Application of Text Mining Method for Classification of Work Order in Power Grid Production

With the completion of the State Grid Corporation’s maintenance system, the number of substations has increased dramatically, the grid structure has become increasingly complex, and there have been internal and external reasons such as the contingency of emergencies, and equipment failures have occurred from time to time. This paper aims to explore the potential value of massive data, show the laws of business data, and further give full play to the comprehensive support of data for enterprise operation and production management, and promote the realization of intelligent and lean power grid core business. This paper uses power system data to provide reliable data support for equipment defect full cycle management and equipment state analysis through ANOVA and neural network statistical analysis. At the same time, we use Term Frequency-Inverse Document Frequency(TF-IDF)Algorithm to calculate the importance of keywords and construct the power keyword library. By constructing Bayesian text classification model, we can classify the defect parts, defect categories and defect causes automatically. This method can be applied to the construction of power grid production work order text analysis system, improve the data quality and system automation level, help the business department to improve work efficiency and provide the basis for power grid business analysis. This method is applied to the data cleaning of the primary production equipment of power grid enterprises, and the accuracy of data error correction for equipment defects with voltages above 110kV is between 93% and 95%, and good results have been achieved.

Cyber-Physical LPG Debutanizer Distillation Columns: Machine-Learning-Based Soft Sensors for Product Quality Monitoring

Refineries execute a series of interlinked processes, where the product of one unit serves as the input to another process. Potential failures within these processes affect the quality of the end products, operational efficiency, and revenue of the entire refinery. In this context, implementation of a real-time cognitive module, referring to predictive machine learning models, enables the provision of equipment state monitoring services and the generation of decision-making for equipment operations. In this paper, we propose two machine learning models: (1) to forecast the amount of pentane (C5) content in the final product mixture; (2) to identify if C5 content exceeds the specification thresholds for the final product quality. We validate our approach using a use case from a real-world refinery. In addition, we develop a visualization to assess which features are considered most important during feature selection, and later by the machine learning models. Finally, we provide insights on the sensor values in the dataset, which help to identify the operational conditions for using such machine learning models.

Evaluation of the sensitizing effect of a disinfectant based on glutaraldehyde upon inhalation

When investigating new disinfectants, it is necessary to assess the degree of their danger by modeling the conditions of use. As is known, the main danger of disinfectants, especially volatiles, is posed by the inhalation route of entry into the body, exerting an irritating effect on the respiratory organs, as well as a general toxic and sensitizing effect. Despite reasonable restrictions on the use of disinfectants from the aldehyde class for decontamination of the surfaces of premises and equipment, state registration receives disinfectants based on glutaraldehyde (GA), intended for wide use in various areas, including everyday life, which is unacceptable. Experiments were carried out to confirm the danger of using GA-based disinfectants. Sensitizing effect of GA-based disinfectant under conditions of surface treatment by rubbing method is studied. Significant release of GA into the air after surface treatment (27.7±4.1 mg/m3) was detected. The concentration of GA in the air decreased by 4.5 times after 6 hours of ventilation. The sensitizing effect of the GA-based agent was determined on guinea pigs by a provocative ear swelling test and a leukocyte specific lysis reaction. The results of the study indicate a pronounced sensitizing effect of the agent. Keywords: glutaraldehyde, toxicity, disinfectants, disinfection, sensitization, inhalation, guinea pigs.

Intelligent Fault Diagnosis of Photoelectric Pod Bearing Based on Multi-information Fusion

Intelligent fault diagnosis technology has become the focus of research in various fields. Its realization depends on the acquisition of equipment state by sensors. Because the fault information provided by a single sensor has limitations and cannot fully reflect the fault state of the tested object, we need to use multiple sensors to collect and fuse the fault information of rolling bearings to ensure the accuracy and accuracy of intelligent fault diagnosis. Based on this, this paper analyzes the application of fuzzy rules of multi-sensor information fusion technology in the fault diagnosis of bearings in the optoelectronic pod, so as to provide a reference for the realization of intelligent fault diagnosis of each structure in the optoelectronic pod.

Indirect estimation of the production rate of wells with low flow rate

The problem of measuring the flow rate of wells with low production rates is relevant for many oil fields. Conventional flow meters are not suitable for such cases, and installing an additional flow meter for each well is impractical. At the same time, wells with sucker-rod pumping units (the majority of wells) are outfitted with dynamographs for continuous diagnostics of the pumping equipment state. Dynamograms allow determining the theoretical flow rate of the well easily, however, a mathematical model is required to estimate the actual flow rate. For the correction of flow rate obtained from dynamograms, the authors of this study propose using models based on regression equations that link the calculated valueswith the measurements made by a reference instrument. The results of the experiments have confirmed the eligibility of this approach.

Cyber-Physical LPG Debutanizer Distillation Columns: Machine Learning-Based Soft Sensors for Product Quality Monitoring

Refineries execute a series of interlinked processes, where the product of one unit serves as the input to another process. Potential failures within these processes affect the quality of the end products, operational efficiency, and revenue of the entire refinery. In this context, implementation of a real-time cognitive module, referring to predictive machine learning models, enables to provide equipment state monitoring services and to generate decision-making for equipment operations. In this paper, we propose two machine learning models: 1) to forecast the amount of pentane (C5) content in the final product mixture; 2) to identify if C5 content exceeds the specification thresholds for the final product quality. We validate our approach by using a use case from a real-world refinery. In addition, we develop a visualization to assess which features are considered most important during feature selection, and later by the machine learning models. Finally, we provide insights on the sensor values in the dataset, which help to identify the operational conditions for using such machine learning models.

D-dCNN: A Novel Hybrid Deep Learning-Based Tool for Vibration-Based Diagnostics

This paper develops a novel hybrid feature learner and classifier for vibration-based fault detection and isolation (FDI) of industrial apartments. The trained model extracts high-level discriminative features from vibration signals and predicts equipment state. Against the limitations of traditional machine learning (ML)-based classifiers, the convolutional neural network (CNN) and deep neural network (DNN) are not only superior for real-time applications, but they also come with other benefits including ease-of-use, automated feature learning, and higher predictive accuracies. This study proposes a hybrid DNN and one-dimensional CNN diagnostics model (D-dCNN) which automatically extracts high-level discriminative features from vibration signals for FDI. Via Softmax averaging at the output layer, the model mitigates the limitations of the standalone classifiers. A diagnostic case study demonstrates the efficiency of the model with a significant accuracy of 92% (F1 score) and extensive comparative empirical validations.

Computer wireless network integration technology in equipment state data collection

The paper introduces a handheld integrated power data acquisition and analysis equipment based on a computer wireless network mobile platform, an intelligent transportation inspection box. The intelligent transportation inspection box acts as a bridge between the transportation inspection work site side and the power grid company center to realize data and resource sharing between the site and center sides. The application of intelligent operation and inspection boxes based on computer wireless network integration technology is essential for improving the professional and intelligent level of on-site operation and maintenance, improving the efficiency of operation and inspection work and the ability to control equipment status, strengthening state maintenance and auxiliary decision-making, and promoting the construction of smart grids significance.