Connect
Bring time-aligned plant measurements into the process model. Establish the quality and meaning of each signal.
+ 01 / THE TECHNOLOGY
Understand the process.
Then improve it.
A measurement tells you what a sensor sees. A process model helps explain what the whole system is doing.
+ A CONTINUOUS ENGINEERING CYCLE
Measure. Estimate.
Optimise. Repeat.
Estimate
Combine measurements with process relationships to infer unmeasured conditions and reconcile conflicting information.
Optimise
Evaluate operating choices against an objective, subject to the model and defined operating limits.
Apply
Present operating targets for review or integrate with the site’s control system through an agreed interface.
+ APPLIED MATHEMATICAL OPTIMISATION
Turn process understanding
into operating setpoints.
Estimata evaluates operating choices mathematically, connecting the plant objective to the settings that influence it.
A model of the process. An explicit objective. A feasible operating point.
Estimata uses estimated process conditions to calculate operating targets. The objective may be higher throughput, improved recovery or a better balance between production and product quality. Equipment capacity, operating bounds and quality requirements define the feasible region.
As measurements and conditions change, the application can update its estimates and recalculate targets. Recommendations can be presented to operators or passed to the site’s control system through an approved integration.
Explore circuit-wide optimisationFROM OBJECTIVE TO ACTION
- Define the outcome
- For example: maximise recovered copper across the flotation circuit.
- Respect the limits
- Maintain concentrate grade and operate within cleaner capacity.
- Calculate the targets
- Evaluate the combined effect of operating settings and recycle streams.
- Review or apply
- Provide advisory targets or enable approved automatic setpoint updates.
The preferred operating point depends on the model, available measurements, objective and configured constraints.
+ VISIBILITY BEYOND THE INSTRUMENT
Understand what
the plant is doing.
Bring measurements, estimated conditions and operating targets into a connected view of equipment performance.
Act with context.
See the current operating condition alongside estimated variables and recommended targets. Understand what is limiting performance and how a proposed change relates to the process objective.
Investigate the process.
Examine estimates, measurement discrepancies and uncertainty. Use the model to connect equipment behaviour across the circuit and investigate conditions that individual instruments cannot explain.
Connect condition to performance.
Build a consistent view of process performance from the underlying balances and estimates. Interpret results in the context of feed conditions, operating constraints and data quality.
+ KEY FEATURES
Confidence is part
of the answer.
Estimation is more than producing a value. It also means understanding how much confidence to place in that value and how it relates to the rest of the process.
Prediction and correction
A continuously updated process picture.
The process model predicts how conditions evolve from the previous estimated state and the available inputs. When new measurements arrive, Estimata compares them with the model’s predicted measurements and uses the differences to correct the state estimate.
The estimator tracks uncertainty in the states, including the correlations between their errors. A correction to one variable can therefore influence other connected estimates; a flow measurement, for example, may affect the balance of several linked streams.
Measurement resilience
Continue with an honest view of uncertainty.
A missing measurement should not be treated as a valid, precise reading. For a failed measurement that is not critical to estimator operation, Estimata can substitute a configured default and assign a larger “fail-error” to represent the greater uncertainty of that value.
The estimator consequently places less weight on the substituted measurement, relying more heavily on the process model and the remaining information. This can allow the estimation application to continue without reformulating the entire estimator for each non-critical instrument failure.
Calibration with a reference
Connect the instrument to an independent reference.
Where suitable reference measurements exist, Estimata can estimate an instrument’s calibration relationship. A laboratory assay, for example, can provide an independent reference against which an online measurement is compared.
The process must vary enough to identify the relationship. This sufficient excitation is essential: measurements clustered around one operating point may not distinguish a gain error from an offset or reveal the behaviour across the full operating range.
A service that evolves
Optimisation expertise, maintained over time.
Estimata combines models, estimation, optimisation and software in an ongoing cloud-based service. Model and software maintenance can be managed centrally, with application expertise supporting the operation as its conditions and objectives evolve.
Rapid deployment
Cloud-based application instances can be deployed remotely. The practical rollout depends on data availability, approved connectivity and integration with the site.
Real-time and on-demand computation
Plant Historian data provides the input for recurring estimation and optimisation calculations. Outputs are refreshed at the application’s configured cadence, with data latency and computational requirements considered.
Subscription-based access
The subscription model provides access to optimisation capability and specialist support as an ongoing service. Scope and commercial terms are defined for the application and the level of support required.
+ MODEL-BASED INTELLIGENCE
Anticipate the response.
Then choose the action.
Estimata is a mathematical optimisation system grounded in process models.
For well-understood unit processes, models describe the relationships between operating settings and process outcomes. This allows candidate changes to be evaluated before selecting a target, rather than relying only on rules triggered by the current measurements.
SAG mills, ball mill circuits, flotation circuits and thickeners are examples where model-based estimation and optimisation can be applied. The benefit depends on how well the model represents the process, the information available and the constraints defined for the application.
Expert rules can encode useful operating knowledge. Model-based optimisation adds the ability to evaluate competing objectives and interacting constraints explicitly, using a predicted process response.
See the applications+ THE NEXT OPERATING POINT