Brad Staley PE, Mark Gross, and Grant Baldwin
Salt River Project, Vaisala, and AES Indiana
SRP (Salt River Project) and AES Indiana evaluated a mobile gas-in-oil monitor manufactured by Vaisala (model OPT100M). The value proposition being evaluated was to find a tool to provide online (near real-time) DGA data for rapid response to lab-based gas results/alarms that required further analysis, that could also be compared to other factors, such as load, temperature, and other environmental conditions, for fault evaluation, better maintenance planning, and more cost-efficient operation. Two case studies will be discussed in this paper, one from Salt River Project (SRP) and one from AES Indiana which set out to achieve similar online DGA monitoring goals.
Case 1: SRP evaluated a mobile online gas-in-oil monitor (model OPT100M) manufactured by Vaisala to determine its viability and function to give SRP’s Apparatus and Substation Maintenance Engineering groups a way to test and monitor gassing power distribution substation transformers (22 and 28 MVA). In particular, SRP wanted to evaluate the setup, installation, and commissioning time, the software capabilities and hardware performance, as well as the general ruggedness of the OPT100 multi-gas mobile platform. Additional goals were to look at best practices for I/T, cybersecurity, and communications, to establish a future solution and procedures for the evaluation of “sick”, gassing distribution substation transformers.
Case 2: AES Indiana, formally known as Indianapolis Power & Light (IP&L) tested the Vaisala OPT100M mobile online DGA device for use on a questionable gassing transformer. This case study shows how AES Indiana used the Vaisala OPT100 Mobile unit, combined with the Delta X TOA™ analytical software to identify an issue, analyze the data and then plan the required maintenance steps before a serious fault or outage occurred, in order to prove the value of mobile online DGA monitoring. Based on this evaluation, AES Indiana approved the use of the OPT100 in its fleet, both for mobile rapid response to DGA alarms as well as fixed online DGA for critical asset online monitoring.
SRP Case Study: Using Mobile Online DGA to Support DGA Analysis of Gassing Distribution Substation Transformers
The Vaisala OPT100M unit was evaluated to give SRP’s Apparatus and Substation Maintenance Engineering groups a way to test and evaluate a mobile online DGA unit for monitoring gassing in power distribution substation transformers (22 and 28 MVA). In particular, these groups wanted to evaluate the performance of the NDIR (non-dispersive infrared) technology of the monitor, the setup, installation, and commissioning time, the software analytical capabilities, the hardware performance, the ruggedness of its mobile platform as well as review and define best practices for I/T, cybersecurity, and communications related to using online mobile DGA for “sick”, gassing distribution substation transformers.
The Vaisala OPT100M Transformer DGA Monitor Pilot R&D Project was initiated to perform dissolved gas analysis (DGA) remotely and in near real-time for distribution substation power transformers that are gassing. The mobile online platform included a small two-wheel trailer, with the DGA measurement and power systems mounted on top, that can be laid flat for transport, to allow ease of transport from one gassing transformer to another.
NDIR (non-dispersive infrared technology) based on light absorption is the core of the DGA monitor’s technology and requires no consumables, so it can be run 24/7 and will provide a new measurement every 70 minutes. The system also includes autocalibration for the measurement system, which eliminates the need for recalibration and ongoing maintenance resulting in a ruggedized unit that can be transported from site to site. As an added benefit for mobile measurements, when the unit is initially powered up, it starts an autocalibration on oil process as well, looking for disturbing gasses to ensure better accuracy and long-term repeatability.
Based on this potential benefit, SRP’s Substation Maintenance Engineering (SME) was asked by SRP’s Apparatus Engineering if they could utilize a mobile online transformer gas-in-oil monitor. SME decided that they would like to test/evaluate (for 90 days) a mobile unit for monitoring gassing distribution substation power transformers to evaluate it.
This transformer DGA monitor would be located where Substation Maintenance Engineering needed to monitor dissolved gases in transformers where severe gassing was present. The Grasmoen Bay 2 transformer was selected as the mobile DGA monitor evaluation site.
Funding from SRP’s R&D group was approved for the installation, commissioning, and evaluation of the DGA monitor.
Vaisala also agreed to provide SRP with a software that would show trending of any O2 leaks in the transformer tank, as leak detection was an SRP requirement for a DGA monitor, and the Vaisala unit offered a Total Gas Pressure measurement for this.
A cybersecurity questionnaire was also satisfactorily completed by Vaisala before the OPT100M monitor was installed.
The R&D project scope was developed for the transformer DGA monitor evaluation that included the following:
- Install/commission the pilot Vaisala OPT100M mobile dissolved gas-in-oil monitor at Grasmoen bay 2. Apparatus and Substation Maintenance Engineering to evaluate/verify the commissioning.
- Evaluate performance of NDIR (non-dispersive infrared) light absorption technology of the DGA monitor and its integrated auto-calibration features.
- Assess performance of Vaisala monitoring software – including installation of gas pressure trending (new feature) measurement to detect oil/air tank leaks allowing the ingress of O2/N2.
- Assess performance of hardware (pump, electronics, and sensors) in an SRP substation high-temperature ambient environment.
- Evaluate ruggedness of mobile platform for use with other “sick” gassing distribution substation transformers.
- Define initial threshold and trending alarms for engineering and dispatching center alerts.
- Determine design changes to be made to the trailer to meet SRP requirements.
- Determine monitor “best practices” for I/T, cybersecurity, and communications.
- Test the OPT100M’s performance for user-friendliness, alarm and communications performance, and data analytics/output/integrity.
Transformer DGA Monitor Installation and Evaluation
The installation and commissioning of the Vaisala OPT100M mobile DGA transformer monitor was completed in about 3 hours. It took about three days for the monitor to complete its initial auto-calibration-on-oil process.
Once the calibration process was completed, the individual gas levels were stable and remained that way for more than 90 days until the monitor was removed from service. For example, ethylene (C2H4) was measured at 38 ppm, ethane (C2H6) was measured at 470 ppm and methane (CH4) was measured at 380ppm. These are averaged values. The stability of these measurements indicated that the performance of the NDIR (non-dispersive infrared) technology of the DGA monitor was working well.
DGA oil samples were taken at biweekly intervals throughout the 90-day trial period. The historical lab-tested DGA oil sample DGA Data was compared to the DGA monitor data.
The historical DGA oil lab test results are summarized below:
The August 2020 lab-based DGA test results show where the gassing levels were before the monitor DGA pilot started. Higher than normal levels of methane and ethane were present in the transformer.
For comparison purposes, a snapshot of the gas-in-oil measurement data from the OPT100M is shown below:
Again, you can see that the values of methane and ethane are higher than normal. Notice that the results from the lab-tested DGA oil samples correlate well with the online DGA monitor’s measurement data.
The OPT100M uses a magnetic drive gear pump to move the oil through the monitor. There were no problems found with it. The DGA monitor used high-quality valves and was maintenance-free.
The DGA analytics from the monitor, DGA oil samples and TOA test results all stated that methane and ethane levels were high (see Appendix A) indicating a gassing status equal to four. (This means that the transformer is nearing the end of its life.) The Duval triangle indicated a “T1” status meaning that a low range thermal fault was ongoing.
The total gas pressure (including Oxygen and Nitrogen) was measured and was stable indicating that there were no leaks (either vacuum or pressure) in the transformer’s sealed system.
The DGA monitor uses a user-friendly clear, browser-based interface. Access to the monitor’s homepage was difficult at first because of SRP’s communications through multiple security firewalls. Some bugs were worked out and eventually all the necessary users were able to easily access the monitor dashboards and data remotely.
Vaisala configured a DIGI modem for communications during installation. It worked well. Apparatus Engineering asked SRP’s telecommunication group about using SRP’s fiber area network (FAN) connection that would be more secure than the Vaisala-provided DIGI modem. Since the cost for the FAN network connection would be significantly less costly to design, procure and install for future DGA monitor installations, a decision was made to attempt a cyber secure communications connection to the DGA monitor in this way. After a wireless communications tower signal strength survey and some adjustments by the I/T and telecommunications groups (including using the configurable Ethernet port zero to access the webpage set up via the user interface IP so the radio setup could be completed and making a policy change in the firewall which would directly allow the network connection) to get a secure SRP network connection, a successful connection to the SRP FAN network was made. A local PC was able to collect data on the DGA measurements too. This was a significant achievement for the mobile DGA monitoring R&D project.
The Transportation group inspected the trailer for the mobile DGA monitor and overall felt it complied with the SRP trailer standard. That’s aid, they made a few recommendations for the Vaisala monitor’s trailer for future units. They included:
1. Make sure data plate is not blocked.
2. Trailer safety chains should be 3/8” Grade 70 chain.
3. Install Buyers 3” ID drawbar hitch in lieu of ball coupler.
4. Hitch height needs to be adjustable 23.5” – 32.5”
5. Equip with electric trailer brakes and Hopkins Engager PN 20099 Breakaway System. 6. Equip with Pollak 12-706 7 Way RV plug and wire.
6. Equip with James King Model 300 registration holder
The performance of Vaisala monitoring hardware (electronics and sensors) in the Phoenix, Arizona substation high-temperature ambient environment (pilot started with daily high temperatures exceeding 105 degrees F) and software met expectations. The monitor was user-friendly and had useful analytical tools and output data.
From SRP’s perspective, the Vaisala OPT100M is a rugged, mobile DGA monitor with features not previously available on the market. This DGA monitor allows online monitoring on a mobile platform and effective online diagnostic evaluation of “sick” gassing distribution substation transformers identified by lab DGA sampling and other DGA analytical software, like TOA. The OPT100M DGA monitor has quick setup, installation, and commissioning capability and has design features that should enable a longer DGA monitor lifetime even in the hot desert ambient. The monitor also can be integrated into the SRP fiber-optic area network (FAN) using secure wireless communications.
In conclusion, the SRP Apparatus and Substation Maintenance Engineering group was very satisfied with the OPT100 mobile online DGA device from Vaisala. It was easy to set up and install, had a very friendly user interface and acceptable communications protocol and security for the IT group, and provided reliable DGA and leak detection (O2) data for trend analysis.
AES Indiana Case Study: The Value of Mobile Online DGA to support DGA Analysis of Questionable Gassing Transformers
AES Indiana has around 350 transformers in its power system but has only 3 fixed multi-gas DGA devices and 15 singles gas monitors. The rest of these transformers are sampled and lab-tested, either quarterly, bi-annually or annually depending on the perceived need. But the majority are sampled annually. This sampling data is fed into an analytical software tool, Transformer Oil Analyst (TOA), which evaluates the health and identifies transformers that may have developing faults.
In particular, AES Indiana identified a 50-year-old transformer showing signs of gassing and potentially degrading health. It was a 1961 step-up transformer, 230-69kV with a 67 MVA rating. This transformer had historical gases from lab results that were ranked with TOA as questionable and requiring deeper analysis to determine if the gassing was from a thermal fault that continued to develop or was from other isolated events. Sampling was providing insufficient data, so the AES maintenance team wanted to see how the gases were behaving when compared to other real-time factors, such as load, and temperature swings.
The question for the AES Maintenance team was to decide if this was an issue where the transformer had an ongoing fault that was continuing and seriously degrading the transformer health or was it gassing due to separate isolated events that were not ongoing?
Bigger picture – the AES Indiana maintenance team felt they needed another effective way to respond to these lab sample-based alarms, as well as those that came from single gas alarms, rather than just more sampling. When the lab results indicated a fault condition arising or that a gas alarm limit came in from the single gas monitors, there was no clear way to address them for a more in-depth analysis. On top of that, there was data from the PI Historian on H2 trend alarms.
While the Delta-X Research TOA4 decision support tool was able to give rankings on abnormal conditions, the company decided they needed a tool to address the need for real-time data for better analysis of gassing issues and needed a way to prepare next steps with their maintenance group.
AES Indiana had been evaluating mobile online DGA solutions and decided to purchase an OPT100M Mobile online DGA and moisture monitor from Vaisala. They saw this tool as potentially addressing the need for a “middleware” solution to gather live data. (See Figure 3.) But there was the question about how easy this type of a system would be to install and move and how reliable were the gas measurement results would be.
The secondary goal for AES was to validate the cost-benefit of mobile online DGA systems and how they would integrate with their existing monitoring solution (TOA).
This case study shows how AES Indiana was able to confirm both the value of a mobile online DGA solution and, just as importantly, better understand the condition of the questionable gassing transformer.
Transformer DGA History
AES Indiana uses TOA from Delta-X Research for assisting with tabulating DGA lab results and ranking transformers for further monitoring or maintenance actions. In particular, the TOA had identified in this 1961 Generator Step-Up Transformer concerning gassing events starting in around 2002 that had been continuing to sporadically accumulate in short bursts over time. (See Figure 4.)
There was not much DGA data prior to 1995, but it appears to have been stable. Based on an assessment of the current risk and perceived health of this transformer, in particular, AES Indiana decided to put the OPT100 Mobile online monitoring device from Vaisala in place to verify the situation and decide what corrective actions may need to be taken. This decision was supported by the lab-tested samples. The TOA software analysis had shown this transformer was having high cumulative energy (NEI1). Both the hydrocarbon and carbon oxides NEI as well as the carbon oxide ratios were showing concerning periods. (See Figure 4.) But even more concerning was the high hazard factor in event 11. That NEI-CO event – at 1.14%/year indicated a significantly higher risk as compared to background risk (see Figure 5). Based on this analysis and cross-referencing the trending to the IEEE C57.104-2019 (Annex F) both the NEI-HC and NEI-CO are increasing along with the CO/CO2, which would indicate that the paper insulation in the transformer is degrading and that continued oil sampling should be performed as the transformer gassing is trending upward. Further supporting this theory, the TOA Duval Triangle analysis was showing a potential T2 fault2, indicating that there might be a thermal hotspot. Given the NEI-CO was also trending up, it indicates paper insulation is involved which was raising concerns at AES Indiana, as the degree of carbon oxide gassing alone is enough to conclude that the paper was being affected in the previous gassing event (see Figure 4).
Moreover, the gassing status was at 4. Gassing status is a short status code similar to IEEE status codes but defined differently. Gassing status 1 is a transformer that has never gassed before. Gassing status 2 is a transformer that has a gassing event in its history but is not actively gassing. Gassing status 3 and 4 is a transformer that is actively gassing. Gassing status 4 are considered higher risk than Gassing status 3. This is because each gassing transformer has an estimated risk of failure from the Hazard factor. The 90th percentile of observed Hazard factors is used to denote 3 vs 4 as high risk. The Hazard factor estimates risk of failure from a large population of transformer failure cases and the transformer’s rate of gassing.
Based on this analysis, the decision was made by AES Indiana to install the OPT100 mobile online DGA monitor so AES could get real-time data and run that through the TOA4 system for comparative analysis. Again, the goal was to determine if this was an issue that was continuing to produce higher gas levels or was gassing due to an isolated event.
Mobile Online Monitoring Device Installation
The OPT100 mobile online device was installed on the transformer in February of 2020 just south of Indianapolis. The device was left in service for 4 months, gathering DGA data for analysis. During the installation, the transformer was left energized during installation as the connection points were low enough on the transformer to be reached safely. The installation and commissioning were completed by 2 people working for a half-day, which included setting up the remote comms. No special equipment or connectors were needed, nor was any proprietary software required. Just a laptop and a browser. (See Figure 6.)
The OPT100M mobile DGA monitor was set up using a Secomea modem running on the Verizon network, which provided the firewall and IT security layer for the wireless communications channel. The OPT100M was also protected with Username/Password protection. The AES Indiana staff used this connection to access the OPT100 software with its analytical features.
This OPT100 device was also connected to and fed data into the TOA software utilizing Monitor Watch from Delta-X Research which is an add-on option to TOA. This connection was made through a cellular Ethernet connection. TOA with Monitor Watch conducted a separate analysis on this data, analyzing it as monitor data versus lab data for more accurate interpretation of the results.
DGA Findings from Online Monitoring
Based on the 4 months of real-time data provided by the online DGA monitor, it was determined that while there had been gassing events that had occurred in the past, the situation did not appear to be getting worse and the decision was made to leave the transformer operating with annual sampling. The TOA software did indicate a status 4 gassing pattern during the pilot but there were no clear events. (See Figure 7.) While the gassing was higher than IEEE 2008 guidelines may indicate, the events did not indicate any critical issues. This was supported by the OPT100M DGA monitor data, which showed stable measurements (See Figure 8.), and the Duval triangle analysis based on the online monitoring period. The Duval triangle still indicated a potential T2 fault, but it seemed stable and was not evolving to another fault type. (See Figure 7.)
AES was able to use the online mobile DGA device to confirm that there were no significant gassing patterns occurring and faults developing in this transformer by using the real-time data gathered during the pilot. While DGA analysis did indicate that there was an active T2 fault which was worthy of continued monitoring, the added data from the Online DGA Monitor gave the AES Indiana maintenance team the confidence that the transformer was in a stable and safe condition and could continue to operate.
The OPT100 Mobile Online DGA device was then moved to another critical transformer site for evaluation of its gassing patterns.
The goal of validating that the T1/T2 fault was not worsening significantly in the transformer was achieved.
- Validated in real-time. Showed there was historical gassing and apparent drifting from T1 to T2 status but nothing that was presently of concern, as minimal upward trending (ROC (Rate of Change)) overall.
- Further supporting this conclusion, the TOA Gassing status has since dropped to Status 2. Transformers that are Gassing status 3 or 4 will naturally de-escalate to a status code 2 and a Hazard Factor of zero, if the gassing stops.
- Verified the transformer was safe to operate. The AES Maintenance team had comfort to continue load the transformer.
- Still in service with no issues.
Just as importantly for the AES maintenance team, this case study validated the cost-benefit of having mobile online DGA as a great tool for deeper analysis on problem transformers identified through sampling and single-gas monitors.
- Validated operation and use case for OPT100M by Vaisala. Tested out mobile installation and confirmed that the monitor was easy to set up and commission, as well as break down and transport to new sites. Also tested how well it integrated with TOA4, which proved to be seamless.
- Validated measurements. Tested out how well it measured the gases that AES knew were in there. Tracked well with laboratory data trend indications.
- “IEEE Guide for the Interpretation of Gases Generated in Mineral Oil Transformers”. Piscataway, New Jersey: Institute of Electrical and Electronics Engineers, June 2019. C57.104-2019.
- “Advances in DGA Interpretation”. Cigre, 2019. TB771.