Condition-based maintenance (CBM) is becoming more commonplace within the petrochemical industry. While we find that previous research leveraging machine learning has provided high accuracy in the predictive aspect of machine breakdowns, the diagnostic aspect of these approaches is often lacking. This paper implements a supervised machine learning approach, with the goal of both prediction and diagnosis of machinery breakdowns, emphasizing the latter. To achieve this, it uses an XGBoost model trained on a combination of sensor and report data, and enriches the model with Shapley values for diagnostic insights. We show that this combination of statistical methods, combined with a proper data treatment, can be used to great effect and can vastly improve the diagnostic value of machine learning approaches. The insights that follow from the analysis can subsequently be leveraged by plant operators in CBM strategies or root-cause analyses. (c) 2021 Elsevier Ltd. All rights reserved.

In this research, we propose a novel data-centric way of optimizing a catalytic cracking unit. We first design a soft sensor to predict catalyst saturation levels within a Fluid Catalytic Cracking Unit (FCCU). To achieve this, we implement an established method and combine it with modern algorithms for accurate and robust results. The input for this model is data from a number of sensors throughout the refinery, combined with laboratory data. Catalyst saturation level is measured by way of manual refraction analysis and lookup tables. These manual measurements were combined with laboratory data to provide training input for our soft sensor models. Subsequently, we utilize this new soft sensor model in an input mix optimization in order to continuously optimize the use of the catalyst within the FCCU. This model leads to a higher product yield, less catalyst consumption, and a more efficient process. This proposed optimization pipeline can be introduced as smart process control tying into the development towards Industry 4.0.

The machine learning revolution is starting to be implemented in machinery maintenance and has become inevitable in highly industrialized and integrated plants. These measures save time, money and effort through new and dynamic condition-monitoring strategies. Moreover, knowledge can be extracted from these models about the potential root causes of machinery breakdowns. This results in key information to prevent similar situations in the future. Finding patterns in these breakdowns and root causes through advanced data analytics is not commonplace however. This case study then develops a strategy to implement predictive modeling and to perform root cause analysis on a compressor unit running at one of the largest refineries in Europe. Findings are presented to field experts and are deemed to add to their intuition due to the presence of unanticipated triggers. We show that root causes can be identified by constructing an intelligent data pipeline based on a multitude of readily available sensor data.

This paper describes a benchmark for stream processing frameworks allowing accurate latency benchmarking of fine-grained individual stages of a processing pipeline. By determining the latency of distinct common operations in the processing flow instead of the end-to-end latency, we can form guidelines for efficient processing pipeline design. Additionally, we address the issue of defining time in distributed systems by capturing time on one machine and defining the baseline latency. We validate our benchmark for Apache Flink using a processing pipeline comprising common stream processing operations. Our results show that joins are the most time consuming operation in our processing pipeline. The latency incurred by adding a join operation is 4.5 times higher than for a parsing operation, and the latency gradually becomes more dispersed after adding additional stages.

Automated Metering Infrastructure (AMI) has gradually become commonplace within the utilities industry and has brought with it numerous improvements in all related fields. Specifically in tariff setting and demand response models, classification of smart meter readings into load profiles helps in finding the right segments to target. This paper addresses the issue of assigning new customers, for whom no AMI readings are available, to one of these load profiles. This post-clustering phase has received little attention in the past. Our framework combines commercial, government and open data with the internal company data to accurately predict the load profile of a new customer using high performing classification models. The daily load profiles are generated using Spectral Clustering and are used as the dependent variable in our model. The framework was tested on over 6000 customers from GDF SUEZ in Belgium and six relevant load profiles were identified. The results show that the combination of internal data with commercial and cartographic data achieves the highest accuracy. Using external data alone, the model was still able to adequately place customers into their relevant load profile.