Integrating wireless into an industrial Ethernet application: Questions and answers
Audience questions were lively related to the Control Engineering March 10 webcast, "Integrating wireless into an industrial Ethernet application," falling into general categories of wireless lifecycle and applications, network design, products, weather, and standards.
The webcast discussed how to integrate two wireless systems into an industrial Ethernet application in two days. An industrial wireless network with a new supervisory control and data acquisition (SCADA) system is part of a 5-year, $47 million capital improvement plan at the Stamford Water Pollution Control Authority (SWPCA) of Stamford, Conn. The webcast covered wireless selection criteria including equipment reliability and capabilities, availability, ease of management, conformance to open standards, and integration with an existing Ethernet network from Daniel Capano, president of Diversified Technical Services Inc. of Stamford, Conn., and SWPCA vice chairman. Key points from Control Engineering research on wireless and Ethernet also were reviewed by Mark T. Hoske, Control Engineering content manager, also moderator for this RCEP-accredited webcast, which has an associated professional development hour for passing the quiz after the webcast and question-and-answer session.
Wireless lifecycle, applications
What do you see for supervised type control such as fire detection systems?
Hoske: Wireless does serve in motion control applications, such as for life-safety applications, crane control, and for large attractions. Redundancy can be built in. Sometimes technology capabilities can exceed standards organizations’ abilities to keep up. (See links below.)
How are you creating a mesh with two access points (APs)?
Capano: At this point, mesh redundancy for the backbone only exists where both APs provide service. Future implementations will add access points, a creating a full mesh design in line with the diagram shown (slide 33 in the webcast). The instrumentation wireless mesh network has five transmitters. Project details made live after the webcast have more information. [See links at bottom.]
How long is this new wireless system expected to last? The last system was only 8 years old.
Capano: Unsure which "last system" you’re talking about, but using standards should help extend useful equipment life. As for the hardware and software, we expect that software will be upgraded regularly and equipment will be upgraded to conform to newer wireless standards as the need arises.
Is there a concern that ancillary uses of the plant network (voice, video, asset care, etc.) eventually could use up the available bandwidth and cause delays in process response?
Capano: We have a good handle on present and projected use by staff. We will control access and thereby control our bandwidth "budget." At this time, however, it appears that adequate bandwidth is available to allow reliable voice and video traffic.
Would you use this for process control, that is, stop the filling of tanks?
Capano: Current uses are for monitoring. Future uses could include control, with appropriate redundancy.
Did you have to install towers for the mesh points, etc.?
Capano: APs were affixed to existing structures. While towers could certainly be used where appropriate, the typical plant affords convenient locations for elevated installations of APs.
How do you feel about the "move to the cloud" as it pertains to traditional on-premise networking hardware, such as wireless products that move the controller to the "cloud"? Such a design might increase risk, giving a manufacturer of networking hardware the ability to change the cloud infrastructure and make working hardware incompatible with updated new cloud structure/features. In the future, we may not have much choice.
Hoske: Consumers always can exert influence in standards organizations and in requests for proposals. Forced obsolescence generally has been less of an issue (perhaps to a fault) for industrial products and applications, compared to manufacturing IT products. Not changing technologies often enough often limits users’ efficiency. Sometimes letting something continue to work without an upgrade, just because no one wants to touch it, misses added features and capabilities that could significantly add quality, efficiency, and optimization.
While wireless is easier to deploy, I struggle with wireless optimization/speed (200 sq ft manufacturing facility with 10 APs that provide great coverage) mainly because of radio frequency (RF) interference and high AP usage, both of which cause communications rates to drop significantly for clients connected to an affected AP. WLAN optimization tools I’ve found can cost as much as $5,000 alone. Any recommendations on tools/tips for optimizing the WiFi speed due to interference reduction, radio power adjustments, etc.? I’m forcing 5 GHz capable devices to use the 5 GHz band, and I’ve disabled IEEE 802.11b support to try to optimize the 2.4 GHz band as well as manually set radio channels (1,6,11) on APs in 1-6-11 groups to try to minimize interference.
I also have our WLAN setup so that clients can roam from AP to AP and keep the same IP on their device. Some devices hold onto an AP much farther away, rather than switching to the closest AP; which causes communications issues on the device from AP to AP. This can be solved, I suppose, by optimizing radio power per AP. What tools do you recommend for such analysis and optimization as it pertains to roaming?
Capano: It would be advisable to do a full site survey to determine the sources of interference and to get a good idea of your propagation environment. Most problems with a WLAN implementation are due to electromagnetic interference (EMI), followed by attenuation caused by building materials or other obstructions. You’ve made some good changes in forcing devices to the 5 GHz spectrum and disabling 802.11b. Using the non-overlapping ISM channels is also a good basic design. Using lower power on multiple APs will give better coverage and throughput than using fewer, higher powered APs. I would suggest renting a spectrum analyzer to root out and eliminate the sources of interference as a first step. Project details made live after the webcast have more information about the wireless network management interface. [See links at bottom.]
Do you recommend a different frequency band between the sensors and the remote AP and both Aps?
Capano: Two networks were chosen in this case because one network couldn’t provide features and capabilities sought. IEEE 802.11 devices are not available at this time, so a decision was made to incorporate 802.15.4 devices as primary sensors; also communications and instrument communications used different methods of modulation. Different frequency bands were used to provide both local and backhaul connectivity. The ideal wireless network would be "flat" and use a single wireless standard and the two unlicensed frequency bands.
Great presentation, Dan. I work for a propylene plant and worked for an automation vendor for 4 years as a field service engineer. A good practice is redundancy on the network. One vendor uses a fault tolerance Ethernet for redundancy. What if one of the APs failed? Have you considered using redundant APs?
Capano: Thank you. The primary reason we chose a mesh architecture is because of its inherent redundancy. The network is self-forming and self-healing. In a well-designed network, mesh nodes with overlapping service would serve multiple areas. If one AP fails, other APs will attempt to recover the lost network segment and continue operating.
How well can a wireless network perform for fast data exchange that monitors critical processes? We recently migrated an old programmable logic controller (PLC) using Modbus TCP wired communications. I integrated it into a control system via OPC, but signals for a dehydration unit are for monitoring, not control.
Capano: Depending upon the implementation, wireless systems can exceed the speed of wired systems. This depends upon many factors, such as design, EMI, and path loss. For critical systems, particularly at this juncture, a wired backup should be used.
How much does an industry need to depend on public wireless carrier services for data transmission?
Capano: It comes down to cost. Using a common carrier introduces a recurring cost component, while using an on-site wireless system requires only the initial capital cost. No wireless carrier was used for these applications, though that can be an option, again, at an additional recurring cost.
The application example shows how wireless has been implemented in a wastewater treatment plant, and apparently it works very well. Have there been many implementations in manufacturing facilities, such as automotive, food and beverage, etc.? If so, are there problems with interferences of nearby motors, variable frequency drives, etc.?
Capano: I addressed this briefly when answering another question, in the verbal Q&A portion of the webcast. Wireless is used extensively in practically every industry except the public water and wastewater industry. There is always some instance of EMI; a proper site survey and system design can drastically reduce the effect of EMI. It is possible to shield some electromagnetic interference sources to reduce or eliminate their effect on the wireless system.
Hoske: And newer VFDs often create much less interference.
What software/hardware did you use for the site survey of WiFi coverage?
Capano: Project details made live after the webcast have more information. [See links at bottom.]
Maybe I missed it in the presentation, but what wireless mesh was used? Are you allowed to specify the brand/model of wireless hardware used for this WLAN project?
Hoske: RCEP webcasts avoid product mentions. As noted, project details in a separate article made live after the webcast provide more information about products used. [See links at bottom.]
What is the battery life on that sensor? What power source(s) were used for the instruments? Hardwired? Battery? What are advantages / disadvantages?
Capano: Power over Ethernet (POE) was used at both 802.11 AP locations. The primary sensors either scavenged power from the instrument loop or used internal batteries. According to the manufacturer, battery life can be as long as 9 years, based upon update frequency.
Hoske: Power scavenging devices can extend that, which can be especially useful in areas where access is difficult.
What type of software and/or IT engineering was used and what was that cost?
Capano: Project details in a separate article made live after the webcast provide more information about products used. A thorough knowledge of both wired and wireless technology is highly recommended to achieve a proper system design and implementation. Cost varies by your relationship with the vendor. In many cases, the vendor may provide equipment free on an evaluation basis, then charge only when the user or system integrator completes a successful evaluation and decides to purchase the equipment. [See links at bottom.]
What did you do to reduce security risks (eavesdropping)?
Capano: By manipulating AP power output to manage the coverage area, we were able to limit propagation to within the facility fence line. While not a security method, we will also disable the service set identifier (SSID) broadcast to reduce the ability of a casual eavesdropper to intercept our signal.
Do you use firewalls at the AP?
Capano: In this application, we did not use a firewall at the AP. Please note other security measures discussed here and in the webcast.
Are there security concerns with BYOD? Can someone capture and walk out with sensitive data?
Capano: The reason we specified the 802.11 open standard was because of the robust security specified. Any device, mobile or otherwise, that conforms to this standard was considered secure for the purposes of this project. To your second question: If an authorized and authenticated device has access to sensitive data, then it is possible that data could be stolen; this is, however, no different from any other system, whether it is wired, wireless, or shoved into a briefcase.
Are wireless devices more secure than wired devices?
Capano: Wireless devices conforming to IEEE 802.11 are encrypted using 128-bit Advanced Encryption Standard (AES), based on the Rijaendal algorithm. It is estimated that a 10-character password could take as long as 30 years to crack. Wired systems use similar encryption algorithms, but are much easier to hack into or compromise.
How well does the wireless work in all weather? How does rain/snow affect the reliability?
Capano: We experienced no performance problems at all during this very harsh winter. At this range, rain/snow fade were not a consideration.
Are the frequencies of the RF signals immune to sun spot activity?
Capano: No more or less than any other electromagnetic phenomenon.
Do you see any emerging or concrete standards for wireless communications?
Capano: IEEE 802.11 has become the de facto WLAN standard. We are now at the first rollout of 802.11ac, which will boost aggregate throughput to about 7GBPS using multiple radios and 160 Mhz wide dynamic channels. Further iterations under development will be using the 60GHz spectrum for high-speed, short-range connectivity; another being developed will use the "white space" in over-the-air television transmission signals to transmit data at relatively high speeds. The 802.11ac standard will be rolled out in several waves, so we anticipate it will be around for a long while. Hoske: www.controleng.com/ce-research Mobility, Ethernet, Wireless research from Control Engineering shows use of various wireless protocols.
– Edited by Mark T. Hoske, content manager, CFE Media, Control Engineering, email@example.com.
Also from Control Engineering, see:
The archived webcast: Integrating wireless into an industrial Ethernet application
Related wireless cover story: Integrating wireless with wastewater
Two more wireless applications—Case studies: Wireless remote monitoring