Effortless NIR Spectrometric Modeling
Process NIR analyzers are known for their simplicity of installation, high reliability, and minimal maintenance requirements. However, it is important to note that this technique is not linear and may not provide clear resolution for all components in the sample. Consequently, any changes in stream composition or process conditions necessitate model updating. Therefore, reliability of process NIR analysis heavily relies on the accuracy of chemometric models.
The spectral results obtained from NIR analysis can be influenced by fluctuations in the chemical composition of crude oil, particularly the presence of substances such as hetero-atomic molecules. These compounds are present at varying concentrations depending on the origin of the crude oil and can distill over together with heavier distillates. To mitigate this influence, several approaches can be taken, including minimizing crude switching, incorporating these compounds in the calibration samples, or utilizing model validation and corrective software.
Modcon.AI automatic modeling software offers a comprehensive solution for building and automatically updating spectrometric models. Specifically designed for NIR analyzers, this product leverages the power of Machine Learning to achieve the highest possible measurement precision. By continuously integrating the most recent reference data into the predictive model building process, Modcon.AI enables customers to gain independence in model updating.
With Modcon.AI, users can achieve modeling quality that is comparable or even superior to what can be achieved by manually updating the model using dedicated spectrometry modeling software. The software provides an intuitive interface for setting up the model update routine, a graphical representation of the model’s precision development over time, and extensive reporting functionality. Under the hood, the algorithm optimizes and fine-tunes the machine learning model, performing predictions and identifying outliers in the reference data that should be excluded from the sample set used for model training.
NIR Spectrometry has gained widespread acceptance as a non-destructive analysis technology for chemical and physical properties of various process streams across multiple industries. At Modcon Systems, we offer a range of NIR analyzers and have successfully implemented online analysis for petrochemical industry customers in Europe, Asia, North America, and South America. With our new automatic modeling software, NIR analyzer customers can fully leverage their hardware by deploying the most up-to-date models, continuously updated through our advanced software solution.
