https://orcid.org/0009-0001-5485-8956
The industry-focused data analytics approach for process engineering plays a vital role in minimizing errors, a clear and concise approach explored in the book Data Analytics for Process Engineers by Daniela Galatro and Stephen Dawe. This book delves into an industry-oriented data analytics approach for process engineers, specifically aimed at selecting and practically applying data analytics approaches to analyze process engineering problems. The book is effectively divided into six chapters, each seamlessly continuing from the data sources to the latest data analysis techniques in process engineering at the end of the book. It covers methods and data sources, exploratory analysis, data-based modeling for prediction, modeling for monitoring and control, and process optimization. Designed for reader comprehension, the book systematically employs mathematical equations and, in some cases, real-world physical understanding of the analyzed phenomena, making it easily understandable. The book is specially designed to align with Education 4.0 and Industry 4.0, becoming a new trend in the world of data analysis. It highlights crucial content regarding data visualization techniques, prediction, modeling, and optimization for application in process control.
The initial chapter defines data acquisition as the process of sampling physical or electrical phenomena through a Data Acquisition System (DAQ), which converts generated samples into digital numeric values. The DAQ includes sensors, measuring devices, and computers. This chapter focuses on modeling for predicting, controlling, and optimizing chemical processes, using data sources such as factory process data analysis based on Instrumentation and Control (I&C) data acquisition, pilot plant and laboratory data, and process simulation data. The obtained data is termed measurable data, while artificially generated synthetic data from measured or observed data serves as a critical data source, explored in cases where measured data is insufficient or biased. The second chapter discusses data visualization techniques, including Exploratory Data Analysis (EDA), used to evaluate datasets. This statistical modeling technique aims to assess the overall structure of datasets, create descriptive summaries, and serve as the foundation for model formation. The EDA also involves assessing data quality, examining errors, outliers, and missing data points, along with summary statistical calculations. This chapter is complex, addressing the identification and management of unusual data points and missing values, employing techniques like correlograms, clustering, and data complexity reduction.
Entering the core of the book, Chapter 3 elucidates data-based modeling using mathematical models. This differs from first-principle-based models, which provide a fundamental understanding of physical-chemical phenomena like fluid mechanics, heat transfer, and mass transfer. In data-based modeling, quantitative data analysis methods identify and study a set of variables and their relationships. This chapter provides an overview of simple regression models, nonlinear regression models, nonlinear machine learning algorithms, and distribution models. The book also outlines tools to evaluate and validate modeling performance and briefly explores research methods distinguishing predictive mathematical modeling from cause-and-effect modeling. Moving on to the direct application of modeling in industries for control functions, Chapter 4 discusses control theory in industry from chemical plants to using machine learning to enhance control and optimization.
Control functions rely on sensor or control system data, with machine learning used to address complex control issues and improve system performance. The comprehensive optimization process is discussed in Chapter 5, emphasizing the vital role of process optimization in transforming systems and manufacturing processes to enhance efficiency and reduce production costs. In machine learning, optimization aims to improve model accuracy by minimizing errors and strengthening prediction capabilities. As previously explained, data analysis and machine learning are crucial in redefining process prediction, control, and eventually, optimization. This chapter introduces various simple optimization algorithms for process engineers, such as grid search, random search, gradient search, and next-generation techniques, including evolutionary algorithms, particle swarm, Bayesian inference, and optimization. This chapter also covers multi-objective optimization, a vital tool for engineers in decision-making for solving design and monitoring problems. The final chapter summarizes crucial findings from the preceding examples.
The authors provide insights and useful resources regarding data analytics tools, including references to open datasets, distinguishing concepts between data analysis and data analytics, and defining machine learning. They emphasize the importance of evaluating data analysis, modeling, control, and optimization considering the physical aspects of phenomena within datasets. This chapter is intriguing as it outlines the future of data analytics in process engineering, providing several examples of current research.
In our review, this book can serve as a reference for researchers involved in the industry, especially in chemical and mechanical engineering processes. It’s also highly suitable for undergraduate and graduate students seeking relevant data analytics skills for their workplace. Furthermore, it can be a reference for professionals, offering quick guidance on data analytics tools to support their process engineering tasks and the latest trends in data analysis for process engineers.
Department of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Gadjah Mada, Bulaksumur, Indonesia
[email protected]
Taufik Hidayat
Department of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Gadjah Mada, Bulaksumur, Indonesia
Fajri Khatami
Department of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Gadjah Mada, Bulaksumur, Indonesia
Muh. Rizal B
Department of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Gadjah Mada, Bulaksumur, Indonesia
Additional information
Funding
Reference
- Daniela, G., and Stephen, D. (2024), Data Analytics for Process Engineers: Prediction, Control, and Optimazation, Cham: Springer. DOI: 10.1007/978-3-031-46866-7.