KeeWee Tree

OpenPEMS

OpenPEMS™ is a predictive emissions monitoring software (PEMS) developed for use across Canada and internationally. The software uses machine learning to predict air emissions such as nitrogen oxides (NOₓ) and sulfur dioxide (SO₂). OpenPEMS™ is free and open-source, making Artificial Intelligence (AI) and Machine Learning (ML) technologies more accessible to a broader industrial audience, while helping reduce the costs associated with air emissions monitoring.

Data Science Illustration
Data Science Icon

What is PEMS?

Predictive emissions monitoring software (PEMS) is a software based solution for monitoring air emissions, such as nitrogen oxide (NOx) and SO2. It uses a predictive model that analyzes data from existing facility sensors and historical operation data to predict air emissions.

Why predictive monitoring?

Industrial facilities with large stationary combustion sources are typically required to be equipped with one or more continuous emissions monitoring systems (CEMS) to monitor NOx emissions for compliance with regulatory emission limits. Since PEMS works with existing sensors and is software based, it is far more cost effective to run - requiring far fewer resources and manpower.

PEMS are equal to CEMS in terms of accuracy and have been used across the world for more than 50 years.


Cost-effective

Cost-effective

Eliminates need for costly maintenance of specialized hardware.
Software-based

Software-based

Uses data from existing facility sensors and reports emissions based on a predictive model.
Regulatory approved

Regulatory approved

Accepted alternative monitoring method to CEMS at federal and provincial levels in Canada.
Quick set up

Quick set up

Can be installed in less than six weeks.
No extra calibration needed

No extra calibration needed

PEMS relies on existing facility sensors, which are already maintained under standard preventative schedules.
50+ years in use

50+ years in use

PEMS has been in use in the US since the 1970s, and has been approved by EPA for regulatory reporting.

What projects can OpenPEMS™ be used for?

OpenPEMS™ is suitable for use across the supply chain, from the largest combustion sources to the smallest. It can be used to measure a variety of emission types, including:

NOx

Nitrogen oxide

SO2

Sulfur dioxide

CH4

Methane

PM

Particulate matter

CO2

Carbon dioxide

H2S

Hydrogen sulfide

Why switch to OpenPEMS™?

After OpenPEMS™ is trained on four months of CEMS data, it can continuously measure and predict emissions from emission sources. This means your facility will have reliable, continuous data for all emissions sources - even when they do not have a CEMS unit monitoring them.

Continuous monitoring

Continuous monitoring

Emissions output can be constantly monitored for all sources.
Works for large and small sources

Works for large and small sources

OpenPEMS™ can be used for emissions sources of any size throughout the supply chain.
No specialized equipment

No specialized equipment

OpenPEMS™ is software based, so no special sensors or probes are required long-term.

Further Readings

PEMS Development With Neural Network

Development of predictive emissions monitoring system using Keras - an open source machine learning library.

Read More

PEMS Development With Gradient Boosting

Development of predictive emissions model using gradient boosting machine learning method.

Read More

Long-term Evaluation of Machine Learning Based Methods for Air Emission Monitoring

Long-term study reveals ML model accuracy, overfitting, and regulatory challenges.

Read More
predictive emission model

Discuss Your Project

See how KeeWee can support your emissions goals with data you can trust and compliance you can count on.