Modelling and optimisation deliver customer value

Pavilion Technologies, a world leader in model-based software, today announced that it has been awarded fifteen new patents for innovations in the field of modelling, optimisation and control

Pavilion Technologies, a world leader in model-based software, today announced that it has been awarded fifteen new patents for innovations in the field of modelling, optimisation and control.

These new patents enable manufacturers to further improve manufacturing production processes and thereby increase production, enhance in-process quality and reduce variable costs.

These patents also extend the application of Pavilion's modelling technology to the enterprise, generating additional value for manufacturers seeking to bridge plant-to-enterprise gaps with robust model-based production performance metrics.

"Our long history of technical innovation and continued emphasis on innovative research has enabled us to deliver leading-edge manufacturing solutions that provide immediate value to our manufacturing clients," said Pavilion Technologies CEO, Ralph Carter.

"Our new patents not only strengthen the value of our solutions within the plant, but also uniquely help customers who are trying to integrate business and production processes across the enterprise." The fifteen new patents represent the continued Pavilion commitment to deliver customer value in four distinct ways.

* Improve Model Quality While Reducing Deployment Time "Method and Apparatus for Training a System model with Gain Constraints" (US Patent 7,058,617) speeds model development while also increasing model accuracy through the incorporation of gain constraints.

* "Automated Method for Building a Model" (US Patent 6,879,971) creates fast executable models of discrete tabular data relationships.

* "Method and Apparatus for Operating a Neural Network with Missing and/or Incomplete Data" (US Patent number 6,957,203) produces more robust models that incorporate an estimation of output uncertainty.

*"Residual Activation Neural Network" (U.S.

Patent 6,985,781) creates models that accurately preserve the inherent relations between input variables, allowing faster, more accurate control of these models.

* Increase Overall Value of Control and Optimization Solution "System and Method of Adaptive Control of Processes with Varying Dynamics" (US Patent 7,039,475) provides control of very fast systems such as particle accelerators, where accurate dynamics are absolutely critical to successful control.

* "Dynamic Controller for Controlling a System" (US Patent 7,050,866) allows manufacturers to more effectively control a dynamic environment by allowing use of independent static and dynamic nonlinear models in a prediction, control and optimisation environment.

"A Method for Optimising a Plant with Multiple Inputs" (US Patent 6,950,711) combines steady state models with dynamic models in an integrated control and optimisation solution.

"Method and Apparatus for Maximizing Power usage in a Power Plant" (US Patent 7,058,483) allows rules-based objectives to be combined into control and optimisation applications through dynamic targets.

"Kiln Thermal and Combustion Control" (US Patent 7,047,089) extends the capabilities of "Kiln Thermal and Combustion" (US Patent 7,024,252), which allows cement and minerals manufacturers to control and stabilise complex, non-linear kiln processes in order to increase controller uptime and reduce energy use.

Enhance Enterprise Value "System and Method for Enterprise Modelling, Optimisation and Control" (US Patent 6,934,931) extends the applicability of Pavilion technology from the plant environment to complex business systems.

The patent reflects the Pavilion object-oriented framework for modelling, optimisation and control of business processes.

To enhance the speed and quality of proactive decision making, the software makes use of a combination of model types including first-principles models and empirical models, known constraints and process objectives.

Support Next Generation Applications An additional series of patents reflect Pavilion's commitment to next generation data analysis, modelling and control with investment in Support Vector Machines (SVM), considered standard tools along with neural networks for machine learning and data mining.

The patents include: "Pre-Processing Input Data with Outlier Values for a Support Vector Machine" (US Patent 6,941,301), "System and Method for Historical Database Training of Support Vector Machines" (US Patent 6,944,616), "System and Method for Pre-Processing Input Data to a Support Vector Machine" (US Patent 7,020,642) and "System and Method for On-line Training of a Support Vector Machine (US Patent 7,054,847).

The new patent awards bring the Pavilion patent portfolio to 156 patents, reinforcing the company's commitment to innovation that delivers customer value.

Back to top