In order to design and implement a field area network (FAN) to provide smart grid coverage for a community, town, or major city, proper planning and coverage modelling is necessary. While there are many methods and approaches to conduct this sort of planning and modelling, some critical factors need to be addressed before realistic modelling can be accomplished. Realistic modelling is truthful and repeatable as well as proven and validated.
Too often we see modelling that lacks any credibility since it was completed with inappropriate tools, missing data, and by unqualified personnel. The visualized models that result are not much than fancy and expensive wall art. Therefore, getting the coverage planning and modelling done right upfront will save millions of dollars later. This is not normally an in-house skill set for most Utilities so selecting and hiring the right professionals is critical. While some Utilities do own planning and modelling tools of their own, training and currency in the proficiency of these tools is mandatory for them to be useful and justified. Some simple capability is always desired to run “what-if” scenarios before bringing in the professionals for the actual modelling exercises. Knowing what questions to ask and knowing how to understand the models generated is essential to your success.
Towers and Rooftops
One of the first steps is to determine what “height assets” are available for use for mounting base stations equipment and antennas. These height assets can be actual Utility owned communication towers, transmission towers, rooftops, or leased facilities from telecom operators, wireless carriers, broadcasters, internet companies, or tower infrastructure providers. The question is where are these height assets in relationship to the population to be covered and who owns them? Geocoding them onto a map is a great way to better appreciate them and how they can fit into your FAN design.
Rooftop space is at a premium these days and great care needs to be taken to mitigate interference from nearby antennas and services
(MJ Martin, 2009, All Rights Reserved)
The terrain varies dramatically and in some places this can be a challenge and / or an opportunity. Modelling your coverage over a high resolution terrain database is important, the better the resolution, the better the modelling. Even in places with very flat terrain, care must be taken in the modelling as multipath results when it rains and the land retains the water in pools or with saturated grasslands.
This is the study of the features that exist on the terrain. Morphology is from the Greek, and means, “the study of shapes”. In RF coverage modelling, it pertains to land features such as water, trees, foliage, structures (man-made and otherwise), wetlands, airports, roads, streets, low and high intensity residential, commercial and industrial buildings, office towers, bridges, statures and much more. All of these objects can obstruct, absorb, reflect, or diffract radio signals from the FAN. These items must be modelled as a layer over top of the terrain data.
Meter and Concentrator Locations
Smart meters in some countries are mounted on the exterior of the home at about 6 feet / 2 metres above grade (see image above). In other countries they are located inside the home often in some hard to reach place like under a staircase or in a basement. Regardless of where they are located, they never seem to be at the optimized location for supporting a sustainable a radio link. Most meter neighbourhood area networks (NAN) use mesh networking, some use star networking. We rarely see cluster tree networking in the NAN, but it does exist. In most scenarios, the meters aggregate to a concentrator (aka Regional Collector). In the early days of smart meter deployments, the concentrator was collocated within one of the meters. However, the performance was poor so the concentrators were relocated to where we see them today about 50 feet / 15 metres up high on a hydro or telephone pole, normally below the secondary. This height provides much needed improvement in connectivity to the backhaul networks, also typically designed as wireless networks using WiMAX technology. Sometimes landlines are used, but this is not the best option over the long run due to excessive OpEx cost.
The antenna “under the glass” within a smart meter is a poor radiator and spills energy in undesired directions
(Itron, 2011, All Rights Reserved)
Modelling tools fall into two key categories, RF propagation Modelling Programs and Mapping Programs. There are two ranges of tools based upon cost. There are a few excellent tools in the $100k range and some good tools in the $20k range. Beyond the tools, it is easy to spend an additional $10k to $100k for mapping data files for topographic (terrain) and morphology. Popular tools include Softwright’s TAP, Mentum Planet, Pathloss, and Splat!. Even Matlab is helpful in modelling.
Mapping tools aid in the visualization of the coverage by placing the signal model over streets, highways, buildings and permit coding of sites and features. Simple mapping tools like Google Earth are popular but can place your confidential information on uncontrolled servers operated by others, so privacy is not guaranteed. Stand alone programs like MapInfo and ESRI are preferred. Even MapPoint can work for some indicative modelling. There are a number of tools that aid in geocoding assets onto maps as well. Great maps help to effectively communicate and share complex modelling with others.
Again, by outsourcing to a qualified RF modelling company, you can expect to see only the best modelling tools used. The outsourced operators are usually professional engineers, armed with respectable maths degrees or doctorates in modelling. Statistics is a major part of modelling so if you are not keen to revisit these maths from your college days, then RF coverage modelling is not for you.
RF Modelling tools make use of specialized algorithms to predict how the radio signals will cover the land and work with the various morphology aspects that impede its coverage. Knowing and understanding the algorithm used for the modelling is critical. Algorithms are designed for a variety of jobs based upon transmit power, receive thresholds, modulation, FEC, frequency, range, environment and more.
A layered or tiered model that has terrain, morphology, maps, smart meters, and RF propagation modelling all overlaid helps to communicate the coverage story to the team and the senior leadership
(Sensus, 2009, All Rights Reserved)
The most popular RF modelling algorithm is the Longley-Rice model. It is popular since it is able to support a wide range of spectrum, from 30 MHz to 30 GHz. Other algorithms like the Bullington model and the Carey model are optimized for mobility. The Okamura model is ideal for land mobile radio design. There are many models and variations of these algorithms that try to extend the model in ways to enhance performance. Selecting the best model is important to match your needs. Consult the vendors of the modelling tools for recommendations.
Using these tools does not guarantee that the output will be successful. You need to appreciate that this is a process and it is an iterative one. RF modelling is a process of continuous refinement and updating. It is not unusual to run 5 to 8 passes to optimize the coverage.
Since the process is iterative, you can expect to make 3-5 passes at the design before beginning the build of the FAN. However, be prepared to continue the modelling after the build as there is always a need to fill in the gaps left from the surprises that inevitably effect your design. Best practices suggest 2-3 more iterative passes to backfill in the holes in your coverage. Know this from the beginning and appreciate that additional equipment will be installed for the backfill activities.
No modelling exercise is perfect. But, the goal is to get it as optimized as possible in order to trust the model and use it for subsequent deployments in other communities as you build out your network. In order to enhance the model, you should plan to perform drive testing. This is a process of equipping a vehicle with a series of test devices, antennas, and commercial GPS to record signal levels and build logs of the routes in the vicinity of your first base station(s). The results of the drive test are then compared to the modelling results and the model is then tweaked to match the real life data harvested from the drive test efforts.
Driving at night is best due to reduced traffic levels. Most drive test take three days and gather about 6 million data points.
(CelPlan, 2011, All Rights Reserved)
It is critical in some parts of the world to perform two drive tests, one in the spring and another in the winter to understand the impact of trees and foliage on propagation. In one project, we noted an astounding 4.6 dB difference from when the leaves were on the trees and when they were off the trees. Frequency plays a big part in this discussion, so this effect will not be consistent from one Utility to the next.
Once the drive test is completed, the data gathered will need to be cleansed as there will be errors. The GPS has dither and will result in what appears to be irregular driving as if you are driving across front lawns and over top of buildings. These errors need to be removed before the data analysis stage.
GPS dither creates errors that needs to be cleansed as a part of the data gather process. Statistical analysis will be performed on a subset of the master data set and will be statistically accurate if done properly.
(CelPlan, 2011, All Rights Reserved)
RF modelling can save millions, will speed up a project, and will result in a better level of performance and quality from the FAN. Mapping helps to visualize the coverage and communicate complex engineering ideas with the team, subcontractors, partners, vendors, and the senior leadership. Remember, RF can not be seen, heard, felt, tasted, or smelled, so it is a difficult concept for more than 80% of the population to understand as they are visual learners. So, visualization is the perfect answer to help others comprehend the coverage results even before the networks are built.
Michael Martin has more than 35 years of experience in broadband networks, optical fibre, wireless and digital communications technologies. He is a Senior Executive Consultant with IBM’s Global Center of Excellence for Energy and Utilities. He was previously a founding partner and President of MICAN Communications and earlier was President of Comlink Systems Limited and Ensat Broadcast Services, Inc., both divisions of Cygnal Technologies Corporation. He holds three Masters level degrees, in business (MBA), communication (MA), and education (MEd). As well, he has diplomas and certifications in business, computer programming, internetworking, project management, media, photography, and communication technology.