This guest post is from Daniel Capano, a consultant specializing in Instrumentation and Control Systems in the municipal construction industry in and around New York City. He is the owner of Diversified Technical Services, Inc. of Stamford, CT. Dan is an advocate of the use of wireless technology in industrial communication networks, presenting regularly to professional and industry organizations and conferences on the use of WLANs as a viable replacement for costly wired networks.
This is not as easy as you would think. In the SOHO and carpeted world, wireless is a given; any network administrator or small businessman worth his salt will almost instinctively gravitate towards WLANs. The cost savings are huge, for starters, and the mobility, utility and efficiency gains afforded allow for an almost immediate ROI. One would think that wireless is ubiquitous – not so. There are industries and facilities that are solidly, even staunchly, wired-only. These are usually older facilities with little incentive for changing over or upgrading – or so they believe.
Unlike many readers of this blog, I am not wholly a wireless professional. I split my time between my primary calling, Instrumentation and Control Systems, and their attendant, and varied, communication systems. These systems range from simple, two-wire loops to sophisticated multi-protocol mixed media networks. All are installed in industrial settings and can rightly be called “Industrial Control Systems” (ICS). As one could expect, these environments are extremely hostile to physical media. Even though the media is usually enclosed in a metal conduit, toxic and corrosive gases eventually find their way into supposed closed system and gradually deteriorate the insulation or jacketing, finally leading to failure.
The cost to replace these conduit and cable systems can be huge. In my neck of the woods, New York City, the cost to run cables and install conduit can run to over $1000 per foot. In an explosive environment, that figure can double. There is obviously a need to avoid this cost. But there are several factors to the resistance for a changeover, not the least of which is organized labor. The chief obstacle to upgrading to wireless is education: most people, even the most learned, look askance at wireless as a serious communications medium. They know wireless as the way their daughter texts her friends, or how they can get on the internet at home. It is an exotic concept that a wireless system could be used for anything serious – as serious as monitoring industrial processes.
To be fair, wireless has suffered from a lot of bad press. All of us reading know this is all in the past, but no one, and I mean no one, wants to be the engineer who specified a wireless system only to have some neighborhood geek break in and start turning pumps on and off. This is the perception: the classic malicious hacker in a van. Try as I may, I cannot convince some very intelligent people that WLAN security is incredibly sophisticated and robust – they can’t understand how something in the air can’t be just snatched up and used against them. Consider also that many of the responsible players are older – I have found a very real and palpable generational bias towards the use of this technology. They understand wiring, they can hold it in their hands, RF is black magic. I took it upon myself to clear things up a bit.
In 2014, I undertook to demonstrate the viability of using 802.11 wireless to monitor a remote process area in a wastewater treatment plant in Stamford, Connecticut. The plant is a publicly owned facility and I sit on the governing board. We had interminable meetings putting together a five year capital improvement plan that would upgrade the various process components (mostly big tanks full of human waste) along with attendant control and communication systems. The control and reduction of short and long term capital costs was a major concern because we wanted to stabilize user rates over the long term. One of the largest expenses was the physical wiring plant, which was deteriorating for many reasons I prefer not to go into. Wireless technology presented itself as a possible answer to this problem.
I designed a system using two dual band access points, one AP would be mounted outside of the control room, while the other would be mounted at the remote process area. The plan was to use the ISM band for local connectivity with mobile clients and process instruments, and to use the UNII band channels for the backhaul to the control room. Surveys showed that there would be no problem with propagation over the 330 foot distance we had to traverse. With the assistance of the plant electrician, we had the wireless network up and running in less than half a day. But we had a small problem: the instruments used to monitor the process were not Wi-Fi capable.
Enter 802.15.4, you may know it as Bluetooth. In this iteration, however, it is called “WiHART” Without getting too far into the weeds, WiHART is the wireless implementation of an old wireline protocol called “HART” or Highway Addressable Remote Transducer. This was an instrument level network, however, and we were looking for more utility from the wireless network, which I will get to a little later. the use of WiHART required that we do a protocol translation from 802.15.4 to 802.11. We scratched our heads for a few minutes and then, with the assistance of the instrument vendors, came up with what we called a “translational bridge” which did the conversion seamlessly. After cobbling together a few components, we were in business – the process data began populating the screens in the control room.
It really was that simple. We had actually tried to prove the WLAN would NOT work, but were very pleasantly surprised by the results and now WLANs will be included in the current plant upgrades. However, not everyone was pleased. I mentioned the generational bias; the older staff couldn’t understand why we needed “this Wi-Fi crap”, while the younger staff practically begged for access. Remember the distance covered? The 330 feet between the process area and the control room was a convenient number to work with – it is the maximum length of a single wired Ethernet segment (100m). This also gave us a very good apples-to-apples comparison for costing purposes. Also, placing two APs at that distance gave us wireless coverage over sixty percent of the plant area (1220 x 800 feet). I gave several of the staff credentials and told them, in effect, to “break it”.
They did their best, but the system held up very well. Available bandwidth was adequate. Video performance while roaming within coverage was solid. Process data was reliable and accurate. We declared the test, the “proof of concept” to be a success, despite all of our misgivings. To my knowledge, we are one of the first facilities of this nature to attempt this innovation. We were also interested in other benefits of a plant-wide wireless network. Being as the process data consumed negligible bandwidth, the idea of using VOIP (VoWIFI) for intraplant communication was discussed and will be explored. The general rule of thumb for process data is the faster it changes, the more sampling required for useable data, hence, more bandwidth. Water treatment processes change very slowly, leaving a lot of room for other utilization.
We envision that we will use the wireless network for multimedia delivery, including voice, and for asset tracking, remote security, remote document retrieval, mobile worker platforms, and a whole raft of other time saving techniques. We expect an increase in productivity and efficiency, resulting in lower operating costs. Our staff is looking forward to being able to monitor and control any plant process from their tablets, anywhere in the plant. This is the promise of WLANs as implemented with Industrial Control Systems.
The use of WLANs also allowed for flexibility in placing instrumentation. For a wired instrument to be moved to a more favorable monitoring location, the entire installation would have to be removed and rewired, at considerable cost. With a wireless instrument, it simply has to be unbolted and moved to the new location, probably never powering down or disassociating. The network and control system can also be scaled up almost indefinitely using a mesh architecture, which is particularly well-suited to ICS communications. In this situation, a mesh architecture offers the redundancy and fault tolerance that is required and fits nicely with the prevalent Distributed Control System Architecture that most facilities employ to control their processes. It is apparent from our analyses that the cost savings for future control systems expansion would be enormous if we used WLANs instead of conventional wired networks.
Now a word to the relative costs of implementing a WLAN. First, the capital expense, or short term, up front cost. The entire WLAN cost a staggering $5600, and took less than a half day to put into operation. By comparison, installing a single Cat 6 cable in a conduit, in a trench to the process area would have cost over $34,000, and would have taken two weeks and caused considerable disturbance to the landscape in order to install. In addition, the wired network would have connectivity at only the two ends of the cable, whereas the WLAN provided coverage to about sixty percent of the plant. It was truly a no-brainer.
As to long term costs: maintenance costs on either wired or wireless are negligible, however, wiring and conduit deteriorate over time; RF does not. To replace wiring and conduit is very costly in both time and money; for a WLAN, simply replacing an AP is all that is really needed. Add to this the fact that technology costs trend down, while labor costs trend up. Payback on a wired system is measured in years; WLANs typically pay for themselves before installation and revenue service, by virtue of labor savings alone. Return on investment is measured similarly: wired systems are designed for the maximum expected utilization and therefore cost is inflated. ROI is typically spread over some method of utilization, and never quite achieves the ROI desired, or is not considered. However, WLANs realize ROI much faster due to accelerated adoption by users: as each new user accesses the WLAN, the cost per user goes down.
I have been a wireless partisan ever since this experiment and regularly evangelize about the virtues of WLANs. My audience is typically made up of engineers and technicians with little, if any knowledge of the technology.
Which brings me back to my original point: the largest single factor in preventing an even more widespread adoption of WLANs is education.
More about Daniel: Dan has been published in several peer reviewed journals and has published an Industrial wireless communication blog that can be accessed at http://www.controleng.com/blogs/industrial-wireless-tutorials.html. The proof of concept project described above was the cover feature in the February 2015 issue of Control Engineering magazine (http://bt.e-ditionsbyfry.com/publication/?i=245749). Dan holds a CWNA and a CWSP, along with an FCC GROL and several vendor specific wireless certifications.