Correct
and timely fault diagnosis is of great importance for enhancing
the safety and reliability of modern chemical industrial processes.
With the arrival of the big data era, data-driven fault detection
and diagnosis (FDD) methods offer enormous potential for complex chemical
processes. Deep learning-based data-driven FDD methods, which extract
features from raw data using an artificial neural network (ANN), are
attracting widespread attention. Among various types of neural networks,
recurrent neural network (RNN) performs excellently when dealing with
time-series data. However, a regular unidirectional RNN proceeds only
in the positive time direction, resulting in insufficient feature
extraction and inferior fault diagnosis performance. In this study,
a bidirectional RNN (BiRNN) was employed to construct FDD models with
sophisticated RNN cells. When applied to the benchmark Tennessee Eastman
process, BiRNN-based FDD models exhibited a dramatically impressive
performance, demonstrating the effectiveness of implementing BiRNN
in chemical process fault diagnosis.
Naphtha
is an important product of crude oil and has widespread
industrial applications. Additionally, the composition of naphtha
is influenced by the procurement and operations of crude oil, which
has different characteristics in each region. In this paper, a novel
molecular reconstruction process is proposed to provide an accurate
composition for local refineries. An effective probability density
function composed of the gamma distribution trend, regional features,
weight features, and uncertainty is constructed using a self-adaptive
cloud model and used in this process. A hybrid genetic algorithm–particle
swarm optimization method is applied in this process after comparing
with other optimization methods. The results of the proposed molecular
reconstruction process are much closer to the actual composition than
that of other methods, which verifies that this process has high performance
and is feasible for industrial application.
Food flavor quality evaluation is attracting continuous attention, but a suitable evaluation system is severely lacking. Gas chromatography-mass spectrometry/olfactometry (GC-MS/O) is widely used to solve the food flavor evaluation problem, but the olfactometry evaluation is unfeasible to be carried out in large batches and is unreliable due to potential issue of an operator or systematic laboratory effect. Thus, a novel fingerprint modeling and profiling process was proposed based on several machine learning models including convolutional neural network (CNN). The fingerprint template was created by the data analysis of existing GC-MS spectrum dataset. Then the fingerprint image generation program was applied for structuring the complex instrumental data. Food olfactometry result was obtained by a machine learning method based on CNN using fingerprint image as the input. The case study on peanut oil samples demonstrated the model accuracy of around 93%. By structure optimization and further dataset expansion, the whole process has the potential to be utilized by sensory laboratories for aroma analysis instead of humans.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.