Low-Cost System for Weld Tracking Based on Artificial Vision

A new laser stripe sensor system for the automation of welding processes in heavy industries is presented. Conventional laser stripe sensors use expensive multiprocessor boards with up to six digital signal processors (DSP) to cope with the high data rate of 14 Mbyte/s delivered by the sensor's CCD-camera. A new solution using only one highly integrated logic IC (FPGA) has been developed offering the same performance as multiprocessor solutions at a drastically lower price making laser stripe sensors attractive for cost sensitive applications in the heavy industries. Integrating the FPGA in the sensor head facilitates data transmission to the sensor computer as the data rate is decreased by a factor of 700 and can thus be transmitted as a robust RS485 serial signal over large distances. For the detection of the location of the welding groove a new robust fuzzy algorithm has been developed that guarantees a satisfying tracking result even when the welding gap geometry strongly varies or is distorted by noise. Using this new method, investigations carried out on industrial samples from the ship-building industry showed that the fuzzy approach was able to cope with strong seam geometry variations and proved to be superior to conventional approaches.

Background: The occurrence of fire and explosion accidents in the process industries is accompanied by the gradual release of large amounts of energy. The DOW Index is one of the methods of evaluating fire and explosion in the process industries. The present study used the fuzzy logic method to improve the results of the fire and explosion risk analysis method and reduce their uncertainty (error in achieving the correct result). Methods: In the present cross-sectional study, having analyzed the gas separation process, the DOW index was applied to study high-risk equipment. Then, to optimize the behavior of the parameters affecting the DOW index, each weight parameter was assigned to between 0 and 1, and the membership functions of each degree of risk were determined based on the fuzzy rules classification system. The five selected linguistic variables based on membership functions were used to assess the risk level. Results: Considering butane, the level of fire and explosion risk was 231.3, which was less risky than the fuzzy logic result of 248.6. Methane risk was calculated according to the fire and explosion index of 262.1 while using fuzzy logic that was 265.6. The lowest risk difference was observed between the two methods for ethane (258.9 conventional method and 259.1 fuzzy method results). The risk level calculated by DOW for propane was 243.6, and the risk level was 255.1 while using fuzzy logic. Conclusion: Although both methods yielded a high degree of risk, fuzzy logic results indicated higher numerical values comparing to the conventional DOW method. Compared to the conventional DOW method, fuzzy logic results are closer to reality with higher confidence levels.

特集 製造業におけるニューコンセプト 未来製造技術をもとめて 重工業におけるニューコンセプト

With recent advances in robustness and design predictability of additive manufacturing (AM), this production method has taken its first steps towards implementation in heavy duty industrial applications such as gas turbines and heat exchangers. Among the benefits of using AM, the most important may be the significant higher degree of freedom in design than when using conventional manufacturing methods. In order to take advantage of these new possible solutions and to consider the surface roughness that arises from AM, it is necessary to have reliable correlations for pressure losses and heat transfer. In this study, the thermal performance for additively manufactured circular channels in Inconel 939 using the Selective Laser Melting (SLM) process are experimentally investigated.

An vision inspection sensor system for the automation of laser welding processes in heavy industries is presented. The system consists of PC based vision camera and stripe type laser diode. For image de-noising, a new surface inspection method of Tailor-welded blanks including in: median filtering, bi-level thresholding, opening and closing, mid-line of the laser stripe extraction, has been developed. After extracting the middle line of the laser stripe, the algorithms of calculating bead width, mismatch, weld slope, convexity, concavity, overthickness and undercut are proposed. Finally, A series of experiments show that the proposed method in this paper is effective.

Multipoint pressure sensing at up to 900 °C using a fiber optic multimode interferometer

The creation and application of classification category labels are essential tasks for transforming complex information into structured knowledge. Categories are used for summary and reporting purposes and have historically been identified by domain experts based on their past experiences and norms. Our interest lies in the general case where expert-generated category lists require improvement, and unsupervised learning, on its own, struggles to effectively identify categories for multi-class classification of human-generated texts. We hypothesise that including an annotated knowledge graph (KG) in an embedding process will positively impact unsupervised clustering performance. Our goal is to identify clusters that can be labelled and used for classification. We look at unsupervised clustering of Maintenance Work Order (MWO) texts. MWOs capture vital observations about equipment failures in process and heavy industries. The selected KG contains a mapping of equipment types to their inherent function based on the IEC 81346-2 international standard for classification of objects in industrial systems. Performance is assessed by statistical analysis, subject matter experts, and Normalized Mutual Information score. We demonstrate that Word2Vec Bi-LSTM and Sentence-BERT NN embedding methods can leverage equipment inherent function information in the KG to improve failure mode cluster identification for the MWO. Organisations seeking to use AI to automate assignment of a failure mode code to each MWO currently need test sets classified by humans. The results of this work suggest that a semantic layer containing a knowledge graph mapping equipment types to inherent function, and inherent function to failure modes could assist in quality control for automated failure mode classification.

Coke drums experience one of the most severe cyclic services of any common pressure vessel type in the process or heavy industries. While the severe loading is thermal in nature, with an inherently random and statistical aspect to it, the end results are very predictable. Namely that cracking adjacent to the girth (or circumferential) weld seams, initiating overwhelmingly from the inside surface, tends to set the economic life of the drums. The concept of calibrating loading to known cracking can be leveraged, but to do so requires meaningful fatigue initiation and crack growth data for typical coke drum welded joints. Such data is reviewed and analyzed, and property recommendations made for both the inside (with clad restoration) and outside (simple matching weld joint) including appropriate test-based fatigue strength reduction factors. Focus is given to the common coke drum material SA-387 Grade 11 Cl. 2 (1.25Cr-0.5Mo) due to the volume of data available for it, and typical coke drum welded joints. Coke drum thermal loading is reviewed to develop realistic load ranges for use with the property recommendations to validate that they are capable of giving meaningful life predictions.

The plant cell wall is a very complex and dynamic system, similar in importance to both the extracellular and intracellular processes which are recognised nowadays. The cell wall is a “vanguard” an effective barrier in the way of different negative chemical and biotic factors, including pathogens and wounding. The defence functions of plant cell walls are associated with the construction of physical barriers consisting of ligninand suberincontaining polymers on the path of pathogens inside a plant. This reaction develops more or less automatically, and barriers are only formed in the zone of pathogen penetration during active pathogen expansion into the host plant’s tissues. Yet the mechanisms of these events are still unclear. It is well-known that peroxidases (PO) are key enzymes involved in lignification (Cosio & Dunand, 2009) and one of the few proteins secreted into the plant cell wall. However, POs have numerous applications in industry and one of the most important of these is the use of POs for lignin degradation. Therefore, both analytics and industry require a great volume of stable PO preparations of a high quality and at a low price, and the search for new methods and substrates for their extraction and purification has great commercial importance. Thus, the ability of plant oxidoreductases to interact with the biopolymers of the cell walls of plants and fungi has been studied for several decades (Siegel, 1957; McDougall, 2001). Moreover, it has been shown that plant POs can bind electrostatically with calcium pectate (Dunand et al., 2002) and chitin (Khairullin et al., 2000). Plants are likely to contain a whole subclass of these “polysaccharide-specific” isoPOs and their encoding genes. This subclass should be characterised by the ability to bind with polysaccharides and the defence function focused on strengthening the cell wall of the host and isolating the non-infected host tissues from the pathogen with the help of lignin. We suppose that the ability of plant POs to interact with some biopolymers without losing their activity can be applied for the isolation and purification of these enzymes. The possibility of the application of chitin in agriculture, biomedicine, biotechnology and the food industry has received much attention due to its biocompactibility, biodegradability and bioactivity. The low price and the ecological safety of this biopolymer define it as an available matrix for technological processes. As such, it may be possible to produce the high-quality preparations of POs that are needed for various fields of industry and analytical methods with the use of chitin (or other

In heavy industries, power curtailment due to any reason is never tolerable. Transformers are critical component in the supply of power. This paper presents a smart online condition monitoring system for industrial applications of power/distribution transformer monitoring and condition analysis. The health index (HI) of transformer is calculated based on weightage associated with healthy sensor input optimized by genetic algorithm (GA). The experimentation is performed on 15 kVA, 400/400 volts transformer. Health indices are computed by conventional numerical method and GA method. Latter offers improved weighing approach than the former. Therefore, proposed system offers development of a smart distribution transformer.

The recent innovations in both industrialization and informatisation technologies have culminated into a new generation of manufacturing paradigm. This manufacturing innovation promises mass customization due to flexibility in manufacturing, improved product quality and safety, new productivity levels, creation of smart products and services. Industry 4.0 is gradually making its presence felt in Indian Industries. Indian Engineering Industries being the largest segment in the Industry, holds immense potential if it embraces Industry 4.0. This research conceptually examines the effect of Industry 4.0 on Indian Engineering Industries on both the heavy and light engineering industries. ABCD framework is used to investigate the effect of Industry 4.0 in these Industries. The study finds that Indian engineering industry can be transformed to produce smart products and services, by the application of Industry 4.0. This is the first study to evaluate the effect of Industry 4.0 on Indian Engineering Industries. It also contributes to the theoretical need for evaluating the effects of Industry 4.0 in the developing world.

Particle Swarm Optimization (PSO) with highly adaptive parameters and Chaotic Local Search (CLS) has been developed to obtain superior and robust convergence pattern. Depending on the objective function values of the current and best solutions in the present iteration, unique and innovative formulae are designed for two sets of PSO parameters, inertia weight & learning factors, to make them adaptive. To enrich the searching behavior and to avoid being trapped into local optimum, CLS is incorporated treating each individual particle as separate entity. Considering recent necessity and to prove the robustness and better effectiveness of the Chaotic Particle Swarm Optimization (CPSO) based algorithm, authors choose its application in power industry, as power flow has complex and non-linear characteristics. To the best of our knowledge, there is no published work on the CPSO to solve the power flow problems. PSO parameters are set to give better and reliable convergence characteristics for power flow under critical conditions like high R/X ratio and loadability limits. Conventional methods like Newton Raphson/Fast-decoupled load flow can not give multiple power flow solutions which are essential for voltage stability analysis. Proposed algorithm can overcome that limitation. The effectiveness and efficiency has been established showing results for standard and ill-conditioned systems.

In the present study, the alginate was extracted from Sargassum ilicifolium using conventional (CON), enzyme (Alcalase), ultrasound (UAE), and Alcalase-ultrasound (ALUAE) methods. Subsequently, the extracted alginates were used for the preparation of edible and biodegradable films. Results demonstrated that the application of different green extraction methods decreased the molecular weight (Mw) of extracted alginates compared to the CON method, which among them the UAE-extracted alginate had the lowest Mw. Green methods applied to alginate extraction also showed significant effects on different properties of alginate films. In this regard, films that are prepared from alginate with lower Mw (UAE-extracted ones) showed lower water vapor permeability (WVP) and light transmittance as well as higher tensile strength (TS) values. Conversely, the highest WVP and lowest TS were measured for films prepared from ALUAE-extracted alginate. These results are consistent with the SEM micrograph. Novelty impact statement Alginate is a polysaccharide with various applications in different industries. This polysaccharide is extracted using the conventional method (CON). Finding a greener and faster method is preferred by the industry. In the current study, alginate of Sargassum ilicifolium was extracted using different green extraction methods further to CON. The impact of green methods on the properties of films produced from extracted alginate was investigated for the first time. The results showed that the green extraction methods led to films with different properties and among them, the films produced from ultrasound-extracted alginate showed the lowest WVP and highest TS.

A new fast sensor for simultaneous high temperature diagnostics (above 800 K) of nitrogen oxide (NO) concentration and gas temperature (T) was developed based on the spectral fitting of low-resolution NO UV absorption near 226 nm. The sensor was intended for process control in future low-carbon footprint heavy process industries using renewable powered electro fuels (e.g. H2, NH3) or plasma torches as heat source. Due to excitation of molecular vibration, the shape of the selected NO feature, including (0, 0), (1, 1), and (2, 2) vibrational transitions of the A2Σ+ − X2Π2 electronic system had a strong temperature sensitivity at temperatures above 800 K. The fitting was made using the well-known NO molecular constants of the A2Σ+ − X2Π2 electronic system. To reduce the computational time, a library of the molecular spectra calculated at different temperatures was created. The fitting of an experimental spectrum representing the convolution of the instrument line function of the spectrometer with the molecular spectra was performed using the pre-calculated library spectra. Based on comparison with conventional measurement methods, the accuracy of the developed sensor was within 15% for NO and about 40 K for T, clearly showing the potential for fast in situ diagnostics in hot process gases.

Goal: Over the decades, there have been improvements in the sequence alignment algorithm, with significant advances in various aspects such as complexity and accuracy. However, human-defined algorithms have an explicit limitation in view of developmental completeness. This paper introduces a novel local alignment method to obtain optimal sequence alignment based on reinforcement learning. Methods: There is a DQNalign algorithm that learns and performs sequence alignment through deep reinforcement learning. This paper proposes a DQN x-drop algorithm that performs local alignment without human intervention by combining the x-drop algorithm with this DQNalign algorithm. The proposed algorithm performs local alignment by repeatedly observing the subsequences and selecting the next alignment direction until the x-drop algorithm terminates the DQNalign algorithm. This proposed algorithm has an advantage in view of linear computational complexity compared to conventional local alignment algorithms. Results: This paper compares alignment performance (coverage and identity) and complexity for a fair comparison between the proposed DQN x-drop algorithm and the conventional greedy x-drop algorithm. Firstly, we prove the proposed algorithm's superiority by comparing the two algorithms’ computational complexity through numerical analysis. After that, we tested the alignment performance actual HEV and E.coli sequence datasets. The proposed method shows the comparable identity and coverage performance to the conventional alignment method while having linear complexity for the n}{}X parameter. Conclusions: Through this study, it was possible to confirm the possibility of a new local alignment algorithm that minimizes computational complexity without human intervention.