What is a digital twin?
A digital twin is a powerful technique that used statistical approaches to mimic the plant to which it is assigned. There are a range of statistical approaches that include principal component analysis, Bayesian analysis and Kalman filtering to continually update and assess accuracy of the model fit.
Better decisions from the data you already collect
Normally plant records include feedstock inputs, gas production, methane content, temperature and laboratory results. With help these can be turned into a clearer picture of performance.
Is a change worth investigating?
Given todays economic circumstances any small advantage is worth pursuing. For example in some situations gas output and composition varies. Statistical analysis helps separate everyday variation from a developing problem—and can check whether an apparent improvement lasts.
Understand your plant’s baseline
We use operating history to establish what normal performance looks like for your digester over time and this tracking is performed by a digital twin. That makes unusual trends easier to spot and changes easier to assess.
A digital twin that observes, not controls
A digital twin is a model updated with data from current data from the plant. It can gather readings from existing sensors, laboratory tests and operating records as they become available. It works in the background. It does not have to change feed rates or plant settings. Operators remain in control.
Explore changes before making them
Once sufficient data is collected and confirmed as replicating the actual plant performance, the model can help explore questions such as:
What might a different feedstock mix mean for gas production?
How could increased loading affect the process?
Does a new reading fit normal behaviour, or need investigation?
These are scenarios to inform decisions—not guarantees or automatic instructions.
The right level of analysis for your plant
A useful starting point may be a focused review of existing data. More detailed modelling depends on the quality and frequency of the measurements available. A digital twin cannot reveal what the plant has not measured nor identify further operating parameters if there is insufficient data.
A digital twin is a powerful technique that used statistical approaches to mimic the plant to which it is assigned. There are a range of statistical approaches that include principal component analysis, Bayesian analysis and Kalman filtering to continually update and assess accuracy of the model fit.
Better decisions from the data you already collect
Normally plant records include feedstock inputs, gas production, methane content, temperature and laboratory results. With help these can be turned into a clearer picture of performance.
Is a change worth investigating?
Given todays economic circumstances any small advantage is worth pursuing. For example in some situations gas output and composition varies. Statistical analysis helps separate everyday variation from a developing problem—and can check whether an apparent improvement lasts.
Understand your plant’s baseline
We use operating history to establish what normal performance looks like for your digester over time and this tracking is performed by a digital twin. That makes unusual trends easier to spot and changes easier to assess.
A digital twin that observes, not controls
A digital twin is a model updated with data from current data from the plant. It can gather readings from existing sensors, laboratory tests and operating records as they become available. It works in the background. It does not have to change feed rates or plant settings. Operators remain in control.
Explore changes before making them
Once sufficient data is collected and confirmed as replicating the actual plant performance, the model can help explore questions such as:
What might a different feedstock mix mean for gas production?
How could increased loading affect the process?
Does a new reading fit normal behaviour, or need investigation?
These are scenarios to inform decisions—not guarantees or automatic instructions.
The right level of analysis for your plant
A useful starting point may be a focused review of existing data. More detailed modelling depends on the quality and frequency of the measurements available. A digital twin cannot reveal what the plant has not measured nor identify further operating parameters if there is insufficient data.
Early tracking with a digital twin configured as Bayesian analysis with new inputs and Kalman filter to 'calibrate' sensor and input responses
