Hybrid Inferential Hard Analyzer for Hydrocarbon Stream Property Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inferential analyzers for hydrocarbon streams are insufficiently robust due to indirect and incomplete representation of process chemistry, leading to inaccurate property predictions, and require frequent updates with laboratory measurements to maintain accuracy, which limits process optimization and efficiency.
Innovation Solution
A system integrating a hard analyzer with an inferential analyzer to capture and compare dependent variables, allowing for real-time adjustment of model parameters and prediction refinement using statistical validation techniques, enhancing the accuracy and frequency of property measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional inferential analyzers are used to predict properties of hydrocarbon streams, then the measurement frequency can be increased and sample integrity issues are eliminated, but the prediction accuracy deteriorates due to indirect and incomplete representation of process chemistry
Solution Approach 1:
The patent combines inferential analysis (soft sensor) with hard sensor measurements into a hybrid system. The soft sensor provides high-frequency predictions while the hard sensor periodically validates and calibrates the model, merging the advantages of both approaches to maintain accuracy at high measurement frequencies.
Solution Approach 2:
The system implements feedback mechanisms where hard sensor measurements are used to continuously update and recalibrate the inferential model. This feedback loop ensures that the soft sensor predictions remain accurate over time despite changes in process conditions or chemistry.
2Measurement precision
If frequent hard analyzer measurements are performed to maintain model accuracy, then prediction accuracy is improved, but the time delay and cost associated with laboratory measurements increase
Solution Approach 1:
Instead of continuous hard analyzer measurements, the system uses periodic measurements at strategically determined intervals. The soft sensor bridges the gaps between hard measurements, providing continuous predictions without the time delay and cost of frequent laboratory analysis.
Solution Approach 2:
The inferential model performs preliminary predictions using available process data before hard analyzer results are ready. This allows the system to provide timely predictions based on the most recent model state, reducing the effective time delay while maintaining accuracy through subsequent calibration.
3Reliability
If more independent variables are captured to improve model robustness, then the completeness of chemistry representation is improved, but the device complexity and data processing requirements increase
Solution Approach 1:
The system uses a unified data acquisition framework that captures multiple types of variables (process conditions, analytical measurements, operational parameters) through a single integrated architecture. This multi-functional approach improves model robustness without proportionally increasing system complexity.
Data Source
AI summary
Systems, methods, and apparatuses are provided for determining properties of process streams, in particular, hydrocarbon processing streams. The systems, methods, and apparatuses frequently, for example, substantially in real-time, determine measurements for the properties of the process stream. The systems, methods, and apparatuses provide features that allow such properties of process streams to be accurately measured even as process conditions and other parameters that affect process operations change. More specifically, an analyzer having a measurement device configured to detect one or more independent variables of a process stream, a model configured to determine one or more analyzer measurements from the one or more independent variables, and a procedure to adjust the model using a corresponding primary measurement is disclosed.