Machine Learning Methods for Prediction of Changes in Material Flow Networks

Complex networks of material flow in manufacturing and logistics: Modeling, analysis, and prediction using stochastic block models

The application of machine learning methods for prediction of metal immobilization remediation by biochar amendment in soil

Nowadays significant part of plastic and, in particular, thermoplastic products of different sizes is manufactured using injection molding process. Due to the complex nature of changes that thermoplastic materials undergo during different stages of the injection molding process, it is critically important to control parameters that influence final part quality. In addition, injection molding process requires high repeatability due to its wide application for mass-production. As a result, it is necessary to be able to predict the final product quality based on critical process parameters values. The following paper investigates possibility of using Artificial Neural Networks (ANN) and, in particular, Multilayered Perceptron (MLP), as well as Decision Trees, such as J48, to create models for prediction of quality of dog bone specimens manufactured from high density polyethylene. Short theory overview for these two machine learning methods is provided, as well as comparison of obtained models’ quality.

This research aimed to map the application of machine learning (ML) methods in the metallurgical industry from 2000 to 2024, identifying key topics, trends, and future directions. A mixed-methods approach, combining bibliometric analysis, text mining, and content analysis, was employed. A total of 341 articles from Scopus and 249 from Taylor & Francis were reviewed following the Preferred Reporting Items for Systematic reviews and Meta-Analyses method and an ad hoc selection of 10 key articles. The analysis revealed five main research areas: (1) Advances in ML and materials science; (2) innovation in additive manufacturing; (3) applications of ML algorithms in steel metallurgy; (4) predictive analytics and modeling; and (5) artificial intelligence and deep learning applied to metallurgy. Based on these findings, five future research directions were proposed, including process optimization using deep learning, prediction of alloy behavior, sustainable waste management, integration of digital twins, and addressing ethical and regulatory challenges.

Eco-balance is the basis for the assessing production impact on the environment. Eco-balancing is divided into four phases: goal and scope definition, inventory analysis, impact assessment, interpretation. However, during the creation and evaluation of material flow networks some defects appear which inhibit or make it more difficult to establish realistic statements towards the environmental impact. In order to make a reliable statement about the environmental impact with the help of eco-balance it is necessary to consider all material and energy flows. This chapter gives an overview of the classification of defects in material flow networks. After the classification of data defects the causes of these defects are discussed. In order to resolve the causes of defects some solutions will be presented using the advantages of Petri nets and the application of Fuzzy sets and Rough sets to the Petri nets.

Based on the combination of the U.S. economic input-output table and the stocks and flows framework for characterizing anthropogenic metal cycles, this study presents a methodology for building material flow networks of bulk metals in the U.S. economy and applies it to aluminum. The results, which we term the Input-Output Material Flow Networks (IO-MFNs), achieve a complete picture of aluminum flow in the entire U.S. economy and for any chosen industrial sector (illustrated for the Automobile Manufacturing sector). The results are compared with information from our former study on U.S. aluminum stocks and flows to demonstrate the robustness and value of this new methodology. We find that the IO-MFN approach has the following advantages: (1) it helps to uncover the network of material flows in the manufacturing stage in the life cycle of metals; (2) it provides a method that may be less time-consuming but more complete and accurate in estimating new scrap generation, process loss, domestic final demand, and trade of final products of metals, than existing material flow analysis approaches; and, most importantly, (3) it enables the analysis of the material flows of metals in the U.S. economy from a network perspective, rather than merely that of a life cycle chain.

This study describes the preparation of an eco‐balance to document the materials and energy which are part of the beer production process of the Hasseröder Brewery Ltd. A general introduction to eco‐balances and material flow networks is followed by the description of the concrete case study. The modelling of material flows and the derivation of eco‐balances is illustrated with the help of a selected process from beer production. Finally, the possibilities and limitations of the balance evaluation is discussed with the aid of performance measurement systems. The article closes with a preview of future studies.

The deployment of additive manufacturing (AM) processes had a rapid and broad increase in the last years, and the same trend is expected to hold in the near future. A way to better exploit the advantages of such technology is the use of appropriate information tools. However, today there is a lack of software applications devoted to this innovative manufacturing process. To overcome this issue, in the present work the application of manufacturing execution systems (MES), a tool commonly used in traditional manufacturing processes, is extended to AM. Furthermore, a framework for the deployment of shop-floor data, acquired through a monitoring system, in the design phase is presented: hence, MES should cooperate with design for additive manufacturing (DFAM), a set of methods and tools helpful to design a product and its manufacturing process taking into account AM specificities from the early design stages. In order to better understand the advantages of such cooperation, a case study for a proof of concept has been developed: the obtained results are promising, thus an online implementation would be recommended.

Predictive modeling for US commercial building energy use: A comparison of existing statistical and machine learning algorithms using CBECS microdata

Data reconciliation consists in modifying noisy or unreliable data in order to make them consistent with a mathematical model (herein a material flow network). The conventional approach relies on least-squares minimization. Here, we use a fuzzy set-based approach, replacing Gaussian likelihood functions by fuzzy intervals, and a leximin criterion. We show that the setting of fuzzy sets provides a generalized approach to the choice of estimated values, that is more flexible and less dependent on oftentimes debatable probabilistic justifications. It potentially encompasses interval-based formulations and the least squares method, by choosing appropriate membership functions and aggregation operations. This paper also lays bare the fact that data reconciliation under the fuzzy set approach is viewed as an information fusion problem, as opposed to the statistical tradition which solves an estimation problem.

Material flow networks are a specific technique for material flow analysis. Software tools based on this method have been used by many companies to analyze and improve complex production systems. While their intuitive modeling approach strongly enhances the identification of optimization potentials, automatic model optimization is not yet possible. The KOMSA project has closed this important methodological and instrumental gap. This article presents the developed prototype of a material flow‐based optimization tool. We provide numerical results on the convergence of the implemented optimization algorithms. The added value of integrating optimization with material flow analysis is illustrated by applying the software tool to a real optimization problem from the waste industry.

PurposeThis paper aims to offer a comprehensive review and categorization of production optimization throughout the additive manufacturing lifecycle in a cloud environment. It aims to provide a structured approach to identifying and addressing issues.Design/methodology/approachThis paper systematically reviews 75 technical papers on cloud manufacturing, nesting, scheduling and postprocessing in additive manufacturing. This includes a detailed discussion of the key issues.FindingsThis paper introduces a production framework for the entire lifecycle of additive manufacturing in a cloud environment. This framework aids in problem identification and decision-making based on the process flow. It provides an integrated view from cloud to postprocessing, examining decision interdependencies and enhancing problem identification and organization.Originality/valueTo the best of the authors’ knowledge, this paper is the first to review the complete lifecycle of additive manufacturing, emphasizing the often-overlooked aspects of postprocessing and cloud manufacturing. It offers a comprehensive study of lifecycle optimization challenges and suggests ways to streamline the production process.

Direct air capture (DAC) of carbon dioxide is a promising technology to enable climate change mitigation. The liquid solvent DAC (LSDAC) process is one of the leading technologies being piloted. However, LSDAC uses a high-temperature regeneration process which requires a lot of thermal energy. Although current LSDAC designs incorporate pre-heat cyclones and a heat recovery steam generator to enable heat recovery, these do not maximize the use of the heat in the products of calcination. In this paper, a linear optimization model is developed to minimize energy cost in a LSDAC that is powered by renewable energy and natural gas. First, the material flow network is modified to include a heat exchanger (HX) and water supply to a proton exchange membrane (PEM) electrolyser. Mass and energy balance constraints are then developed to include the water flow as well as the energy balance at the PEM and the HX. Results show that about 911 tonnes of hydrogen could be produced over 336 hours of operation using a 136MW PEM. Further analysis reveals that hydrogen production is only prioritized if the value is higher than the cost of natural gas.

This article intends to answer what manufacturing execution systems (MES) are and describe their relationship with other information systems. It summarizes the features of those significant MESs being used today and discusses their limitations and where they should evolve in order to accommodate customer needs for faster real-time response, the expanded functionality, and increased complexity, as well as computing technologies and research areas that support such evolution. The main message is that it is essential for a semiconductor manufacturer to invest into advanced MESs to keep them competitive in the severely intensified global marketplace. As an MES is deployed on the shop floor in a manufacturing enterprise, the manufacturing process becomes information driven, contributing to both overall productivity and the financial viability of the manufacturing enterprise. Lastly, recognizing that MES deployments are not restricted to the semiconductor industry, the advancements could be applied with little or no refinement to other automated manufacturing environments.

Controlling and regulation of integrated aquaponic production systems – An approach for a management execution system (MES)