Beneath the surface of every landfill lies an invisible economy of gas — methane rising through decomposing waste, either lost to the atmosphere or captured as fuel. Loci Controls, drawing on a decade of wellhead data, has trained an AI to help operators see what the numbers alone cannot easily tell them: which wells are drifting, which are failing, and where to act before a renewable energy plant goes offline. It is a quiet but consequential convergence of machine learning and environmental stewardship, where better data interpretation translates directly into fewer emissions and more resilie
Loci Controls deploys AI to optimize landfill methane capture
tease out some things that would have been harder to find otherwise
Mark
Why does it matter that Loci has a decade of data? Couldn't any AI company train a model on landfill data?
Mimi
The data is anonymized, so it's not proprietary in that sense. But it's the *scale* and *continuity*—real-time measurements from dozens of wells over ten years. That's rare. Most AI models in waste are trained on smaller datasets or synthetic data. Loci's model learned what normal looks like, what drift looks like, what failure looks like.
Mark
So the AI isn't doing anything revolutionary. It's just pattern-matching at scale.
Mimi
Exactly. But that's the point. Operators don't need revolutionary. They need to walk in at 6 a.m. and know which three wells to check instead of staring at fifty rows of numbers. The AI is doing the tedious cognitive work.
Mark
Why is renewable natural gas such a big deal for landfills?
Mimi
Pipeline operators are strict about gas composition. If your methane or CO2 content drifts, you're out. You lose revenue instantly. An RNG operator can't afford to guess which well is the problem. They need to know now.
Mark
And the climate angle—is this actually reducing emissions, or just capturing them for fuel?
Mimi
It's capturing them. Methane that would have vented into the atmosphere is being burned as fuel instead. That's not zero impact, but it's better than the alternative. And it funds the whole operation.
Mark
What's the next step for Loci?
Mimi
They want to add a chatbot that can troubleshoot in real time. But that costs more to run. They're also expanding to operators who have monitoring equipment but not full control systems—a bigger addressable market.
O Pulso
- Landfill operators begin every shift facing walls of pressure readings and flow rates, with no easy way to know which of dozens of wells demands immediate attention.
- Methane — one of the most potent greenhouse gases — escapes uncaptured when operators miss subtle system drift, costing both climate and revenue.
- Loci Controls deployed an LLM trained on anonymized wellhead data to surface the patterns that matter, flagging problems before they cascade into pipeline violations or plant shutdowns.
- Early pilots showed operators could pinpoint the exact wells pulling their gas out of specification — critical intelligence for renewable natural gas plants facing strict pipeline compliance.
- The company has already achieved a 15% methane capture improvement through its hardware; AI is now the next layer, with success measured not in percentages but in plant uptime and operational resilience.
- A chatbot troubleshooting agent is on the horizon, and Loci plans to extend AI access to monitor-only clients — widening the reach of a system built on ten years of landfill operational secrets.
Beneath the surface of every landfill lies an invisible economy of gas — methane rising through decomposing waste, either lost to the atmosphere or captured as fuel. Loci Controls, drawing on a decade of wellhead data, has trained an AI to help operators see what the numbers alone cannot easily tell them: which wells are drifting, which are failing, and where to act before a renewable energy plant goes offline. It is a quiet but consequential convergence of machine learning and environmental stewardship, where better data interpretation translates directly into fewer emissions and more resilient energy systems.
Loci Controls has spent a decade collecting real-time data from wellheads at landfill sites across the country. Now the company is using artificial intelligence to turn that archive into something operators can actually act on.
The challenge is familiar to anyone working in data-heavy environments: arriving at a shift to find tables of numbers — gas flow rates, pressure readings, composition data — with no clear signal about where to focus. Melinda Sims, Loci's co-founder and director of product development, trained a large language model on anonymized data from the company's WellWatcher platform to surface the patterns that matter. The AI layer, rolled out to all clients with Loci hardware installed earlier this month, flags problems that might otherwise disappear into the noise.
The stakes are higher than they might appear. Landfills rank among the largest sources of methane emissions in the United States. That gas, if captured, can be converted into renewable natural gas — a pipeline-quality fuel that generates revenue while reducing climate impact. The better operators manage their wells, the more gas they recover. But renewable natural gas plants operate under strict composition constraints; drift out of specification and they get cut from the pipeline. Knowing exactly which well is the source of a problem means operators can respond in minutes rather than hours.
Loci's hardware has already delivered a 15% increase in methane capture. The company will measure its AI features differently — by plant uptime, the clearest signal that operators are staying ahead of changing conditions. Looking further ahead, Sims is developing a chat agent that could walk operators through troubleshooting using their own wellhead data, and plans to extend AI access to clients who have monitors but not full control systems — broadening the reach of a tool built on a decade of quiet observation.
Loci Controls, a company that helps landfill operators fine-tune their gas collection systems, has spent the last decade collecting real-time data from wellheads across dozens of sites. Now it's using artificial intelligence to make sense of it all.
The problem is straightforward: when an operator arrives for their shift, they face tables of numbers—gas flow rates, pressure readings, composition data—from wells scattered across a landfill. Spotting which ones need attention, which ones are drifting out of spec, which ones are losing efficiency, takes time and expertise. Melinda Sims, Loci's co-founder and director of product development, saw an opportunity. Her team trained a large language model on anonymized data from their WellWatcher platform to teach it what patterns actually matter. The result: an AI layer that can flag problems a human operator might miss in the noise.
"We have the largest database of real-time measurements of what's happening at a landfill," Sims said. "We're looking to put an extra level of AI smarts on top of that to tease out some things that would have been harder to find otherwise." The company rolled out these AI-assisted features to all clients with Loci wellhead monitors and controllers installed earlier this month, with minimal deployment costs so far.
The timing reflects a broader shift in the waste industry. Major haulers are already using AI for routing optimization, expecting millions in annual savings. Recycling facilities are deploying machine learning to improve sorting and reduce contamination. But landfills face a different pressure: methane. They are among the largest sources of methane emissions in the United States, according to the EPA. That gas can be captured and converted to renewable natural gas—a pipeline-quality fuel that generates revenue while addressing climate impact. The better operators can manage their wells, the more gas they can capture, and the more money they make.
Early pilots at a few sites revealed where the AI's value lies. Operators could now identify which individual wells were contributing to broader trends in the system—crucial information for sites running renewable natural gas plants. Those operators work under strict composition constraints; if their gas goes out of spec, they get kicked off the pipeline. Knowing exactly which well is the problem means they can act fast. "If you come in and you're an RNG plant operator and you're about to go out of spec and get kicked off your pipeline, then you're highly motivated to figure out where in my well field do I need to pay attention and go check," Sims said.
Loci's wellhead technology has already delivered a 15% increase in methane capture for landfills using the system. The company will measure the success of its AI features differently: by plant uptime. When operators can respond faster to changing conditions—weather shifts, equipment drift, composition swings—their systems stay online longer and capture more gas. Sims is already exploring a next step: a chat agent that could help operators troubleshoot problems based on their wellhead data, though she acknowledged that would come with higher costs. For now, the company plans to expand AI access to clients who have monitors installed but not full control systems, widening the reach of a tool built on a decade of landfill secrets.
Citações Notáveis
We have the largest database of real-time measurements of what's happening at a landfill. We're looking to put an extra level of AI smarts on top of that to tease out some things that would have been harder to find otherwise.— Melinda Sims, Loci Controls co-founder and director of product development
If you come in and you're an RNG plant operator and you're about to go out of spec and get kicked off your pipeline, then you're highly motivated to figure out where in my well field do I need to pay attention and go check.— Melinda Sims
Quer a matéria completa?
Leia o original em Waste Dive ↗
Technology