Search

With the important role in detecting electrical equipment defect, infrared thermography is widely used in power industries. With the rapidly increasing data of thermal images, the automatic processing for thermal images is more and more important. This paper presents a new approach to adjust the display temperature range of electrical equipment infrared thermal images. We first take advantage of the temperature distribution of thermal images to find two temperature peaks representing foreground and background respectively, and then find the new display temperature range according to the peaks. In the experiments, we collect 2000 images with improper display temperature ranges to validate the performance of our method. The experiment results show that our approach can effectively improve the image quality of thermal images, which helps the analysis of infrared thermal images.

Different decomposition component gases caused by different insulation defects in gas-insulated switchgear will accelerate insulation degradation, which is a major threat to human and equipment safety. In this paper, we proposed online monitoring of GIS based on Fourier transform infrared spectroscopy. Firstly, we analysed the absorption characteristics of infrared light in SO2 and moisture. Secondly, we introduce the design principle and signal stabilization method of the laser. Thirdly, we propose the methods for measuring the concentration of SO2 and moisture. Finally, experiments with various concentration of SO2 and moisture are deployed, which proves that the GIS on-line monitoring method for decomposed gas has small measurement error and can be applied in practical applications.

While the standardization of live detection process helps ensure safe operation of substation equipment significantly, data transmission between detection equipment and the handheld tablet is still a great challenge. In this article, we propose a communication method for live detection in substations with D2D in licensed band and unlicensed band. Firstly, we analyze the scenario of live detection in substation and compare various near field communication methods. Secondly, we study the key technologies of D2D mainly including spectrum resource management and channel access mechanism. Finally, we perform system level simulation on D2D in licensed and unlicensed band, whose results show that modified D2D can achieve high quality communication between detection equipment and the handheld tablet.

Since the traditional relational database is difficult to cope with the massive data and properties produced by thousands of sensors in the smart grid, the HBase database of the new large data platform, as a column oriented key value storage database, has gradually become the mainstream large data platform database. Though HBase can support large data volume, but it is still difficult to cope with the demand of high frequency and short response time in power grid data analysis. In this study, we proposed a query oriented two level index scheme that can speed up the query. The experimental results show that when the two table connection queries are involved, our scheme can provide a minimum of 1.026 to the maximum 4.761 times the acceleration ratio compared to the classic two level index, when it involves three table connection queries, Our scheme provides a minimum of 1.797 to a maximum of 8.581 times the speed ratio of the classic two level index scheme. After further optimization, the scheme can also save much of the storage space of the index table. The two level indexing scheme proposed in this study has a good effect in terms of query performance and storage efficiency.

The tensile strength of cold-drawn commercially pure aluminum wire shows an obvious three-stage characteristic, including the first strengthening stage, steady stage and second strengthening stage, with increasing extrusion strain. The dislocation density, intensity of 〈111〉 texture and percentage of high-angle grain boundaries are apparently increased in the second strengthening stage. Consequently, the final strengthening mechanisms may be well explained by the existence of the high-density dislocation, the 〈111〉 texture and high-angle grain boundary.

In the field of high-voltage electric equipments, the SF 6 gas as insulation material is widely used in various kinds of Gas Insulated Switchgears (GIS). And the decomposition of SF 6 gas under discharge would generates various kinds of gases, which may seriously result in worse insulation performance of the high-voltage electric power equipment. Thus, it is an urgent issue to monitoring these gases. However, existing inspection methods still fail to conduct on-line monitoring of SF 6 decomposed gases. Aimed to solve this problem, this paper proposes an On-line SF 6 decomposed gases monitoring approach based on the Quantum-Cascade Laser Absorption Spectroscopy. In this system, the composition and the concentration of each kind of gas can be obtained by testing the absorption peak value and absorption intensity of the wavelength. Experiment results show that the proposed system is able to perform on-line monitoring of SF6 decomposed gases with high accuracy.

Convolutional networks are powerful visual models that transform images into more effective representations. To make full use of this technique, we propose a new method based on deep learning and convolutional network to effectively get the discharge information of an ultraviolet (UV) image. We firstly segment the equipment region and the UV spot region separately by using the DeepLab network, and then several properties which can show the discharge information are extracted on the basis of the segmentation. The use of the DeepLab network helps to get a reliable segmentation result, and more accurate discharge information. We take 5000 UV images to test our network, and use the concept mean IOU to evaluate its performance. The results show the advantage of the method, and it can meet the demands of further fault diagnosis.

As an important kind of data for device status evaluation, the increasing infrared image data in electrical system puts forward a new challenge to traditional manually processing mode. To overcome this problem, this paper proposes a feasible way to automatically process massive infrared fault images. We take advantage of the imaging characteristics of infrared fault images and detect fault regions together with its belonging device part by our proposed algorithm, which first segment images into superpixels, and then adopt the state-of-the-art convolutional and recursive neural network for intelligent object recognition. In the experiment, we compare several unsupervised pre-training methods considering the importance of a pre-train procedure, and discuss the proper parameters for the proposed network. The experimental results show the good performance of our algorithm, and its efficiency for infrared analysis.

Recently trade-off among producers, customers and society/ environment is one of the most popular concerns in industry. This issue conducts the decision making by manufacturers on the selection between two alternative techniques, ECM and MM. The purpose of this study was to explore the problem in how to assess the value of a product/ service in the early stages of the design activity. Through the simulation tool of System Dynamics, a model helps to quantify all the values which were involved. The result of this model was that the coming 10 years NPVs of two alternatives were assessed. Contrasting the alternatives on different manufacturing processes and market conditions, the decision makers could find out an appropriate scenario with an appropriate technique. This approach was verified to ensure the accuracy and practicability.

With an increasing number of traffic accidents, more and more people are injuring or even dying during the collision. One of the main reasons is frontal collisions. Therefore, the safety system is a very important part of the vehicle. In order to ensure the passengers’ safety, the common way to minimize mortality and property damage in a collision is to install energy absorption device in the vehicle structure called crash box. We propose to implement the principle of the origami pattern to improve the traditional bumper system, let the bumper have folding property. Origami properties make it useful to apply in a variety of engineering fields. The thin-walled tube and beam will undergo elastic deformation to decrease the impact force and absorb more kinetic energy. We use Autodesk Inventor 2012 to model the seamless origami pattern tube, and do the simulation by ABAQUS to analyze the displacement of axial compression and the impact force. The numerical results show that origami patterns perform well. Then we can obtain the plot with relationship between force and displacement in the elastic behavior and in the elastic-ideal plastic behavior. As for manufacturing field, we analyze existing manufacturing methods include stamping and tube hydroforming, compare them to find their features. Meanwhile we try to explore a new way to improve the accuracy and efficiency of manufacturing origami pattern tube. Med ett ökande antal trafikolyckor, fler och fler människor att skada eller till och med dö under kollisionen. Ett av de viktigaste skälen är frontalkollisioner. Därför är säkerhetssystemet en mycket viktig del av fordonet. För att garantera passagerarnas säkerhet, är det vanligaste sättet att minimera dödlighet och skador på egendom vid en kollision för att installera enheten energiupptagning i fordonets struktur som kallas crash box. Vi föreslår att genomföra principen om origami mönstret att förbättra traditionella stötfångaren systemet, låt stötfångaren har fällbara egendom. Origami egenskaper gör det användbart att tillämpa i en mängd olika tekniska områden. Den tunnväggigt rör och strålen kommer att genomgå elastisk deformation för att minska stötkraften och absorbera mer kinetisk energi. Vi använder Autodesk Inventor 2012 att modellera sömlösa rör origami mönster, och göra simuleringen av ABAQUS att analysera förskjutning av axiell kompression och slagstyrka. De numeriska resultaten visar att origami mönster prestera bra. Då kan vi få tomten med förhållandet mellan kraft och förskjutning i den elastiska beteendet och i det elastiska-ideala plast beteende. När det gäller tillverkning område, vi analyserar befintliga tillverkningsmetoder inkluderar stämpling och rör hydroformning, jämföra dem att hitta sina funktioner. Samtidigt försöker vi att utforska ett nytt sätt att förbättra noggrannheten och effektiviteten i tillverkningen origami mönster röret.

Big data platform for equipment condition assessment is built for comprehensive analysis. The platform has various application demands. According to its response time, its application can be divided into offline, interactive and real-time types. For real-time application, its data processing efficiency is important. In general, data cache is one of the most efficient ways to improve query time. However, big data caching is different from the traditional data caching. In the paper we propose a distributed cache management framework of big data for equipment condition assessment. It consists of three parts: cache structure, cache replacement algorithm and cache placement algorithm. Cache structure is the basis of the latter two algorithms. Based on the framework and algorithms, we make full use of the characteristics of just accessing some valuable data during a period of time, and put relevant data on the neighborhood nodes, which largely reduce network transmission cost. We also validate the performance of our proposed approaches through extensive experiments. It demonstrates that the proposed cache replacement algorithm and cache management framework has higher hit rate or lower query time than LRU algorithm and round-robin algorithm.