"Oil coats the same vehicle parked outside S&WB Carrollton Plant 13 months apart. Turbine 5 exhaust stack in background. " Credit: Courtesy the Louisiana Department of Environmental Quality and Ariane Livaudais

Oil sprayed the neighborhood around the Carrollton Power Plant

The pattern and distribution mimic a 2022 venting of lubrication oil from the exhaust stack of the utility's turbine generator 5, an event the Board blamed on operator error when a tank was overfilled.

Residents living across Spruce Street from the Sewerage and Water Board’s Carrollton Power Plant found their homes and property coated in oily residue over the weekend of April 8-9 — for the second time in just over a year.

The pattern and distribution mimic a 2022 venting of lubrication oil from the exhaust stack of the utility’s turbine generator 5, an event the Board blamed on operator error when a tank was overfilled.

If confirmed, the discharge would mark the sixth time in thirteen months the agency has emitted chemicals past its fence line in violation of its state operating permits. In addition to the oil spraying in March 2022, four spills between December 2022 and February 2023 resulted from over 4,000 gallons of diesel fuel and highly corrosive hydrofluorosilicic acid overflowing tanks, all of which were attributed to operator error by the Board.

Nearly 3,000 gallons from those spills flowed into storm drains which lead to Lake Pontchartrain. One of the spills was not reported until three days after it occurred.

While human actions very likely played a role in all the incidents, investigation by The Lens shows other undisclosed factors were just as important in allowing the discharges to escape the Carrollton Plant.

In the case of the acid spills, the Board had been warned repeatedly by the state department of health since 2015 about lack of required containment around the acid tanks to prevent infiltration into the city’s drainage system.

For the turbine oil spraying, despite public statements by the agency executive director that nothing could be done besides increased operator training and monitoring to prevent further incidents, Board documents show the turbine control system has had equipment in place for years which could stop the spraying as soon as it began or perhaps even before it started. 

Oil falls on Carrollton Plant neighbors again

The latest coating of oil on the Board neighbors living along Spruce Street has not yet been confirmed as coming from the Board’s facility, but it does bear a resemblance to the incident in 2022. On March 3, 2022 Spruce Street homeowner Ariane Livaudais filed a complaint with the Louisiana Department of Environmental Quality (LDEQ), the agency tasked to oversee air and surface water emissions from the Board’s plant. 

Livaudais said that a turbine inside the facility had been flinging material that appeared to be diesel fuel “all over the surrounding neighborhood,” damaging property and making it unsafe to sit outside. LDEQ dispatched an inspector the next day to look into the complaint, and found spots of oily residue on cars, sidewalks, and the street. A significant amount of oily substance was also found at the base of the turbine 5 exhaust stack. 

Photographs taken by Livaudais after the latest oil misting event bear a striking resemblance to pictures taken by the LDEQ inspector on March 4, 2022, with both sets of images showing spots of oily residue coating cars, concrete, and property. After a six month investigation into the residue – during which LDEQ issued a Notice of Deficiency and a Compliance Order – the Board wrote LDEQ with their conclusion in August 2022: the oil, identified as Mobil DTE 732 lubrication oil used to cool turbine 5’s bearings, had likely come from the turbine when the tank holding the oil had been overfilled by an operator.

To ensure oil pumped from the lubrication oil tank to the turbine’s bearings and back does not spray out of the piping and housings, a slight vacuum is imposed on the tank using a device called a vapor extractor, which acts like a vacuum cleaner. The air from the tank is directed through the vapor extractor to the main exhaust stack for the turbine. Should the level in the tank get too high, it is possible for the vacuum to pick up oil from the surface of the reservoir, and fling it as droplets out the stack and across the street.

In their August letter to LDEQ, the Board said they had increased operator monitoring of the level in the lube oil tank by not relying exclusively on the tank’s level gauge, but also checking the tank’s dipstick. They also said they were investigating installation of a knockout pot in the exhaust stream to capture any vacuumed-up oil before it left the top of the stack.

In a WWL-TV story from August reporting on the conclusion that the March 2022 oil spray event was due to “operator error” in overfilling the lube oil tank, Board Executive Director Ghassan Korban said he didn’t “want to simply rely on employees being diligent to avoid another overfilling error.” But, he said, “the refurbishment of Turbine 5 and its exhaust stack did not include improvements to much of the half-century-old technology, including the manually-filled oil tank.” Korban went on to say that the only mitigation against future overfilling would be more training.

Turbine controls already in place could stop oil spraying 

Omitted from the Board’s correspondence with the state regulator and its public statements to the media and neighbors was any detailed mention of the existing control system on turbine 5. 

Following an explosion in the turbine in December 2019, a new modern computer control system was installed by Longmont, Colorado-based Nexus Controls, a division of Baker Hughes. Nexus’ contract was originally valued at $1.5 million, with change orders raising the eventual cost to nearly $2.2 million. The new system completely supplanted the former hydraulic and pneumatic system used since the turbine was installed in the late 1950’s, a process termed “rip and replace” by Nexus in their proposal.

The Lens obtained as-built drawings of the new control system, as well as newly drafted procedures for starting, stopping, and running the turbine drafted by Houston-based firm Ethos Energy. The drawings and procedures date from 2021, before the March 2022 oil spray incident.

The drawings show the lube oil tank has a level switch or switches on it which indicate when oil levels are too low or too high. In fact, the drawing showing the high level switch is one of the originals from the 1950’s install, but it was placed inside a 2021 Nexus Controls border. Such placement of an old drawing in a new frame appears to indicate the contents of the old drawing remain valid. Other Nexus Controls drawings show that the high lube oil level output is indeed still used. It was routed into the new industrial computers to trigger an alarm. An operating procedure which documents how to start and stop the turbine includes mention of the high level alarm, as well as a “high-high level” interlock which would shut the turbine down if it were triggered, though Nexus Control’s drawings do not seem to include instrumentation which could indicate such a “high-high” level in the tank.

The presence of the high level switch across multiple drawings of the new turbine 5 control system and the mention of the high level alarm within the startup/shutdown procedure raise certain questions as to why the Board did not make adjustments to the switch and the action it triggered after the March 2022 incident.  

When asked why the high level switch was not set up to shut down the turbine if it was triggered instead of only sounding an alarm, and whether the switch was set to trigger at a level equivalent to the overfilling believed to have caused the 2022 spraying incident, the Board did not offer a direct response. They said, “We are investigating the events of the past weekend and the reports of the droplets. [On April 7 we] were testing the EMDs in preparation for the storm in conjunction with T5, which has been online since Monday afternoon.” They added, “Turbine operations is required to check the oil levels several times a day.”

When presented with The Lens’ findings via email, Livaudais wrote, “This is another example of the SWBNO saying one thing to the public, but doing something completely different.”

The new control system was active and the procedure was in force at the time of the March, 2022 incident as well as on April 8 and 9. The Board said turbine 5 had been running continuously since April 3, 2023, the Monday before the most recent oil droplets incident. Turbine 5 was running during this period – and continues to do so – due to the shutdown of turbine 4 for steam valve problems and inspections of other unspecified components. Turbines 4 and 5 are the only somewhat reliable 25 cycle generators remaining in the Board’s fleet after the decommissioning of turbines 1 and 3 in the past two years. The Board has shown a distinct preference for turbine 4 over turbine 5, running it on over 300 days in 2022 (the units generally only run in tandem during heavy rainstorms, otherwise only one is on). Other Board sources of 25 cycle power – frequency changers which convert Entergy 60 cycle power to 25 cycle, and five electromotive diesel generators – are only used for short stretches, mostly during rainstorms.

Back to back spills in December

The oil spray incident in March 2022 was not the only one when chemicals were released at the Board’s Carrollton Plant because of a tank getting overfilled. Four times between December, 2022 and February, 2023 Board personnel overfilled storage tanks, each time resulting in spills to the ground around the tanks. In three cases, the chemicals flowed into the storm drainage system.

The first incident occurred December 16, 2022, but was not immediately reported as required by regulation. The Board uses a number of chemicals in its water treatment system, one of which is hydrofluorosilicic acid, also known as HFS. The acid adds fluoride to drinking water. As part of its operations, the Board moves acid from a main storage tank to a smaller “day” tank, from which the acid is injected into the water.

As it is stored, the acid has a pH of 1, which is highly corrosive. Stomach acid has a pH of about 2, or ten times less acidic than HFS. It was this corrosivity which led Board personnel to conclude some had spilled: on December 19 a Board employee noticed concrete around the HFS tanks showed signs of recent damage, which triggered a report to the state. Further investigation of computerized level monitoring of the tanks helped determine that approximately 2,662 gallons of acid had spilled three days earlier during a two hour period. All the acid went down the drain into the storm sewer. Damage to the subsurface drainage by the acid is unknown. The agency blamed the spill on operator error.

Normally, such hazardous chemicals are protected from not only entering nearby water bodies, but also from injuring personnel by placement of a containment structure such as a low concrete wall or an earthen berm around their tanks. Such barriers hold spills from the tank until it can be neutralized and pumped away safely. The Board has no such barrier around its HFS tanks, or its nearby anhydrous ammonia tank. Additionally, the tanks are located immediately adjacent to both the engineering building and the main roadway within the Carrollton Plant, where dozens of employees and visitors pass by daily.

The Board was aware of the lack of containment for their acid tanks long before last December. In 2015, 2017, and 2019 the state Department of Health specifically cited the lack of fluoride tank containment as a regulatory deficiency during their sanitary surveys of the east bank water system, saying such tanks must have “a receiving basin capable of receiving accidental spills or overflows without uncontrolled discharge.” In the report for the December acid spill, the Board claimed construction on a long-delayed new bulk chemical feed facility would begin this year, but made no mention of interim measures to prevent future spills from leaving the plant other than operators receiving refresher training on handling the acid.

Just five days after the discovery of the acid spill, Board operators overfilled another tank. While filling a fuel day tank for one of the five electromotive diesel (EMD) generators the morning of Christmas Eve, 100 gallons of diesel fuel went to the ground and down nearby storm drains. Unlike in the case of the December 16 acid spill, the diesel overfilling was promptly reported to the state and Coast Guard. According to the official spill report, as an operator “manually” filled the tank, diesel flowed out on to Spruce Street and into the catch basins, which lead to the Monticello canal. The Board’s spill response contractor E3 OMI recovered some of the fuel and placed absorbent booms in the canal. The report is unclear if the canal booms caught oil, as it says the booms were placed both “as a precautionary measure” and “as a mitigation effort to prevent the spill from spreading.” The agency again blamed the spill on operator error.

Repeat spills in January and February; LDEQ inspects

Less than three weeks after the December diesel spill, the same EMD day tank was again overfilled, 100 gallons of diesel fuel spilled to the ground and out to Spruce Street, where it again flowed into the catch basins. During this January 11 spill, the fuel appears to have traveled further from the plant than on December 24, as E3 OMI placed absorbent booms in both the nearby Monticello Canal and the downstream Palmetto Canal, which is closer to the lake. The report says the booms were placed to “contain product causing sheening until it naturally dissipated.” The spill occurred during “manual” filling by an operator, just like on December 24, and the report concluded the spill was preventable.

Four weeks later, there was another HFS acid spill during a transfer from the main tank to the day tank. On the afternoon of February 15, an estimated 1,248 gallons of acid spilled to the ground. Unlike in December, the authorities were notified within hours. An LDEQ representative was on site within minutes of receiving the notification and observed the entire cleanup process by E3 OMI. Reports are unclear if any acid drained into the storm sewers, with a great deal of the acid – which tested as a pH of 1 three times – being neutralized with 40 bags of sodium bicarbonate and absorbed by sand. The neutralized mixture was placed in drums for shipment offsite, but not before a plastic two gallon bucket of acid was found sitting unattended under the HFS day tank. It was shipped offsite for neutralization.

This second acid spill, which appears to have happened under exactly the same circumstances as the first acid spill two months earlier, was similarly attributed to operator error. However, the spill report contains evidence there is more the Board could have done. In addition to yet more training for operators – the report does not mention if the refresher training promised after the December spill had taken place – and a repeat mention of the future bulk chemical facility, the Board said “temporary secondary containment and fail-safe transfer mechanisms will be designed and installed in the interim period,” finally addressing the multiple state department of health citations which began nearly a decade earlier, albeit two spills too late.

There has been no mention of adding containment to the diesel fuel storage around the EMD generators, installed in 2018. Aerial photos taken as recently as February 4 of this year show eleven fuel storage tanks – six of them rivaling large SUV’s in their size – arrayed around the five generators. None of the tanks have any containment structures in case of spills or overflows.

But the spills appear to have resulted in added attention from LDEQ: they paid an inspection visit to the Carrollton Plant on February 1 to examine how the Board responds to spills.

Matt McBride is a freelance reporter covering infrastructure for The Lens. He began his reporting career with his blog "Fix the Pumps," covering post-Katrina flood protection construction by the Corps...