Between the Poles

Automated drone systems improve power line inspections and reduce costs

Transmission line inspections are mandatory in North America. Traditionally helicopters are used for these inspections, but with helicopters the risk of accidents is significant and helicopters are costly to maintain and operate.  Drone inspections dramatically reduce the risk of accidents.  Aeriosense Technologies has developed an automated drone inspection system that can reduce the cost of inspections by up to 55%. 

Background

Transmission line inspections are essential in ensuring grid reliability and resilience.  They are generally performed by manned helicopters or by a ground crew.  Data may be collected with cameras and analyzed to detect a variety of conditions including corrosion, evidence of flash over, cracks in cross arms, and right-of-way issues such as vegetation encroachment. Inspections every year or 18 months are mandated by NERC in North America and are not optional. 

With over 200,000 miles of high-voltage transmission lines and 5.5 million miles of distribution lines in the United States, improving the efficiency and reducing the risk of inspections would have a major impact on the reliability of the power grid.  Drone technology, legislation and regulations enabling beyond visual line of sight (BVLOS) flights, software for planning flights and analytics for feature identification are making it conceivable that the expensive process of inspections could be completely automated.  Even now current technologies make it possible to dramatically reduce the cost of power line inspections. According to Navigant Research, globally electric utilities are expected to spend over $13 billion a year on drones and robotics by 2026, a dramatic increase from about $2 billion currently. 

Automated drone inspections

Flight plan for transmission tower inspection

Using automated unmanned aerial systems (UAS) Aeriosense Technologies can reduce the cost and improve the safety of vertical infrastructure inspection.  Unlike many other companies which offer drone services to many different industries, Aeriosense focuses only on the electric power sectors, currently transmission but in the future distribution.  I had a chance to chat with Alex Babakov, a co-founder of Aeriosense, who has many years of experience in the electric power sector, including working for BC Hydro Research/ Powertech’s High Voltage Laboratory (HVL). 

Alex explained to me that Aeriosense provides a complete visual inspection of pylons and poles, cables, and right of way including vegetation encroachment using a UAS typically with a photo camera.  Examples of the types of problems that can be identified include broken strands in lines crossing canyons and rivers, corrosion on insulators and other equipment, evidence of flash over, cracks in cross arms, and problems in the right of way such as human or vegetation encroachment.  Aeriosense also conducts inspections as part of storm response to identify damage resulting from extreme weather conditions.

Flight plan for transmission line inspection

Automation enables Aeriosense's drone systems to provide higher quality inspections at lower cost. In preparation for an inspection flight, the available GIS and LiDAR data is analyzed with the aid of machine learning algorithms to create an optimal flight path.  With a flight path optimized in this way cost reductions of about 36% can be achieved compared to manually operated drone-based inspections.  Furthermore compared to manual drone inspections automated inspections reduce pilot fatigue and safety risks while conducting inspections.

Aeriosense is also active in distribution inspections. The sheer size of the distribution network in terms of total number of kilometers and number of pieces of equipment makes it clear that an automated approach is required. With current technology and regulations, automated inspections can be applied to some but not all distribution circuits. In the future with new technology developments and regulations automated inspections will be able to be applied to more distribution circuits.

Cost reduction

Based on data of inspection efficiencies from our past projects and the data coming from the field trials done by Electric Power Research Institute (EPRI) the reduction in cost with using an automated UAS compared to a helicopter inspection is between 55% and 28%.  The challenge is doing these cost comparisons is that there are intangibles that are difficult to quantify. Safety is one of these. Helicopter accidents are not infrequent in the electric utility industry and more often than not they lead to fatalities.  BC Hydro, for example, has a requirement for helicopters to be dual engine for redundancy and rates for this equipment are much higher.  Other utilities may not be willing to pay for the extra security whichl reduces the rates of the helicopters.

Improved safety

One of the most important benefits of Aeriosense's automated UAS approach is safety. Because of the relatively high risk of helicopter inspections, utilities have been looking for alternatives. Drone-based inspections are much safer and this provides additional motivation for utilities to move toward adopting drone-based inspections.

Development of training datasets for machine learning

Another benefit of the drone-based inspections are that utilities get a historical, unbiased record with the imagery that is captured by the drones. This enables engineers to go back and review the condition of assets not from a written record but from photos.  This provides a more objective view of how the condition of an asset has changed over time.

Collecting historical data has another important benefit in that it can provide a body of imagery that can be used to train machine learning algorithms. Collecting the data now will allow them to apply the AI tools currently being developed in the future.

Future applications of automated drone systems

The impact of more frequent, less costly inspections is improved reliability. Combining the efficiency of automated UAS inspections with defect detection using machine learning will allow utilities to get a very clear, timely situation assessment of their distribution system.  Aeriosense is currently working with the Centre for Energy Advancement through Technological Innovation on a project to help utilities identify current opportunities of automated drone applications and to develop best practices in implementing automated drone inspection programs.  Aeriosense is currently working in the transmission and distribution sector, but is actively exploring applications in the substations and generation verticals.

Another area that Aeriosense is currently piloting is the use of beyond visual line of sight (BVLOS) UAS flights for transmission inspections. FAA rules permit beyond visual line of sight (BVLOS) UAVs through Part 107 waivers.  I have blogged about Xcel Energy which obtained one of the first three waivers to fly UAVs for beyond visual line of sight operations. Xcel Energy estimates that flying drones beyond line of sight will eventually cost between $200 and $300 per mile, compared to helicopter flights that cost an average of $1,200 to $1,600 per mile.  More than 30 of these waivers have been granted and more than 20 companies are operating beyond BVLOS drones, still with some restrictions.  FAA very recently began allowing BVLOS operation using radar to track the location of the UAV.

Summary

Automated drone technology is already impacting transmission line inspections by improving safety, reducing the cost and time required to complete inspections, and by collecting an objective historical record that can provide training data for future machine learning applications. It not only enables more frequent inspections and improved reliability, but it also enables more rapid inspections, which is especially critical after a storm or other emergency situation. Research is underway at Aeriosense to determine the applicability of automated drone technology in other sectors, such as substations and generation, of the electric power industry.

Is legacy GIS failing network infrastructure operating companies ?

I thought it would be worthwhile to include here an expanded version of my contribution to a webinar that IQGeo presented last week with a provocative title that I think goes right to the nub of the challenge that telecom and utility companies are facing in developing digital twins of their network assets.  In this new world there is an exponentially increasing volume of real-time data and less and less time to conduct analyses and make decisions.  Telecoms and utilities and finding that their legacy GISs are simply not up to the task.  The webinar aimed to shed some light on the problems that telecoms and utilities are facing with their legacy GISs.  The webinar speakers were Linda Stevens, Managing Partner at 51by1, Peter Batty, Chief Technology Officer at IQGeo, and myself. It was moderated by Steve Tongish of IQGeo.  The full webinar is available on demand here.

Geoff Zeiss opening comments

Common symptoms of the failure of legacy GIS include as-built backlogs stretching into months, inaccurate as-builts submitted by contractors, limited updates from the field, frequent underground infrastructure damage during construction, inability to connect customers to transformers or feeders, and repeats (multiple truck rolls because of inaccurate information about field assets).

As-built backlogs are the result of a construction information flow involving different organizations with different tools.  Engineering uses CAD, construction relies on paper drawings, and records management manages data in a GIS.  The major factor contributing to as-built backlogs is the decades old lack of interoperability between CAD and GIS. The result is that the GIS is inaccurate and permanently out-of-date.

During the webinar there was a poll Poll asking webinar participants how long is their average as-built update backlog time?
8 % Less than 1 week
8 % Less than 1 month
21% Less than 3 months
25% More than 3 months
38% Don’t know or not relevant

It is indicative of the problem that of those from utilities or telecoms only a quarter had a backlog of less than a week.  (There were a significant number of people participating in session who were vendors or service organizations for whom the question did not apply hence the 38% response to the last option.)

Another problem relating to as-builts is lack of quality control. Construction contractors often submit as-designeds rather than what was actually constructed.  Contractors may also red-line as-builts to indicate the difference between the design and what was constructed, but these are often sketches rather than accurately surveyed data.  The result is that sometimes it is uncertain on which side of a street a cable or pipe has been installed or whether it was actually installed on private property instead of on the public right-of-way.

If there was an effective feedback mechanism these problems might be ameliorated over time, but getting updates from the field into the GIS in a timely fashion is a challenge for many utilities and telecoms.  I have encountered telecoms and utilities where updates from field staff did not get into the GIS at all or were seriously backlogged. Not seeing updates reflected in the GIS in a timely manner discourages field staff from reporting changes and errors.  The result is that over time the quality of GIS data degrades.

Inaccurate GIS data is responsible for 400,000 incidents of underground infrastructure damage every year in the U.S.  This represents a serious risk for workers and the public. Just a week ago there was an explosion in Murrieta, Los Angeles which left one person dead and fifteen injured.  In the last few months there have been explosions in downtown San Francisco, Durham NC, and I could go on.  It is also a multi billion dollar drag on the economy and is responsible for many construction projects running over budget and over schedule.

Missing or incomplete connectivity linking customers to transformers and feeders directly affects outage duration.  It is not uncommon for a utility not to know the correct connectivity for 15-20 % of its customers ( I would add that knowing the phase for each customer is often a bigger problem).  Not being able to identify a damaged upstream device when a customer reports an outage or being able to identify which customers are affected by a failed device directly impacts outage duration and restoration (SAIDI and CAIDI).

Inaccurate GIS data is also responsible for "repeats", where more than one crew trip is required to maintain a piece of equipment or fix a problem.  Typically it involves one trip to determine where an asset is and its details and another to return with the required equipment to resolve the problem.  This also impacts outage duration statistics.

Other symptoms include redundant data where the same information is maintained simultaneously in different GISs by different groups within a utility and difficulty in reporting accurate situation assessments of network assets to management and regulators.

Question: - Utilities devote huge resources to maintaining redundant geospatial data in different systems, much of which is inaccurate and out of date. This unreliable GIS data directly affects outage statistics, field staff efficiency, the reliability of reporting to regulators, and the ability of management to get an accurate perspective on operations. Why has this situation persisted for decades?

There are several reasons for this problem persisting as long as it has.  Among the top technical reasons is the incompatibility between CAD and GIS data. An important contributing factor is that multiple organizations and departments are involved in the workflow.  A fundamental problem is procurement practices that don't treat data as important as the physical assets themselves. An enlightened procurement policy can resolve this problem. I worked with a large Brazilian telecom that enforced a policy that contractors were only paid when as-builts were in the GIS and verified.  Another fundamental problem is that many regulators have not recognized the value of accurate, up-to-date GIS data.  I only know of a few that recognized how important this is for utility operations.  For example, about a decade ago ANEEL, the electric power regulator in Brazil, imposed a rule that all utility GISs had to be 95% accurate or better.  This was enforced by requiring frequent audits of utility GISs with penalties imposed if 95% was achieved.

Another important reason that this problem has persisted is that in some countries including the U.S. and the U.K. there is a liability structure that invariably assigns responsibility to the construction contractor for damage to underground infrastructure during construction.  This provides little motivation for telecom and utility network operators to improve the reliability of their GIS data.  In some countries such as France liability is shared between the telecom or utility operator and the contractor.  For example, if a contractor hits a utility facility that is shown accurately on utility GIS maps, it is the contractor's liability.  Conversely if a contractor hits a utility facility that placed incorrectly on utility GIS maps, it becomes the utility's liability.

Question: What is the impact of reality capture on how we create and maintain geospatial data? Can capture of network connectivity be fully automated ?

I have seen estimates of telecom and utility GIS accuracy of 70-80 % or even lower.  With recent advances in reality capture a question that utilities are asking themselves is whether it is more cost effective to conduct a complete reinventory of its network assets, to recapture the location and attributes of its field equipment, or to try to correct the existing GIS data.  Recently a very interesting pilot was carried out by Duquesne Light in western Pennsylvania.  It compared the network model derived from a standard walking inventory with a model derived from LiDAR/photo data captured with a truck and post-processed in the office.  It found that the model based on LiDAR scans was just as reliable as the walking inventory and much safer.  In addition the LiDAR scan provided significant side benefits. For example, It could be used for a joint-use audit. The LiDAR data also allowed clearances between poles and cables to buildings and structures to be accurately measured to ensure regulatory clearances and to verify third party attachment heights. 

In a related development machine learning is being applied to automatically recognize different types of telecom and utility equipment in LiDAR and photo imagery including high resolution satellite imagery and to automate inspections for maintenance issues such as corrosion, vegetation encroachment, and other issues. To support this application of AI, EPRI is collecting training data sets for utility applications of machine learning.  A startup SiteSee claims that it can inspect telecommunications towers with a UAV and automatically analyze the resulting imagery for problems with antennas. I don't think that we are quite at the stage of network connectivity being routinely determined using these techniques, but I suspect that the day is fast approaching when this will be possible.

Question: Poor GIS data quality for underground assets within the utility industry can have direct and very serious consequences in injuries and lost lives, not to mention the huge financial impact in damages and fines [400 lives, thousands of injuries and a $50 billion drag on the U.S. economy arising from 400,000 cases of utility damage every year]. Can new tools for underground asset capture and the sharing of this data begin to mitigate these risks?

There are two major challenges in improving the reliability of our maps of underground utilities, remote sensing technology to detect them without excavation and sharing the information that is captured everyday about utilities exposed during construction excavation.  In the last few years there have been important advances in both areas.

One of the most versatile technologies for detecting underground utilities is ground penetrating radar (GPR).  It can detect underground objects that electromagnetic techniques can't and it is better at estimating depth than other technologies.  This makes it possible to generate 3D maps of underground utilities from GPR scans.  However, until now it has required a trained geotechnologist to interpret the radar scans. GPR arrays have typically been mounted on a pushcart making scans slow to acquire and hazardous for the operator.  In the last year or two new technologies have been developed that allow GPR scans to be acquired much faster using a vehicle.  With this approach there are no boots on the pavement making it much safer for the operator.  It is also much faster.  The most recent technology advance makes it possible to capture GPR scans using a vehicle travelling at up to 135 km/hour. Another advance is combining a GPR array with a LiDAR scanner on the same vehicle to capture below- and above-ground scans simultaneously.  A major advance is new GPR software processes multiple GPR scans to present a view of the underground that enables non-geotechnical professionals to interpret GPR scans.  These are very important advances in the detection technology and in democratizing its use.

In North America every state and most provinces have a government mandated one-call centre which requires anyone planning to excavate to call 811.  It requires all utilities and telecoms with services in the area of the excavation to paint the ground indicating the probable location of their cables and pipes. In the U.S. the locate industry is estimated to amount to billions of dollars annually.  But in general this information is not shared and the same underground assets may be located over and over again.  Recently there are initiatives at the city, regional, and national levels to share information about the underground.  The most recent is the National Underground Asset Registry (NUAR) in the U.K., supported by the Geospatial Commission, that is currently running two pilots in the Northeast of England and in central London.  In Belgium the KLIP initiative which has been running for a number of years was motivated by the 2003 gas explosion which caused multiple fatalities.  In the Netherlands a similar initiative KLIC for sharing the location of underground utilities was augmented in 2018 by a "key registry of the underground" (BRO) which mandates sharing geotechnical information discovered during drilling and construction excavation.  There has also been a pilot by the City of Chicago to share information captured during excavation with contractors.  An interesting experiment by Bentley placed low cost consumer cameras on a backhoe to photograph exposed utilities and used software to identify and georeference underground utilities exposed during excavation.  In another experiment exposed underground infrastructure together with accurately surveyed control points were photographed with smart phones and the images processed to identify and georeference underground utilities.

When viewed together the technical advances in GPR and the initiatives to share information about the underground are evidence of accelerating vendor investment in remote sensing hardware and software. In addition governments are beginning to show serious interest in finding ways to encourage the sharing of information about the location of underground assets.

Question: Why has there been so little change in the GIS industry in utilities and communications over the past twenty years? Why have there been no new entrants in this space?

A fundamental problem is that data has been valued less than the physical infrastructure itself.  The challenge of changing traditional construction practices is another.  I also believe that there has not been an awareness of the cost of unreliable GIS data among utilities and regulators. But with the growing interest in developing digital twins of network assets there is a growing awareness of the importance of accurate, real-time GIS data among regulators and utilities and telecoms.  The example of ANEEL in Brazil is evidence of how regulators can contribute to changing this.  As I have mentioned previously enlightened procurement policies can also have an important impact. Another factor that is harder to change is a liability structure relating to utility damage during construction that provides little motivation for telecoms and utilities to improve the quality of their location information about their underground assets.

Audience question: What are the principle differences between the GIS requirements for telecom and utility network operators? What can these different industries learn from each other?

In the U.S. since the breakup of AT&T, the development of cell technologies and the rise of the internet, telecom (including cable) has become extremely competitive.  This has engendered a focus on the customer that is still rare among electric power and gas utilities, and even more so among water utilities. This has led to the explosion in smart phone capabilities and smart phone apps. It also means that telecoms are reluctant to share information about the location of their network infrastructure.  This makes them reluctant to share information about their underground networks without assurance that this no more of this information is shared than is necessary.  Electric and gas utilities are generally more cooperative and more willing to share data.  Another major technical difference is that large telecom companies have led most industries in the development of IT technologies - remember the developments at Bell Labs.  Most electric power, gas and water utilities have not had much IT capability in-house, which is why there is high level of interest among small to medium sized utilities in software as a service (SaaS).  Another technical difference is that telecom data models are more much complex than utility models.  My rule of thumb has been, telecom data models involve hundreds of classes, electric maybe a maximum of 150, and gas and water less than 50.  This is changing with the rise of the smart grid with utility networks evolving to something closer to communications networks.

This suggests that utilities can learn a lot from the IT expertise of telecoms and in particular with the growth of smart grid from the complex data models and supporting applications developed by telecom firms.  They can also learn from the consumer-oriented focus represented by smart phones and smart phone apps.  Telecoms could learn from utilities about the benefits of sharing data, especially about the underground.

Audience question: What is the role of governments to improve the accessibility and interoperability of utility

As I have mentioned there is a growing awareness of the social and economic cost of not knowing where network assets are located or their condition.  Collecting and sharing this information does not necessarily require government involvement, but there are reasons why governments should be involved.  As I mentioned this is a sensitive issue for the very competitive telecom industry.  Any policy has to protect critical competitive information.  Secondly it can't simply push all the cost off on construction firms who already suffer from very low margins.  In the U.S. an important additional reason for government involvement is the Critical Infrastructure Protection act which restricts access to location and other information about network assets.  As I have mentioned a growing number of governments have gotten or are getting involved in encouraging or mandating the sharing of information about network assets. Furthermore in some countries there is discussion underway about a national infrastructure map (U.S.) or a national digital twin (U.K., Singapore and others).

Key takeaways

• There are important technology innovations among them reality capture for above and below ground and machine learning to detect and assess network assets that are poised to dramatically change the utility operating business.
• Governments are recognizing the economic and social costs of not knowing the location of network assets and are implementing policies to encourage or require the sharing of location information about network assets.

High resolution satellite imagery helps innovative electric utility optimize vegetation management

Vegetation is a significant source of outages for many utilities. In some regions from one quarter to one half of all outages can be ascribed to vegetation.  Vegetation encroachment initiated the massive North American northeastern outage of 2003.  It is so critical for utility operations that some utilities have a vice president responsible for vegetation management.  It can be costly for utilities to identify areas of high risk vegetation encroachment.  Scheduling and routing crews to prune the trees that require attention is often sub-optimal adding to the cost of vegetation management.  Optimizing these processes can produce significant benefits in the form of fewer outages and reduced costs. At Distributech 2018 Pedro Lemes, and engineer at Energisa (Brazil) described an innovative project which used high resolution satellite imagery to optimize vegetation management. Energisa is the sixth largest Brazilian power utility with 6.6 million customers.

Vegetation is the source of many outages at Energisa and the budget for vegetation management represents a significant proportion of its operating budget.   To reduce the number of outages due to vegetation and the costs of vegetation management, Energisa initiated a project to develop a tool that would allow it to identify and map high risk areas for vegetation encroachment and secondly to optimize routes and crew allocation for tree pruning.  The objective of the project was to decrease the number of outages due to vegetation encroachment, reduce vegetation management costs especially third-party firms hired to identify risk areas for vegetation encroachment, reduce the level of risk associated with vegetation near power lines, and to bring more IT into operations at Energisa.

The project innovated in using satellite imagery together with GIS data to solve the problem of identifying high risk areas for vegetation encroachment in a distribution grid.  By analyzing satellite imagery it was possible to distinguish vegetation from other objects, to differentiate trees from grass and bushes, measure the height of trees, and using GIS data to measure the distance between trees and poles and cables. Energisa contracted with Airbus Pléiades to provide 50-cm resolution, colour, orthorectified as well as near infrared (NIR) imagery. The contract with Airbus specified that all imagery supplied to Energisa had 15% or less clouds.  The imagery enabled identification of vegetation using the RGB and near infrared (NIR) bands.  In addition Pleiades provides high-resolution stereo imagery enabling generating digital surface and terrain models.  Trees heights could be determined by comparing surface and terrain models.    Trees were separated from other vegetation by a minimum height threshold.  Even when there were many trees together, it was possible to separately identify individual trees. NDVI (Normalized Difference Vegetation Index) was used to differentiate between vegetation and other objects and to distinguish healthy and diseased trees.  Coconut palms were the most challenging because of the lack of a central crown.  Energisa's GIS provided the location of line segments, switches, poles and structures.  Based on the information derived from the two sources it was possible to calculate the distance between trees and cables and identify high risk areas of vegetation encroachment.

The second part of the project involved developing a pruning management system to optimize pruning.  Once the risk areas had been identified, based on the available budget, team locations, the desired level of risk and the cost of pruning a tree, the algorithm was able to determine an optimized route and schedule for pruning operations for each team.  The algorithm was sufficiently robust to be able to be applied to a neighbourhood or an entire city.

Nowadays there is a plethora of satellite imagery with resolution as high as a foot and a frequency of several times a day.  One company is now imaging the entire landmass of the Earth every day. Satellite imagery is becoming increasingly competitive which is lowering the cost of imagery.  Energisa's creative use of satellite imagery and another innovative application that uses satellite-based radar interferometry to detect subsidence, which often can be sourced to a water leak, are two examples demonstrating how satellite imagery can be applied by innovative utilities to improve performance and reduce costs.

Are point clouds going away ?

At the SPAR3D 2017 conference in Houston, one of the interesting questions posed to a panel comprised of Greg Bentley (CEO of Bentley), Burkhard Boeckem (CTO of Hexagon Geosystems), and Shabtay Negry (Senior VP at Mantis Vision) is whether point clouds, whether from digital photography or LiDAR, will effectively disappear, being replaced by meshes for most reality modeling applications.  Meshes are mathematical constructs, typically 3D triangular networks, that are smaller and much easier to manipulate than point clouds, which are typically huge and a challenge to edit.
 
I blogged previously about a digital model comprised of meshes of downtown Philadelphia that was used to orchestrate and secure the Papal visit.  All the data was collected as photographs which were then used to create a single reality mesh using Bentley ContextCapture.  At SPAR3D there are presentations on two other projects which used a similar approach to create models of the city of Helsinki and the Cologne Cathedral.

Another fascinating example is the USS Arizona digital project which combined underwater lidar, sonar, aerial imaging, and existing photography and surveys to produce a survey-grade 3D model of the USS Arizona in the form of a mesh using Autodesk Remake.  Surveys were conducted in 2014 and 2016.  Just recently the two digital mesh models of the ship were compared to create a difference mesh to show areas where there has been change in the two years since 2014.

At SPAR3D there was a consensus among the panelists that meshes are replacing point clouds for most applications. Point clouds are just too big and have too little intelligence. They will remain as an intermediate data type because most scanners generate point clouds.  Replacing point clouds for many applications meshes will grow in importance in the future.  But realizing their full potential will require interactive, editable meshes that allow small bits to be changed or replaced without having to regenerate the entire mesh (which is often the case with triangulated surfaces).  Easy to use analytic tools like those used to create the difference map of the Arizona will also be required.  The panelists thought that with the appropriate edit and analytic tools people will be able to manipulate surfaces easily and may not  even be aware of what the form of the underlying data structures is.  Smaller files will also make cloud processing easier because in the future huge point cloud files won't have to be uploaded to the cloud.

Utility industry vendors urged to adopt data model standard

Software comes and goes, but a utility's data model persists. A shared electric power data model enables electric power utilities to move away from incompatible point solutions toward tightly integrated solutions which are essential for real-time monitoring, situational assessment and decision making.  Based on today's smart devices and sensors and utility-wide communications network, decision making in a real-time, big data world requires tight integration between applications.  To realize the full potential of the smart grid, utility industry vendors need to move rapidly to adopt data model standards, that's the recommendation of Geoff Cameron of AGSI. which has developed a platform based on a utility data model standard that provides real-time access to critical infrastructure, customer and operational information. 

 

Traditional electric utilities are typically departmentalized and rely on manual, paper-based processes. Sharing knowledge is relatively slow, cumbersome, and reactive in nature and characterized by redundant data, many incompatible point solutions and non-real-time.  Facilities are inspected periodically, data is accumulated and analyzed, a report is written and sometime in the future do something is done.  Today's intelligent electronic devices, bidirectional communications networks, and automation provide the basis for dramatically improving a utility's SAIDI and SAIFI outage statistics.  It also provides a foundation for transactive energy.  What is required to realize the full potential of the smart grid is a platform that enables real-time integrated applications.

 

At Distributech 2017 Dan Guatto, COO and VP of Engineering and Operations, described how Burlington Hydro (BHI), a distribution utility with 65,500 customers, has evolved from a point solution environment to an integrated, secure, and real-time utility.  As is typical of many distribution utilities, BHI was facing increased technology sophistication, the need integrated business workflows (operations, engineering, customer engagement, financial) to improve efficiency.  BHI had to address siloed legacy systems and business processes.  Many industry solution offerings fixed individual problems but were expensive and did not take into account integration and long term data management.  Many “enterprise” offerings in the marketplace were proprietary black box solutions with little flexibility that made BHI feel that it was surrendering control of its data and processes.  One of the major problems BHI identified was the lack of O.T. / I.T. convergence for true enterprise efficiency.   Solutions required quite expensive capital outlay and deployment.  Projects became too big and and bogged down and were unable to address fast paced industry change.  In Ontario some of the larger utilities made significant technology investments, but at the end of the day these projects were extremely expensive, disjointed and largely underutilized (or even abandoned). BHI was also concerned about the infancy of smart grid technology and rapid future changes and didn't want to be tied to an inflexible system as the speed of industry changes accelerated.  BHI was convinced that a sound system architecture had to be based on electric industry standards and was aware that AGSI had been working with EPRI on a real-time platform based on Common Information Model (CIM) and Open Geospatial Consortium (OGC) standards.

 

Over the past decade open electric power data model standards have been developed for transmission, distribution, smart substations and other electric power functions.  Worldwide the Common Information Model (CIM) has become the standard for transmission.  CIM is also widely used outside of North America for distribution.  In North America the Multispeak standard is the most widely used distribution data model standard, although CIM is used by larger utilities.  These data models are mature and cover virtually all aspects of electric power utility operations.

CIM was originally developed for standardizing electric utility information exchange.  It was designed by electrical engineers, adopted by the IEC and supported by EPRI and multiple electric utility partners, vendors and organizations.  The IEC, EPRI and partners continue to extend the model to cover asset health, station design, work, smart grid integration, IoT Integration, and network operating protocols.  CIM is able to support all aspects of electrical operations (OMS/DMS), electric analysis / smart grid, distributed energy (DER), work orders, crew management, and assets.  AGSI has implemented the CIM Core, Topology, Wires, & Assets object models in a relational database (Oracle Spatial).  Its implementation represents a real-time network model of the CIM electric system data model and provides a centralized view and interaction with the network in real-time.

 

Today this technology connects and integrates engineering, SCADA, AMI, GIS, load analysis, other smart grid deployments such as EV charging, customer, and asset databases into a single accessible platform and enabled BHI to evolve from a point solution environment to an integrated, secure, and real-time utility.  Real-time data unification provides a basis for transparent workflows and a centralized reliable source of truth enables advanced analytics for optimized investment planning.  An open architecture based on industry standards provides the flexibility to adapt to rapid changes in technology as the smart grid evolves. Real-time information sharing enables faster decision making and increased customer satisfaction via real-time situational awareness about outages and planned work.  Don Guatto believes that it provides the foundation for transactive energy which is looming on every utility's horizon.

U.S. governors urge Trump to support renewable energy and grid modernization

The Governors' Wind and Solar Energy Coalition (GWSC), representing 20 U.S. state governors, both Democratic and Republican, sent a letter to President Trump presenting the benefits of renewable energy, arguing for a Federal commitment to support renewable energy and grid modernization, analytics and simulation and to streamline regulation of the grid.

The letter outlines how renewable energy is providing jobs and investment to low-income U.S. counties. For example, U.S. wind facilities pay rural landowners $222 million a year, with most of it going to landowners in areas with below-average incomes. $100 billion has been invested by companies in low-income counties, where some 70 percent of the nation’s wind farms are located.  Wind energy currently generates five percent of U.S. power demand. Last year, the country’s solar industry employed over 200,000 and added 31,000 new  jobs. Most of the installations are in rural areas and have provided landowners another income option.

Furthemore, the letter argues that any national infrastructure legislation introduced in Congress should provide significant funding for grid modernization. The complexity of the U.S. aging, multi-state grid makes the need for sophisticated modeling and analysis critically important. The GWSC asks the President to create a new state-federal task force, in conjunction with the Federal Energy Regulatory Commission (FERC) and the National Laboratories, to examine options to  modernize and streamline state and federal regulatory processes.

The GWSC suggests that if the United States does not continue robust federal research and development programs in wind  and solar energy, China will assume leadership and see most of the economic benefits in these critical technologies. The Governors make the case that the growth of the renewable energy industry is an American success story built on federal research and development, state policy leadership, private sector investment, and ingenuity.

Citizen geodata science an emerging big data trend

Gartner's latest hype cycle identifies a new trend "citizen data science" that has important implications for the geospatial sector.

At the recent Canadian Institute of Geomatics (CIG) Ottawa Branch Workshop on Big Data: What is it? Technology and Applications, Monica Wachowicz from the University of New Brunswick pointed out a brand new entry Data Science in the Gartner Hype Cycle 2014 that wasn't in the Gartner Hype Cycle 2013. Then in the Gartner Hype Cycle 2015 Data Science and Big Data have disappeared and a new entry with the intriguing name Citizen Data Science has appeared.

2013 Big Data

2014 Big Data and Data Science

2015 Citizen Data Science

What is data science ?

Monica referenced research by Harlan Harris et al “Analyzing the Analyzers” An Introspective Survey of Data Scientists and Their Work which surveyed data scientists to try to determine what data scientist do.  They divided data scientists into five categories based on their self-ranked skill sets; Statistics (which includes spatial statistics), Math/Operations Research, Business, Programming, and Machine Learning/Big Data, and four categories based on their respondents' self-identification; Data Researchers, Data Business people, Data Engineers, and Data Creatives.  

• Data Business people are most focused on the organization and how data projects yield profit.
• Data Creatives excel at applying a wide range of tools and technologies to a problem, or creating innovative prototypes at hackathons.
• Data Developers are focused on the technical problem of managing data — how to get it, store it, and learn from it.
• Data Researchers backgrounds are in statistics or mathematics, 75% of them have published in peer-reviewed journals, and over half have a PhD.

The resulting landscape of data scientists shows a broad range of skills distributed among the self-identified professional categories.  The researchers found that

• a defining feature of data scientists is the breadth of their skills, and their ability to single-handedly do at least prototype-level versions of all the steps needed to derive new insights or build data products.
• the evidence points in favor of a scientific versus a tools-based education for data scientists. 
• 70% of the respondents had at least a Master’s degree. 
• 40% had undergraduate degrees in scientific fields, specifically, physical or social sciences (but not mathematics, computer science, statistics, or engineering).
  

The survey showed that the most successful data scientists are those with substantial, deep expertise in at least one aspect of data science; statistics, big data, or business communication.  The authors concluded that data science is a collaborative and creative field, where the successful professional is able to work with database administrators, business people, and others to get data projects completed in innovative ways. 

What is citizen data science ?

One of the biggest drivers for a new type of data specialist is that there are simply not enough data scientists to satisfy the demand from virtually all sectors of the economy.  Gartner has recommended cultivating “citizen data scientists” which they identify as people on the business side who may have some undergraduate mathematics or social science background and who can be assigned to exploring and analyzing data with the appropriate software tools.  Last year Gartner predicted that the demand for citizen data scientists will increase five times more rapidly than that of the highly skilled data scientists Harlan Harris studied.

Software tools are critical for helping citizen data scientists find real insights and avoid simple statistical mistakes.  For example, a technology that Gartner identified is "smart data discovery" which is a next-generation data discovery capability that provides business users or citizen data scientists with insights from advanced analytics and helps them avoid some of the common statistical pitfalls.

Citizen geodata science

There is at least ancedotal evidence that there has been for some time a trend toward "citizen geodata science".  At the India Geospatial Forum in Hyderabad in 2014, I moderated a session on electric power.   One of the speakers was Arup Ghosh, Chief Technology Officer at Tata Power Delhi Distribution Ltd (TPDDL) who presented his perspective on implementing geospatial technology in a private utility.   The geospatial group at TPDDL had about 60 field personnel and 18 analysts and support staff,  none of which had an educational background in geospatial data and technology.  Twelve were electrical engineers and the rest were people with electric power experience.  All learned geospatial data management and simple analytics "on the fly".

At a recent GoGeomatics Social Jonathan Murphy gave a presentation about his experience working in Northern Alberta preparing terrain for seismic surveys.  Geospatial data and technology was used in all aspects of field operations.   The staff were experienced in seismic surveying, winter drilling programs, wildfire management, and road and facility construction, but had minimal education in geospatial data management and analytics.  They also had picked up enough geospatial knowledge "on the fly" to do their jobs.

These are two examples of what could be called "citizen geodata science" which I suspect is part of Gartner's citizen data science trend and is almost certainly growing more rapidly than traditional "geospatial data science" as geospatial technology is adopted into vertical industries.  The challenge is how to reach these people, perhaps through MOOCs, community colleges, conferences focused on vertical industries and including geospatial technology vendors, presentations and hand-on training, or vendor marketing to help them avoid the common pitfalls of geodata management and geoanalytics.

GeoBuiz 2016: Geospatial and consumer devices, reality modeling, big data and analytics

Reality modeling

In the future, the deliverables at the end of a construction project will not only include the completed building itself, but also a digital "reality model" of the structure.  The digital model will include digital as-builts comprised of georeferenced point clouds captured by laser-scanning or meshes generated from images captured with a digital camera on a smartphone.  That was the prediction of Bhupinder Singh of Bentley at the GeoBuiz 2016 Summit in Bethesda.  This is a vision that is shared by other professionals in the construction industry including Ron Singh, Chief Surveyor at the Oregon Department of Transportation, who sees it as dramatically changing the construction process from surveying through design and construction to operations and maintenance.

Consumer devices and professionals

Consumer mobile devices will increasingly be the platform for many professional activities which currently require specialized devices and a PC computing platform.  Bryn Fosburgh of Trimble pointed out that while currently the location accuracy of GPS equipped smartphones is 10 meters, centimeter precision is surely coming because all that is required is a relatively simple antenna redesign.  Bryn sees a continuum in these devices between consumer and professional applications.   But while consumers tend to blindly accept whatever is available on their smartphones, professionals using these devices will have to be aware of the limits of the hardware, what it can be used for and what not.  For example, while smartphones in the future will be capable of centimeter accuracy in ideal locations, the accuracy will be less where there is a lot of scatter in the RF signals from the GPS satellites.  To ensure reliable results a professional surveyor relies on multiple sources including GPS linked to an earth station and traditional total station measurements.  Increasingly we will require other professionals to be involved.  Bryn reported a rising demand for photogrammetrists, as a result of the growing volume and availability of high resolution imagery which needs professional to help interpret.

Platform as a service

For virtually all of the world's big IT companies including IBM, Oracle, SAP, and Microsoft and the world's big content companies such as Digital Globe, the cloud is central to their vision of the future.  Together with others in the IT industry Xavier Lopez of Oracle sees the future in platform as a services (PaaS), but there are challenges.  A major challenge is integrating data in different databases including traditional SQL databases from Oracle and Microsoft and nonSQL databases such as Hadoop.  Another challenge is semantics.  Frequently different disciplines use different terms for the same thing.  I have blogged about the CB-NL project in the Netherlands aimed at solving the semantics problem in the construction industry.  Another big issue that Xavier identified is data fusion without co-location - how do you enable efficient processing of a logically federated database when the data resides in different locations ?

Monetizing the cloud

But perhaps the biggest challenge is finding an appropriate business model.  It is not completely transparent how the cloud is going to be successfully monetized, though the trend is toward some form of subscription or pay-per-use.   An interesting model of the successful application of the cloud in the construction industry is Textura, just acquired by Oracle.  Textura’s cloud serves as a place in which all of the participants in a project from the funding, through design, to construction can come together and access the shared assets necessary to complete the project.  Textura’s network includes about 85,000 general and subcontractors who using Textura to collaborate on construction projects.  Textura handles on average about 6,000 projects each month. The company’s shared database, which hosts project budgets, contractors, invoices, payments, change orders and compliance documents, resides in the cloud and is accessible to contract administrators, contractors, developers, subcontractors, consultants, title companies, and banks.

Big data and analytics

One of big drivers for the cloud is enabling the extraction of information from the huge volumes of data (most of which includes location) which are now available.  I have blogged about the nearly 100 petabytes of earth observation imagery available from Digital Globe.  According to Dr Richard W. Spinrad, Chief Scientist at the National Oceanic and Atmospheric Administration (NOAA) , NOAA is expecting to have 160 petabytes of archived geospatial data by 2020. 

As another example, Jeff Jonas of IBM described a publicly available database which includes 600 billion transactions per day collected by telephone companies.  This data, which includes location, can be mined (after being anonymized) for very interesting and commercially valuable information.  Jeff Jonas described some examples - using it to analyze where people spend most of their time (their "pattern of life"), how many people go into a store, and how long they remain in a store. 

Dipanshu Sharma, Founder & CEO, xAd illustrated live some of the things his company, which is in the location-based marketing business, does with this type of data.  xAd can customize ads to target people who are inside a store ("There's a great deal on aisle 5"), just outside one ("Come on in, we have great deals on umbrellas") or close to a competitor's store ("Checkout XXXX, our prices are lower than YYYY") - the words are mine.

In all of these example, geospatial data and technology are key to the applications.  Jeff Jonas foresees that the apps of the future on your smartphone will require geolocation and will be so compelling that just about everyone will enable GPS location on their smartphones.

RICS BIM 2016: April 4th BIM mandate looms for the UK construction industry

At this year's RICS BIM conference the elephant in the room was April 4^th 2016, the deadline for all public construction projects to be BIM Level 2 compliant. This is the culmination of a five year process starting in 2011 focussed on developing a set of documents defining BIM Level 2 and providing best practices for achieving compliance.

This year's conference suggests that for many in the industry they are now wrestling with implementation issues and are reporting a shortage of BIM skills.  They are finding that BIM is really about data and are realizing that data management and analytics are essential to handling and interpreting the huge volume of data that is being generated. A shortage of data scientists to help them to do this has become an acute problem.

Transforming the UK Construction Industry in Five Years

The government's focus from the beginning has been on full life cycle BIM.  Back in 2010/2011, the Government's BIM roadmap was perceived to be very aggressive, mandating BIM Level 2 compliance for all national government projects after just five years. In the UK public projects comprise about 50% of the construction industry so this amounted to radically transforming the UK construction industry in half a decade. Put another way the construction industry had to move from thinking analog to thinking digital in just five years. 

The utility industry has been going through a similar transition in moving to what is called a smart grid and distributed generation.  Some think that this process will take 20 years.  But new technology and a younger generation steeped in digital is stepping in to make this happen much sooner than that.  Energy is going digital more rapidly than expected.  I think it likely that something similar will happen in the construction industry because digital enables all sorts of efficiencies that we often cannot even imagine.  Uber and Airbnb are classic examples.  Both are the largest or second largest companies in their respective industries.  Gartner has forecasted a similar type of company will emerge for managing energy flows in the electric power industry by 2020.

For 2012 the Government's strategy focussed on discovery through pilot projects, notably the Ministry of Justice Cookham Wood project, a £20 mllion project where it is estimated that BIM saved hundreds of thousands of pounds in CAPEX costs.  

At the 2013 conference David Philp,Head of BIM Implementation, Cabinet Office, UK Government, announced PAS 1192-2 a key document supporting the Construction BIM Strategy to achieve Level 2 compliance.  It specified requirements for achieving BIM Level 2, sets set out the framework for collaborative working on BIM enabled projects and provided specific guidance for the information management requirements associated with projects delivered using BIM.  David Philp made it clear that although the initial focus is on the design/build part of the lifecycle with the goal of saving 20% of Capex and reducing the 30% waste in most construction projects, the "largest prize for BIM lies in the operational stages of the project life-cycle".

At the 2014 conference Peter Hansford, Government Chief Construction Advisor, made a link between the overwhelming interest in BIM as represented by the sold-out RICS BIM conference and the industry’s enhanced interest and appetite for BIM. He saw a shift in emphasis from asking ‘’should we do BIM?’’ to ‘’how do we do BIM?’’  This was happengin inspite of the fact that as, Rich Saxon, the UK Government's BIM Ambassador for Growth, pointed out In a panel discussion on the future of BIM that some aspects of BIM Level 2 were still  incomplete. 

Deborah Rowland, Head of Facilities Management, Government Property Unit, Cabinet Office in the UK Government outlined the Government's motivation and plans for a program to improve the post construction handover and operation of newly constructed buildings. She described the Government Soft Landings (GSL) program which was intended to rectify this and was designed to put in place a legal, contractual, and technical framework (based on the BSRIA Soft Landings Framework)  incorporating building information modeling (BIM) to fix the problem by ensuring continuity thoughout the buiiding lifecycle from inception, though design, construction, commissioning, training and handover through to operations and maintenance for 1-3 years after handover.

At the 2015 conference seven key BIM Level 2 documents (PAS 1192:2, PAS 1192:3, BS 1192-4, CIC BIM Protocol, Classification (Uniclass), Digital Plan of Works / Levels of Detail (LoD), Government Soft Landings) which collectively represented the current repository of BIM best practices were announced.  Every speaker at the conference referenced these documents.

Seven major government projects using BIM, worth about £ 10 billion, have maintained records of the savings they have seen from using BIM. The measured benefits of BIM and factors associated with BIM range between15-20 %.   By 2014 the government estimated that it had achieved savings of £ 1.4 billion through BIM and related improvements in the 2013/2014 fiscal year.  But with full lifecyle BIM, the government saw an upside potential of 40%+ with the biggest benefits being realized in the operations phase of the building lifecycle.

This year it was announced that Government statistics show that in 2015 Her Majesty's Government saved GBP 3 billion on these demonstration projects as a result of gains in construction productivity related to BIM. 

Full Life Cycle BIM

But BIM Level 3 is where the gold is.  Level 3 will enable the interconnected digital design of different elements in a built environment and will extend BIM into the operation of assets over their lifetime. It will support the accelerated delivery of smart cities, services and grids. As the AGI has pointed out in a recent Foresight Report integrated BIM and geospatial technology will be fundamental for BIM Level 3.

Is the UK construction industry ready for April 14 2016 ? 

An industry survey in the UK in 2013 reported that 54% of respondents said that they were using BIM on projects.  Most of those not yet using BIM said that they would be within one to two years.  70 % of the respondents reporting using BIM said that BIM has given them a competitive advantage. But this year's conference suggests that most in the industry are doing BIM although for some it is not really clear what BIM Level 2 is and how one demonstrates compliance. 

Data and People

This year's RICS BIM conference was quite different from previous conferences.  The focus was heavily on implementation, not on progress in completing specifications or developing and working with 3D models as in previous RICS BIM conferences.

The event was led off by a panel discussion that set the tone for the rest of the conference - the focus was on data and people rather than 3D BIM models.  The industry people on the panel were patently up to their elbows in real world BIM projects including data, the analytics to make sense of it, and the shortage of the required skills (data scientists and people with BIM skills).

Alex Jones, Interserve, said that within the construction industry there is now a focus on data and how data is used downstream.  He said that clients (owners) are now aware of BIM and that both private and public procurements are asking for BIM.  But he qualified this - they are not so much asking for BIM (as in 3D models) as asking for data that can be used downstream. David Throssel, Skanska, added that deliverables now means data.

In the context of data, Simon Rawlinson, Arcadis UK, gave an example of utilities where smart devices and automated capture has resulted in a database of asset conditions, and which also includes geospatial data for geolocating those assets. This information allows utilities to move forward to condition-based maintenance where prioritizing maintenance and replacement depends on asset condition not the calendar.  Real-time monitoring is being adopted by virtually all industries.  For example, devices are being implanted in structures such as bridges to monitor distortions. He added that the leadership has not yet recognized that sensors and analytics enable a major change that will revolutionize asset maintenance.

Also the new digital economy is disintermediating certain skills. For example, digitalized construction requires fewer layers. Simon cited a KPMG study of disintermediation by technology. As an example, he mentioned data such as that collected by utilities from smart devices that allows owners/clients/utilities to do condition-based maintenance, which can lower field staff requirements significantly.

In the construction industry as a whole one of the major issues that emerged is a shortage of BIM related skills.Simon emphasized that now that we have lots of data, what we need urgently are data scientists to help make sense of it.  He identified this as a major problem because in his experience data scientists are as rare as “hens teeth”. 

David Hancock, UK Government Construction, said that construction industry still thinks analog. Nowadays we can all produce data, but analyzing it is the challenge.  We need people who can analyze to produce actionable items. He also identified BIM skills shortage as a major problem, and repeated that data scientists people are “as rare as hens' teeth.  He suggested that one solution is upskilling – training existing construction/bldg maintenance/utility folks in BIM and data science.  Another may be raiding the U.S. baseball industry where teams such as the Oakland Atheletics starting using analytics several years ago.  "Moneyball" analytics has been recommended as a strategy for the utility sector in the U.S.

In the question period a speaker argued that the 3D BIM model has been "massively overplayed".  BIM is really about data and collaboration, not about 3D. To maximize the benefits of BIM clients  (owners) have to be clear about what the model, and much more importantly the data associated with the model, are going to be used for. For example, if for facilities management (FM) then clients (owners) have to be specific about what data is required for maintaining and operating the building as part of defining deliverables.

BIM Level 2 for BIM Novices

What if you are a small company with no more experience with BIM than discovering what the acronym stands for ?  At RICS BIM 2016 there was a sequence of presentations by a group of mostly quantity surveyors (people who do quantity takeoff), all BIM novices, who came together to try to understand and share openly what BIM Level 2 means practically. (As one speaker pointed out, there is no legal definition of BIM Level 2 compliance.  They called their project KT4BIM, which you can find on LinkeIn. The objective is to answer the questions

• how do we do BIM Level 2

• how do we demonstrate compliance.

The target audience is BIM novices and for both the residential and commercial sectors.  The objective is to share all the material used and developed as part of this project with the industry.  More information about this very important project will be provided in my next blog post.

BIM Level 2 Documents

I've included this list just to show how comprehensive this set of BIM and BIM-related documents is.

Key documents

• BS 1192:2007+A1:2015     Collaborative production of architectural, engineering and construction information
• PAS 1192-2:2013     Specification for information management for the capital/delivery phase of construction projects using Building Information Modelling
• PAS 1192-3:2014      (Corrigendum No. 1) Specification for information management for the operational phase of assets using building information modelling
• BS 1192-4:2014     Collaborative production of information Part 4: Fulfilling employer’s information exchange requirements using COBie
• PAS 1192-5:2015     Specification for security minded building information modelling, digital built environments and smart asset management
• BS 7000-4:2013     Design management systems. Part 4. Guide to managing design in construction
• BS 8536-1:2015     Briefing for design and construction. Code of practice for facilities management (Buildings infrastructure)
• Uniclass 2015     Classification embedded within the NBS Toolkit
• Digital Plan of Work     Delivery plan embedded within the NBS Toolkit
• CIC/BIM Pro first edition 2013     Building Information Model (BIM) Protocol

Supporting documents

• CIC/INF MAN/S first edition 2013    Outline Scope of Services for the Role of Information Management
• Government Soft Landings    "Design for operation" to encourage better outcomes using BIM for built assets during the design and construction stages to ensure value is achieved in the operational lifecycle of an asset.
• AEC (UK) BIM Technology Protocol Version 2.1.1 June 2015    Practical implementation of BIM for the UK Architectural, Engineering and Construction (AEC) industry
• BS ISO 16757-1:2015    Data structures for electronic product catalogues for building services. Concepts, architecture and model.
• BS 8541-1:2012, BS 8541-2:2011, BS 8541-3:2012, BS 8541-4:2012, BS 8541-5:2015, BS 8541-6:2015    Library objects for architecture, engineering an construction

For the future of distributed energy watch the Hawaiian power utilities

Keep a close watch on the Hawaiian power utilities (HECO) because Hawaii has committed to 100% renewables by 2045 and 65% by 2030. That was the recommendation of Sharon Allan, CEO of the Smart Grid Interoperability Panel (SGIP) at a DistribuTECH2016 MegaSession on distributed energy resources (DER) in response to a question from the audience asking what should utilities do to get ready for GRID 3.0.

Hawaii already has 487 MW of solar PV capacity, 90% of which is residential rooftop panels. At  DistribuTECH2016, Talin Sakugawa from HECO and Matthew Shawver from in2lytics gave an insightful presentation on distributed energy generation (DER), big data and analytics at the Hawaiian power utility.  Hawaii has seen an exponential rise in rooftop solar PV. As a result since 2010 the utility has been experiencing a reduction in net load and revenue. The most dramatic drop in load is during daytime hours as is readily apparent by comparing the utility's load on clear and cloudy days which reflects the impact of residential solar PV.

Challenges facing utilities

The Rocky Mountain Institute (RMI) has published  an analysis of the economics of leaving the grid.  The central thesis is that solar PV and electricity storage enables consumers to leave the grid completely.  Solar PV has already reached grid parity in about 10% of the U.S. and declining PV prices suggest that this trend will continue.  However, without the ability to store electric power, consumers with rooftop PV still require the grid for nights and cloudy days.  The development of combined solar PV and batteries from Tesla and others promises to make solar + storage accessible for increasing numbers of consumers.  These consumers could form their own microgrid, either by themselves or with their neighbours and disconnect from the grid.  Alternatively they could become a source of dispatchable power and become an energy provider.  Increasingly they could find that either of these options is economically advantageous.  This has serious implications for local utilities because if this trend develops it will seriously erode the traditional utility revenue base.   It could also lead to a completely decentralized grid comprised of many microgrids.  From an operations perspective it increases the complexity of managing the grid compared to the centralized model in use today.

The other challenge for a utility with a high penetration of intermittent, distributed generation (DER) is load balancing, ensuring that generation meets demand. The German power industry, the U.S. Department of Energy, Hawaiian power utilities,  California power utilities, ERCOT, and many others are working to address this challenge.  HECO's presentation at DistribuTECH gave an insightful overview of what they are doing to meet this technical challenge.

Data sources

HECO collects a large amount of data from diverse sources. It is comprised of internal operational data plus data from esternal sources. It includes weather forecasts, data from customer sited PV, consumer/public resource data, phasor (PMU) data, renewables power quality, feeder data, irradiance meters, generator and substations power quality, SCADA, and vendor data such as forecasts and data from Independent Power Producers (IPPs). The data has widely different time resolutions ranging from real time such as PMUs sampling 30 times per second, SCADA reporting every 2 seconds, calculated gross load every 2 seconds, PV inverter data reporting every 5 minutes and aggregated monthly, weather forecast reporting every 15 minutes and aggregated daily, transformers collecting data every 15 mins and aggregated monthly, and smart meters reporting every 15 minutes and aggregated quarterly. Much of the data includes location. The data comes in different formats and has different levels of quality.

Fundamental to distributed energy are weather maps.  Maps of irradiance and wind velocity can be directly used to estimate or predict solar and wind generation in different parts of the islands.  Temperature maps are also important because the efficiency of solar panels depends on temperature.  The more accurate the weather forecasts are, the more predictable the generation.  Knowing that there it will be cloudy over the south of Oahu, but sunny over the rest of the island, or cloudy over the whole island, but with moderate to strong winds allows operations to estimate how much backup generation is required and where. 

EMS Integration

Data from new sources relating to distributed renewable generation are integrated with traditional sources in a distributed energy management system which enables operations and planning visibility into how distributed generation is impacting thte grid. It not only provides a view into the status of the grid (situational awareness) but also allows simulations to assess the impact of different future renewable generation scenarios on the grid. This helps to determine where backup generation may be required because high demand and low generation from renewables or curtailment when there is excess generation and insufficient load.  Location is fundamental to the analytics, because wind and irradiance varies widely over the islands.

Traditional input sources include SCADA and data from transmission, generators, protection, and so on. The new input data sources include GIS-based infrastructure models and data from the distributed generation sources. Also new is weather forecasting which enables predicting generation from renewable sources.  It also includes historical and actual weather reports, and satellite and other weather data.

HECO uses an IT platform from In2lytics (a spinoff from Referentia) which integrates all of these time series data with EMS, SCADA, CIS, and GIS and makes it available to operational users, planners, modelers and the public. in2lytics is a high performance time series database that is specially designed to provide instant data accessibility for planning and operational decision making.  in2lytics enables load, query, analysis using MATLAB, and sharing of the "big data" required to monitor and manage today’s complex electric grid. in2lytics has native high performance interfaces for MATLAB in addition to programming languages.  It translates data from different sources into its internal time series database. All data is archived in a spatially enabled time series database for future analysis. Matthew said that in2lytics is able to analyze large volumes of historical time series data very rapidly. For example, a longitudinal analysis on two years worth of archived data can be run in an hour or two.

Applications

Accounting for renewables requires a lot of data. Some areas that use this data include generation planning, load forecasting, distribution planning, and contract administration. One of the first applications is a customer facing web site called Renewable Watch - Oahu that reports the total renewables generation, solar and wind, on Oahu in real-time.

An example of an application that will be used for contract administration beginning next month estimates how much power is lost from Independent Power Producers (IPPs) in the case of curtailment. It uses weather information to estimate wind speed and irradiance meters to estimate wind and solar generation potential.

Another analytical application was used to study how solar PV variability affects transformer tap changes. This application allows HECO to relate frequency of tap changes to solar PV variability and to determine which transformers are most affected by solar variability.

HECO has a number of projects with the Department of Energy and other partners.

1. Department of Energy's Integrating System to Edge-of-Network Architecture and Management" (SEAMS) A federally-funded research initiative on high penetration grids which aims to streamline grid planning and operations for utilities in regions with high concentrations of distributed generation (DG) resources.

2. Department of Energy's Distributed Resource Energy Analysis and Management System (DREAMS) Department of Energy system to provide useful information about the distributed grid to grid operators.

   Partners: Siemens, Alstom, DNV, AWS Truepower, Referentia, Sunshot

   Advisory Team:SMUD, Southern California Edison, California ISO, TEP, APS, Kaua'i Island Utility Cooperative, Western APA

3. Making sense of synchrophasor data for utilities. The Synchrophasor Visual Integration and Event Evaluation for Utilities (SynchroVIEEU) with High Penetrations of Renewables. Accelerate the integration of synchrophasor information into production grade data visualization and analysis platforms/models. Leverage PMU capability at many substations – explore ways to tap resources and provide real-time visibility and real-time data. Make synchrophasor data accessible for efficient and reliable operations of a modern grid in light of high penetrations of renewable resources

   Partners: SEL, DNV GL, Referentia

4. Monitoring, planning, and modeling high PV penetration microgrid. Decision support for microgrid design and operation at Marine Corps Base Hawaii. The Office of Naval Research (ONR) has launched an ambitious program to demonstrate and evaluate energy technologies using Navy and Marine Corps facilities as test beds, known as the Energy Systems Technology and Evaluation Program (ESTEP). Program management is being handled by the Space and Naval Warfare Systems Command (SPAWAR) Systems Center Pacific (SSC Pacific). ESTEP was established in 2013. It brings together the Department of the Navy, academia, and private industry to investigate and test emerging energy technologies at Navy and Marine Corps installations. At present, ESTEP conducts over 20 in-house government energy projects, ranging from energy management to alternative energy and storage technologies.

Other projects include

• Remote light sensor technology at the Kahuku wind farm on Oahu to measure and forecast windspeed and direction to support reliability;
• Sonic detection and ranging technology to remotely measure wind speed and direction;
• Network of solar irradiance sensors to measure light and to make forecasts;
• EV Race to the Sun, a joint Japan-U.S. showcase event for electric vehicles,to take place on Haleakala;
• Battery storage technology development, including the JUMPSmartMaui project.

The ultimate objective for HECO is to become a company offering diversified services providing value to engaged customers, of course with satisfied regulators and sustainable costs and margins.  

Big data and spatial-temporal analytics

It was clear from this presentation that essential for HECO, or to any utility, with a high penetration of solar or wind are weather maps, forecasted, actual and historical because these enable the utility to project generation. The also allow utilities to analyze the historical record in conjunction with other data to discover trends that may help improve forecasting.  Collecting and analyzing temporal geospatial data is fundamental for distributed energy management.  This is in addition to the expanded application of geospatial technology for utility operations and planning in the smart grid era.