Building Optimal Multiresolution Soft Sensors for Continuous Processes

Abstract

Data-driven models used in soft sensor applications are expected to capture the dominant relationships between the different process variables and the outputs, while accounting for their high-dimensional, dynamic, and multiresolution character. While the first two characteristics are often addressed, the multiresolution aspect is usually disregarded and confused with a multirate scenario: multiresolution occurs when variables have different levels of granularity, because of, for instance, automatic averaging operations over certain time windows; on the other hand, a multirate structure is caused by the existence of different sampling rates, but the granularity of the recorded values is the same. This has two major and immediate implications. First, current methods are unable to handle variables with different resolutions in a consistent and rigorous way, since they tacitly assume that data represent instant observations and not averages over time windows. Second, even if data is available at a single-resolution (i.e., all variables with the same granularity), it is not guaranteed that the native resolution of the predictors is the most appropriate for modeling. Therefore, soft sensor development must address not only the selection of the best set of predictors to be included in the model, but also the optimum resolution to adopt for each predictor. In this work, two novel multiresolution frameworks for soft sensor development are proposed (MRSS-SC and MRSS-DC) that actively introduce multiresolution into the data by searching for the best granularity for each variable. The performance of these methodologies is comparatively assessed against current single-resolution counterparts. The optimized multiresolution soft sensors are bounded to be at least as good as their single-resolution versions, and the results confirm that they almost always perform substantially better.

Machine learning based soft sensor model for BOD estimation using intelligence at edge

“…Several types of research have addressed the development of soft sensors with fairly large numbers of real-time applications [3,[27][28][29]. Different approaches exist to develop a soft sensor like the model-based approach or empirical approach [30]. Model-based approaches describe the fundamental physical and chemical phenomena taking place in the process.…”

“…Examples of methodology used in these approaches are principal components regression [32], artificial neural network [33], neuro-fuzzy systems [34], ML algorithms [35] like IBK, random forest, random tree, Kstar, REPTree, support vector machine (SVM) [21,36], and Gaussian processes [37,38]. The soft sensor concept is now widely being used in different application areas, such as biological wastewater treatment [19], bioprocess monitoring [29], bio-chemical systems [39], and many complex process predictions [16,30].…”

Multiscale and Multi-Granularity Process Analytics: A Review

“…Therefore, it should be in the best interest of the analyst to have the capability of tuning the optimal granularity to use for each variable, in order to maximize the quality of the outputs of data analysis. The optimal selection of variables' resolution or granularity has been implemented with significant success for developing inferential models for quality attributes in both batch [148] as well as continuous processes [149]. Notably, it can be theoretically guaranteed that the derived multiresolution models perform at least as well as their conventional single-resolution counterparts.…”

Outliers detection using an iterative strategy for semi‐supervised learning

“…In terms of data structures, this results in two types of process data: labeled for which there is a corresponding output, ie, measurements for quality characteristics and unlabeled for which there is no output. This type of data is commonly referred as multirate data in chemical and process engineering. In the current work, it is assumed that inputs and outputs are collected at different frequencies and inputs are samples at the same frequency.…”