Large-scale Industrial Environments Research Articles

A precise compensation algorithm for visual measurement in non-uniform refractive index environment based on Hamiltonian ray tracing method

In complex large-scale industrial environments, light rays no longer propagate in a straight line due to the influence of non-uniform refractive index, which will lead to large measurement errors in visual measurement. In order to accurately compensate the error, this paper firstly quantitatively analyzes the influence of non-uniform refractive index distribution on visual measurement. Then the symplectic matrix ray-tracing method of Hamiltonian optics is used to reconstruct spatial light in the reverse direction with known light propagation end point and spatial refractive index distribution. Thus the traditional model of light propagation along a straight line in visual measurement can be changed and the measurement error introduced by non-uniform refractive index field can be corrected. Finally, the proposed Hamiltonian ray tracing method was verified through the spatial non-uniform refractive index field reconstructed by Background Oriented Schlieren(BOS) method. The experimental results verified the effectiveness and feasibility of this method.

Cascading Failure Analysis of Hierarchical Industrial Wireless Sensor Networks under the Impact of Data Overload

As industrialization accelerates, the industrial sensor network environment becomes more complex. Hierarchical multi-cluster wireless sensing network topology is generally used due to large-scale industrial environments, harsh environments, and data overload impact. In industrial wireless sensor networks, the overload of some nodes may lead to the failure of the whole network, which is called cascading failure. This phenomenon has incalculable impact on industrial production. However, cascading failure models have mainly been studied for planar structures, and there is no cascading failure model for hierarchical topologies in industrial environments. Therefore, this paper built a cascading failure model for hierarchical industrial wireless sensor networks (IWSNs) for realistic industrial network topologies. By establishing an evaluation mechanism considering the efficiency of the network and the viability of nodes, the network communication efficiency that is not considered in the traditional evaluation mechanism is solved. In addition, aiming at the problem of network topology changes caused by node failure, dynamic load distribution methods (ADD, SLD) are used to improve network invulnerability. Theoretical analysis and experimental results show that the traditional allocation method (SMLD) does not apply in hierarchical topologies; when the general cluster head node capacity is moderate, increasing the capacity of single-hop cluster head nodes can prevent cascading failures more effectively.

Reconstruction of refractive index maps using photogrammetry

Large volume metrology is a key enabler of autonomous precision manufacturing. For component positioning, the optical-based metrology technique of photogrammetry could be used more widely if its accuracy was improved. These positional measurements are subject to uncertainties which can be greater than manufacturing tolerances. One source of uncertainty is due to thermal gradients, which cause the refraction of the light rays in large-scale industrial environments. This paper uses light-based sensor data to reconstruct a heterogeneous spatial map of the refractive index in air. We use this reconstructed refractive index map to discount the refractive effects and thereby reduce the uncertainty of this positioning problem. This new inverse problem employs Voronoi tessellations to spatially parameterize the refractive index map, the Fast Marching Method to solve the forward problem of calculating the light rays through this medium, and a Bayesian approach in the inversion. Using simulated data, this methodology leads to positioning improvements of up to 37.

An integrated machine scheduling and personnel allocation problem for large-scale industrial facilities using a rolling horizon framework

In this study we present an optimization problem where machine scheduling and personnel allocation decisions are solved simultaneously. The machine scheduling consists of solving a variant of the job shop problem where jobs are allocated in batches and multitasking is allowed. On the other hand, the personnel allocation problem searches for the optimal allocation of human resources to support the facility operation. Although these problems have been extensively studied independently in the literature, the integrated problem has very limited studies, in particular for large-scale industrial environments. We introduce a mathematical formulation for this problem and propose a solution method to evaluate optimal solutions more efficiently. We also implement a rolling horizon framework to test our approach using real-world based instances extracted from an analytical service facility. The results show that, overall improvements of up to 11.6% are achieved compared to the case where no personnel allocation decisions are considered within the scheduling problem.

Taking compact NMR to monitoring real reactions in large-scale chemical industries-General considerations and learnings from a lab-scale test case.

The considerations for use of compact nuclear magnetic resonance in a large-scale industrial environment clearly differ from those in academic and educational settings and even from those in smaller companies. In the first part of this article, these differences will be discussed along with the additional requirements that need to be fulfilled for successful applicability in different use cases. In the second part of the article, outcomes from different research activities aiming to fulfill these requirements will be presented with a focus on an online reaction-monitoring study on a lab-scale nucleophilic chlorination reaction.

Embedded Processing and Compression of 3D Sensor Data for Large Scale Industrial Environments.

This paper presents a scalable embedded solution for processing and transferring 3D point cloud data. Sensors based on the time-of-flight principle generate data which are processed on a local embedded computer and compressed using an octree-based scheme. The compressed data is transferred to a central node where the individual point clouds from several nodes are decompressed and filtered based on a novel method for generating intensity values for sensors which do not natively produce such a value. The paper presents experimental results from a relatively large industrial robot cell with an approximate size of 10 m × 10 m × 4 m. The main advantage of processing point cloud data locally on the nodes is scalability. The proposed solution could, with a dedicated Gigabit Ethernet local network, be scaled up to approximately 440 sensor nodes, only limited by the processing power of the central node that is receiving the compressed data from the local nodes. A compression ratio of 40.5 was obtained when compressing a point cloud stream from a single Microsoft Kinect V2 sensor using an octree resolution of 4 cm.

Customized Industrial Networks: Network Slicing Trial at Hamburg Seaport

Driven by a massive surge in digitization and customization, so-called vertical industries are expected to be a major beneficiary of the 5G of mobile networks. The use cases of such vertical industries define qualitative and quantitative requirements unprecedented in the history of mobile network development. Autonomous vehicles, traffic light control, video surveillance, and IIoT, to only name a few, introduce challenging requirements regarding both conventional performance metrics, such as throughput and coverage, as well as formerly rather subordinate system metrics, such as deterministic latency, ultra-high reliability and resilience, high number of devices, multi-tenant networks, and demanding security mechanisms. Nokia, Deutsche Telekom, and Hamburg Port Authority have deployed a largescale 5G trial testbed in the Hamburg port area. The testbed proves in a real, large-scale industrial environment that basic features of network slicing, namely slice isolation, flexible slice customization, and multitenancy, are already technically feasible. Three exemplary communication services have been selected and are demonstrated in the testbed. Multi-connectivity is implemented as a key component to achieve high reliability throughout the testbed area. The testbed shows that all network domains must be involved in the setup of network slices, that is, user terminals, radio access, core network, and enterprise networks, in order to efficiently operate and manage network slices. Therefore, the discussed life cycle management is key for the interaction between mobile service provider and tenants of the network.

Theoretical and Experimental Study of the Reaction between Ammonium Nitrate and Sodium Salts

Hazards posed by chemical incompatibility, especially in a large-scale industrial environment, warrant a deeper understanding of the mechanisms of the reactions involved in these phenomena. In this study, reactions between ammonium nitrate and two sodium salts, namely, sodium nitrate and sodium nitrite, have been studied by ab initio calculations (with density functional theory, DFT) and experimental calorimetric methods (with differential scanning calorimetry, DSC, and heat flux calorimetry, HFC). The agreement between theoretical and experimental results allows an understanding of the thermal decomposition behaviors of the two sodium salts when exposed to ammonium nitrate. Moreover, this study highlighted the critical role of the water that appears to promote the incompatibility between ammonium nitrate and sodium nitrite.

Learning from incidents: Practices at a Scandinavian refinery

This paper reports on a study of practices used to learn from incidents with the aim of improving safety performance in a Scandinavian refinery. Data for the study was collected during five months of fieldwork at the refinery and interviews with 70 refinery employees. In this paper, we examine how managers, engineers and operators at the refinery participated in activities aimed at learning from incidents. Incident learning did not just happen through formal incident management processes, but also through daily work practices. Hence, workplace learning may be an interesting lens through which to examine employee practices to learn from incidents. We found that employees executed learning-related tasks in different ways from formal presentation of reports and risk reducing measures to informal meetings and discussion raising the reflexivity of employees. We conclude this paper with recommendations for learning practices in large-scale industrial environments.

From Place/Transition Petri nets to B abstract machines for safety critical systems

Fulfilling the safety requirements is one of the most serious and problematic issues in critical systems design and safety critical software development. In this context, thanks to their analysis power, formal methods have been widely used in the various stages of the design and the implementation of safety critical systems. However, some methods such as the B method, although well adapted to safety issues, are still poorly used in large scale industrial environment. The purpose of this paper is to present a methodology of Place/Transition Petri nets transformation into B abstract machines enabling an interesting combination of the graphical modeling power of Petri nets and the verification tools of the B method. In fact, translating a Petri net to a B abstract machine can have many advantages such as the generation of code, the integration of safety invariants or the aggregation with other formal models. Therefore, the B verification will enlarge the scope of its applicability by having a new modeling alternative and passing through model transformation. An illustrative example of the transformation is presented for a railway study case.

Realization Limits of Impulse-Based Localization System for Large-Scale Indoor Applications

In this paper, a 3D indoor localization demonstrator on the basis of impulse-radio ultrawideband (UWB) technology and time-difference-of-arrival (TDOA) principle is developed and analyzed. The parameters of the transmitter and receiver hardware components are investigated to determine their influence on the localization performance. The signal detection method based on a comparator and the precise time measurement unit was examined. Two effects, namely the threshold-trigger offset and the TDOA variance errors, were quantified. Corrections methods of both these phenomena have been proposed, which delivered very good results in the experiments. With the identified and modeled inaccuracies, the Cramer-Rao lower bound for a 3D TDOA localization system is derived for the first time and verified by measurements, performed in a large-scale industrial environment. The results obtained from measurements follow closely the variance predicted by the CRLB. Moreover, the comparison with the literature published up to date proves the excellent performance of the system presented here.

Automatic Requirement Classification: Tackling Inconsistencies Between Requirements and Regulations

Current Requirement Engineering research must face the need to deal with the increasing scale of today's requirement specifications. One important and recent research direction is handling the consistency assurance between large scale specifications and many additional regulations (e.g. national and international norms and standards), which the specifications must consider or satisfy. For example, the specification volume for a single electronic control unit (ECU) in the automotive domain sums up to 3000 to 5000 pages distributed over 30 to 300 individual documents (specification and regulations). In this work, we present an approach to automatically classify the requirements in a set of specification documents and regulations to content topics in order to improve review activities in identifying cross-document inconsistencies. An essential success criteria for this approach from an industrial perspective is a sufficient classification quality with minimal manual effort. In this paper, we show the results of an evaluation in the domain of automotive specifications at Mercedes-Benz passenger cars. The results show that one manually classified specification is sufficient to derive automatic classifications for other documents within this domain with satisfactory recall and precision. So, the approach of using content topics is not only effective but also efficient in large scale industrial environments.

The effect of scale on gene expression: commercial versus laboratory wine fermentations

Molecular and cellular processes that are responsible for industrially relevant phenotypes of fermenting microorganisms are a central focus of biotechnological research. Such research intends to generate insights and solutions for fermentation-based industries with regards to issues such as improving product yield or the quality of the final fermentation product. For logistical reasons, and to ensure data reproducibility, such research is mostly carried out in defined or synthetic media and in small-scale fermentation vessels. Two questions are frequently raised regarding the applicability of this approach to solve problems experienced in industrial fermentations: (1) Is synthetic medium a sufficiently accurate approximation of the generally more complex natural (and frequently highly variable) substrates that are employed in most fermentation-based industries, and (2) can results obtained in small-scale laboratory fermentations be extrapolated to large-scale industrial environments? Here, we address the second question through a comparative transcriptomic approach by assessing the response of an industrial wine yeast strain fermenting a natural grape juice in small-scale laboratory and large-scale industrial conditions. In yeast, transcriptome analysis is arguably the best available tool to holistically assess the physiological state of a population and its response to changing environmental conditions. The data suggest that scale does indeed impact on some environmental parameters such as oxygen availability. However, the data show that small-scale fermentations nevertheless accurately reflect general molecular processes and adaptations during large-scale fermentation and that extrapolation of laboratory datasets to real industrial processes can be justified.

Large Scale Production Planning in the Stainless Steel Industry

We introduce a new formulation for a cutting stock problem in the stainless steel industry. The formulation extends some previous models developed for the paper industry. In this formulation, mother coils are assumed to have finite lengths, and the widths and lengths of the mother coils are considered variables as well as the lengths of the product coils along with the cutting patterns to be used. To enable reuse of scrap material, the trim loss at the end of a mother coil is preferred and collected into a single coil. Finally, the scattering of an order to multiple mother coils is minimized to facilitate the shipping of the order in a single batch. For numerical solution, the inherently nonlinear problem is transformed to a linear one by a decomposition technique. The procedure is illustrated with numerical examples. The presented model has been successfully utilized in large scale industrial environment as a part of the production planning system developed for Outokumpu Stainless by Accenture.

Detecting and identifying artificial acoustic emission signals in an industrial fatigue environment

This paper details progress in the application of a methodology for acoustic emission (AE) detection and interpretation for the monitoring of fatigue fractures in large-scale industrial environments. The approach makes use of a number of novel signal processing techniques. An online radius-based clustering algorithm (ORACAL) is used to identify clusters of data, both in the spatial domain (locating AE sources) and in the feature domain (identifying candidate fracture processes). The paper proposes a new approach to the identification of AE waveforms produced by crack propagation; rather than seeking to identify the waveform features characteristic of a fracture event, the new method looks for specific patterns of clustering in the feature space. The approach is validated by a full-scale experiment. An artificial acoustic emission source, representative of a fatigue fracture, was injected into a test of a substantial landing gear component. A commercial AE monitoring system was then used to successfully locate and identify the source in a blind test using the new signal processing methodology. The method was successful on two of three experiments performed and the position of the artificial source was determined accurately; further analysis shows that the unsuccessful test appears to have occurred due to incorrect mounting of the artificial source.

Detecting and Identifying Artificial Acoustic Emission Signals in an Industrial Fatigue Environment

This paper details progress towards the application of a methodology for Acoustic Emission (AE) detection and interpretation for the monitoring of fatigue fractures in large-scale industrial environments. An artificial acoustic emission source, representative of a fatigue fracture was injected into a test of a substantial landing gear component. An AE monitoring system was then used to successfully locate and identify the source using the new signal processing methodology.

A Critical Review of Recent Large-Scale Experiments on Hydrogen-Air Detonations

Recent large-scale experiments have indicated that scale and geometric effects can strongly influence both the ability of a fuel-air mixture to propagate or sustain a detonation, and the probability that a deflagration will undergo a transition to detonation. The data show that many past concepts concerning detonations, which were derived from small-scale experiments, should be revised. In particular, the occurrence of detonations may not be as unlikely as previously considered, especially in large-scale industrial environments in which obstacles are present. Some of the important recent experimental research is reviewed and some qualitative ideas concerning the interpretation of the experiments and their implications for improving safety are provided.