Inferential Models for Real-Time Compositional Property Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In manufacturing processes like refining and chemical production, compositional properties such as flash point and boiling point are difficult to measure in real-time, leading manufacturers to rely on experience for adjusting process parameters, which can be inefficient and costly.
Innovation Solution
A method and system for building inferential models that estimate compositional properties by cleaning and conditioning data, identifying relevant variables, and fitting them to models, allowing for real-time adjustments to process parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time measurement of compositional properties is implemented, then process control accuracy is improved, but measurement cost and system complexity increase
Solution Approach 1:
The patent introduces an inferential model as an intermediary between easily measurable process parameters (temperature, pressure, flow rates) and difficult-to-measure compositional properties. The model acts as a mediator that translates readily available sensor data into estimates of compositional properties like flash point, boiling point, and viscosity, avoiding the need for complex direct measurement systems while maintaining measurement precision.
Solution Approach 2:
The patent replaces physical/chemical measurement systems (such as flash point testers, viscosity viscometers, or chromatography systems) with a computational model-based system. Instead of using mechanical or chemical apparatus to directly measure compositional properties, the system uses mathematical relationships and data processing to infer these properties from other process parameters, thereby reducing device complexity and measurement cost.
2Measurement precision
If direct measurement of compositional properties is used, then measurement accuracy is improved, but measurement time and response speed worsen
Solution Approach 1:
The patent performs preliminary action by continuously collecting and storing process data (temperature, pressure, flow rates, compositions) during normal operation. The inferential model is pre-trained on historical data and continuously updated, so when compositional property estimation is needed, the system can immediately process existing data without waiting for time-consuming direct measurements. This eliminates measurement delays while maintaining accuracy through continuous model refinement.
Solution Approach 2:
The patent replaces time-consuming physical measurement processes (such as flash point testing, viscosity measurement, or chromatographic analysis) with instantaneous computational inference. The mathematical model can calculate compositional properties in real-time based on current process parameters, providing immediate feedback for process control without the delays inherent in laboratory or online physical measurement methods.
3Ease of operation
If manufacturers rely on experience to adjust process parameters, then operational flexibility is maintained, but process efficiency and productivity decrease
Solution Approach 1:
The patent implements a feedback mechanism where the inferential model continuously estimates compositional properties and feeds this information back to the control system. This closed-loop feedback enables automated or assisted adjustment of process parameters based on actual (or inferred) product quality, replacing reliance on operator experience with data-driven decision-making. The system automatically identifies when compositional properties deviate from targets and triggers appropriate parameter adjustments, improving both efficiency and consistency.
Solution Approach 2:
The patent enables the process control system to serve itself by using the inferential model to automatically monitor compositional properties and trigger parameter adjustments without requiring operator intervention. The system self-diagnoses quality issues and self-corrects process deviations, reducing dependence on operator experience while maintaining operational flexibility. This automation increases productivity by continuously optimizing processes without human delay or error.
Data Source
AI summary
Methods for building models that estimate compositional properties of a process may include (a) receiving data relating to a process; (b) cleaning the data by identifying and removing outlying data points from the data; and conditioning the data; (c) identifying inferential model parameters comprising a parameter selected from the group consisting of: a compositional property of the process as a model output that is not part of the data, operational constraints of the process, interactions between process variables of the chemical process, and any combination thereof; (d) building one or more inferential models based on the cleaned data and the inferential model parameters; and (e) outputting the one or more models and the corresponding validation metric.