of settings so their usefulness can be validated. Of great interest will be the work involved in extracting from existing data repositories the priority features and attributes called for by the data models. For utilities that have poor data or no data at all, it will be necessary to understand the cost and effort to build and validate the necessary data. This section of the report presents an outline of proposed pilot activities to accomplish the necessary model testing and refinement.

For the success of this project it is not enough to build the best possible interoperable data models. Data model work must be done with an eye towards how the data to populate those models will be created. We therefore propose that in addition to data model testing, there be a series of research projects which will develop information to support and facilitate the building of underground infrastructure data inventories. These research projects will focus on state of the art methods and strategies for developing accurate utility information from incomplete, incompatible, and inaccurate older records. We will also be looking at lowering the barriers that in the past have impeded underground infrastructure data development and sharing. We expect to address security concerns, lack of quality return on investment information, legal options for mandating quality data development, and financing alternatives.

TBD

TBD