جستجوی مقالات مرتبط با کلیدواژه "condition monitoring" در نشریات گروه "مکانیک"

The purpose of condition monitoring is to monitor the conditions of an asset in order to predict its failure. The first step in implementing condition monitoring is to establish a health indicator. To construct the health indicator, operational data such as vibrational data should be collected from the asset during its operation, followed by extracting meaningful features from the data.This study introduces an artificial intelligence model (convolutional autoencoder) for feature extraction from the vibration data, which only requires healthy status data for training. For this purpose, the vibration data of the asset is gathered online during operation in a healthy state to establish the healthy dataset. Then, the deep learning model is trained by the healthy dataset. Finally, after the failure stage is detected, the health indicator is established by measuring the differences in the vibrations of the healthy and the failure conditions. The performance of the proposed model is evaluated by vibrational data of a gearbox. The health indicator exhibits a monotonically increasing degradation trend and has good performance in terms of detecting incipient faults.

Lobe compressors are one of the most important compressors used in the steel industry, and the occurrence of any failure in this equipment leads to production stoppage and imposing heavy losses on steelmaking units. Neglecting to monitor the precise condition of lobe compressors has led to the growth of the vibration range of the compressor as well as the adjacent equipment and piping, and in addition to creating misalignment between the compressor, gearbox and electric motor, it has led to the failure of bearings and the failure of precision instrument connections and branching of piping lines upstream. And it will be downstream of the compressors.In addition to the negative consequences of increasing the cost of repairs and reducing production, this issue will cause risks such as process gas leakage. In this article, the study and solution of the vibration problem of the mentioned equipment in one of the country's steelmaking complexes has been discussed.

Detection of rotor asymmetry faults (RAFs) in induction machines (IMs) based on stator current signature of machine due to the presence of low-frequency load torque oscillation (LTOs) can cause false alarm (FM). Therefore, isolating the RAFs from the LTOs can improve the condition-based monitoring system. The methods discussed in the past generally require three-phase current and voltage of stator windings information along with machine angular velocity. In this paper, a new method based on single phase machine data is presented. In this regard, a virtual rotating reference frame based on Hilbert transform is provided which do not need the angular velocity of machine. In order to improve the output spectrum resolution and low computational cost, the proposed method is combined with the Gortzel algorithm. The proposed method is tested and evaluated by synthetic data and then evaluated by means of experimental results. The results show that this method can isolate the RAFs indices from LTOs, effectively.

In this study, a new method for the fault detection of the locomotive engine injector nozzle based on vibration analysis and statistical tests, inside artificial neural networks, is presented. For this point, first the under study received vibration signals in the frequency domain is divided into several smaller ranges and the RMS of each range is then extracted as a frequency property and given as an input to the neural network. Because the high selection of the features reduces the accuracy of the neural network, the extracted feature vector with different levels of significance passes through the T-test filters, firstly, and then enters the neural network as an input. Using of this method, the accuracy of the neural network increases from 78.4 to 94.6%, and also help to detect the frequency ranges. According to the results, the fault of the injector nozzle crack increases the intensity of vibrations in the upper band frequencies of 1500 Hz.

In recent years, with the advent of the Fourth Industrial Revolution concepts and the development of artificial intelligence technologies, new approaches such as the digital twin have been introduced. In a digital twin, a virtual counterpart of the physical system during its whole life is created, with abilities such as analyzing, evaluating, optimizing, and predicting. The first step in creating a digital twin model is to construct a (multi) digital health indicator that describes different aspects of the physical component state during the whole life of the component. In this research, a new method for constructing health indicators based on vibration measurement and a deep learning model has been introduced. For this purpose, the Continuous Wavelet Transform was used to convert the raw vibration signals into two-dimension images; Then, the deep learning model was used to extract features from the images and the health indicator is constructed based on the differences of the images in normal and failure stages. In this article, various Autoencoder architectures are discussed, and it is demonstrated that the Convolutional Autoencoder has better performance in terms of detecting incipient faults. The performance of the proposed model is evaluated by the vibration data of the bearing, and the constructed health indicator exhibited a monotonically increasing degradation trend and had good performance in terms of detecting incipient faults.

Bearings are one of the main components used in the drive-train of electrical machines. Early fault diagnosis and classification of bearing fault for maintenance of electromechanical system are very important. With progresses in measurement and digital systems, extensive range of real-time data can be available in electrical machines. Since fault diagnosis based on signal processing methods from extracted signals may not be possible due to different reasons such as noise level, natural frequencies of system, saturation of core,, severity of fault and load torque, deep learning methods have been considered in recent years. In this paper, time series deep learning method for condition monitoring of bearing in electrical machine for the purpose of detection and classification of fault is considered. Obtained results by means of proposed method have been compared with pervious techniques. Experimental results show that proposed deep learning method can detect and classify bearing fault with accuracy above 90%.

Since the maintenance and repairing costs of mechanical systems, such as structures and rotating machines are significantly high, one way to reduce these costs is to consider some approaches before any operational work to check for damages in such systems. In this study, a new method is presented for damage detection of offshore jacket structures in the presence of various uncertainties, such as modeling errors, measurement errors and environmental noises, based on the simulated model and intact state of the real model. In the proposed method, real intact structure data is used to update the simulated model parameters. Some parts of the signal that are not related to the nature of the system are removed using the complete ensemble empirical mode decomposition method. Frequency data is extracted from the vibrational signals using the frequency domain decomposition method. A deep auto-encoder neural network is designed to learn the damage-sensitive features from the frequency data and to damage detection of the structure. In order to train the proposed deep network, frequency data of the simulated model and real intact state are used; then the frequency data of the real structure is used to test the proposed deep network. The results show that the proposed method is capable for damage detection of the offshore jacket structure with more accurate results than the other comparative methods.

Early fault detection of the rolling element bearings has a very important role in increasing the reliability of rotating machines.It leads to better decision-making for maintenance activities.  In recent decades, the shock pulse method has been developed to detect faults in the early stage of rolling element bearings degradation. In this paper, the accuracy of the remaining useful life estimation using extracted features from vibration signals and that from the shock pulse method are compared. In this regard, a set of accelerated life tests on rolling element bearings were designed and performed. Both shock pulse signals and vibration signals of the under-test rolling element bearings were recorded. Then two models based on feed-forward neural-network are developed to predict the remaining useful life of rolling element bearings. In the first model, only extracted features from vibration signals are fed for remaining useful life prediction. In the second model, the extracted features from shock pulse method are fed too. The results show that using shock pulse method-based features improves the accuracy of remaining useful life estimation. Also, using the health indicators extracted from vibration analysis and shock pulse method leads to a better estimating of the degradation behavior.

Condition monitoring of mechanical systems, such as structures and rotating machines is always a major challenge. This paper is presented a new method for damage detection of real mechanical systems in presence of the uncertainties such as modeling errors, measurement errors, varying loading conditions, and environmental noises based on a simulated model and real healthy state. In this method, data of a real healthy system is used to updating the parameters of the simulated model. Some parts of the signals that are not related to the nature of the system are removed using the complete ensemble empirical mode decomposition method. A deep convolutional network is designed to learn damage-sensitive features from raw frequency data of simulated model and real healthy state. Raw frequency data is extracted from vibration signals using the power spectral density method. In order to train the proposed deep network, raw frequency data of the simulated model andreal healthy state are used. Then, raw frequency data of the real model are used to test the proposed deep network. The proposed method is validated using an experimental beam structure. The results show that using the proposed algorithm for identification and damage detection of the beam-like structure has more accuracy with respect to the other comparative methods

Adhesive joints have been employing in engineering structures such as marine, aerospace, automotive, oil and gas industry. Since the joints are the weakest part of engineering structures, the fracture possibility in adhesive joint is high. Therefore, structural health monitoring of adhesive joint is an important issue. The aim of this paper was condition monitoring of damage in the single lap adhesive joint by multiwall carbon nanotubes. This work is carried out by recording the electrical resistance change during the mechanical loading. This damage propagation is observed as a crack extension in adhesive joint. Firstly, MWCNTs with 9 wt.% are dispersed in an epoxy resin by ultrasonic device. Then, the hardener is added to the nonoadhesive. The obtained material is immediately poured into single lap adhesive joint mold. The defective adhesive joints were manufactured with circular and square defects in different sizes i.e. 10%, 30% and 70% overlap area. The specimen is subjected to tensile test and electrical resistance change are recorded during the shear load. The results showed that the maximum value of relative resistance change is occurring in the adhesive joint comprises of square defect with 30% overlap area of defect. In this situation, the crack was propagated in nanoadhesive and a small part of a crack was extended from defect boundary.