Modified Ensemble Empirical Mode Decomposition Research Articles

Pulse component extraction and its application in simulation of pulse-like ground motions

A procedure for simulating new pulse-like ground motion based on natural pulse-like ground motion record is proposed. Firstly, an extraction method of pulse waveform based on Modified Ensemble Empirical Mode Decomposition is proposed, which reveals the relationship between half-wavelength width and pulse time-frequency information. Secondly, according to the response spectrum of the extracted non-pulse time-history, a triangular series is used to generate a stationary time-history, and then the extracted pulse component is added to the time-history, and the intensity envelope is adjusted to obtain the simulated pulse-like ground motion. Finally, the seismic response results of the Bouc-Wen-Baber-Noori model and the three-dimensional finite element model of the high-rise building are validated. It shows that the nonlinear effect of the simulated pulse-like ground motion by this method is close to that of the original natural pulse-like ground motion. The simulated pulse-like ground motion by this method can be used to analyze the inelastic response of various structures.

Can hybrid model improve the forecasting performance of stock price index amid COVID-19? Contextual evidence from the MEEMD-LSTM-MLP approach

The sudden eruption of COVID-19 has inflicted tremendous damage to the worldwide economy, and stock markets have become violently volatile due to its negative impact. Therefore, accurate forecasting of stock price index has been playing an essential role in maintaining national economic security and formulating related policies. In this paper, a novel decomposition-ensemble model is proposed to predict the highly fluctuating stock price index. To begin with, the modified ensemble empirical mode decomposition (MEEMD) method is adopted to decompose the original stock price index into subsequences with different frequencies. Then, the last high-frequency subsequence and other subsequences are predicted through multilayer perceptron (MLP) and long short-term memory (LSTM), respectively. Finally, the prediction outcomes of different model subsequences are reconstructed into the ultimate prediction results by utilizing the integration method. Compared with the contrast models, the MEEMD-LSTM-MLP model proposed in our paper not only demonstrates significant advantages in multi-step forecasting for both emerging and developed markets, but also achieves excellent prediction performance amidst the severe market fluctuations triggered by COVID-19. Furthermore, the application of the MEEMD-LSTM-MLP model is extended to financial time series with different data characteristics and market types, which further proves its high applicability and reliability. Therefore, the conducted hybrid MEEMD-LSTM-MLP model is an effective and stable multi-step forecasting tool to provide valuable intelligent technical support for governments and enterprises in complex economic conditions.

Fault diagnosis of reciprocating machinery based on improved MEEMD-SqueezeNet

Reciprocating machinery structure is complex, and its signal is nonlinear and nonstationary. Because of the complex working environment, the signal usually contains noise interference. Modified Ensemble Empirical Mode Decomposition(MEEMD) reduces residual noise by introducing permutation entropy. But it directly eliminates components larger than the threshold, which may lose the effective information of the original signal. Deep learning models can solve nonlinear problems. But with the number of network layers increases, it increases the number of parameters, resulting in long model training time. Based on the above analysis, this paper proposes a fault diagnosis of reciprocating machinery based on improved MEEMD-SqueezeNet. By introducing multi-scale permutation entropy to replace the original single-scale permutation entropy, the calculated entropy contains more scale information. Singular value decomposition (SVD) is carried out to denoise the signal larger than the threshold to retain more effective information in the original signal. Convert the noise reduced 1D time series to 2D images by Gram Angle Field(GAF) to preserve the temporal correlation of 1D vibration data. Combined with the idea of Inception series model, an InFire module is proposed to replace the original Fire model of SqueezeNet. Use the improved SqueezeNet to extract features from the generated 2D images, and finally input the softmax classifier for classification to obtain fault diagnosis results. The simulation experiment and noise reduction comparison experiment prove that the improved MEEMD method proposed in this paper can effectively achieve signal noise reduction and improve the overall average accuracy of diagnosis from 89.49% to 94.35%. The fault classification comparison experiments prove that the accuracy of improved SqueezeNet is basically the same as that of VGG19. Compared with SqueezeNet, the accuracy of improved Squeezenet is improved by 1.77%, and the number of model parameters is reduced by nearly 29.4%. In terms of training time, it is proved that the training time of the improved SqueezeNet is basically the same as that of SqueezeNet, accounting for only one-third of the training time of VGG19. The improved SqueezeNet not only ensures lower training time, but also improves the accuracy of SqueezeNet. The above experimental results show that the proposed method reduces the model training time and improves the fault diagnosis accuracy of reciprocating machinery.

An assembly tightness recognition method for bolted connection states with singular-value entropy and GA least-squares support vector machine

The bolted connections are widely used in high-end equipment such as aerospace and energy fields. However, when the equipment is assembled, the assembly tightness of bolted connection structure is easily affected by friction and manual experience, resulting in inconsistent preload. It will affect the dynamic performance of equipment, resulting in abnormal vibration or failure of equipment. Currently, the existing assembly tightness testing methods mainly focus on whether the bolts are loose, ignoring the impact of assembly tightness on the vibration of the whole machine. This paper proposes an assembly tightness recognition method for bolted connection states based on modified ensemble empirical mode decomposition (MEEMD) and genetic algorithm least squares support vector machine (GALSSVM). Firstly, an ultrasonic pre-tightening force test system is used for multi-bolt pre-tightening force calibration and precise control, and subsequently, accurate vibration signal samples are collected. The feature extraction of multi-bolt connection states is carried out based on MEEMD. Secondly, a quantitative evaluation index of assembly tightness, singular-value entropy, is proposed. Finally, experimental verification is performed. Multiple sets of different preloads case experiments are designed to verify the identification accuracy of the proposed algorithm. The results show that with the bolted connection structure from loose to tight, the singular-value entropy of the quantitative index decreases monotonically. It provides some theoretical references for improving the reliability of high-end equipment and the evaluation of assembly performance.

Carbon price prediction based on multi-factor MEEMD-LSTM model

China’s national carbon market has already become the largest carbon market in the world. The prediction of carbon price is extremely important for policymakers and market participants. Therefore, the prediction of carbon price in China is of great significance. To achieve a better prediction effect, a multi-factor hybrid model combined with modified ensemble empirical mode decomposition (MEEMD) and long short-term memory (LSTM) neural network optimized by machine reasoning system on the basis of production rules is proposed in this paper. In addition to historical carbon price, other factors, such as energy, macroeconomy, environmental condition, temperature, exchange rate which affect carbon price fluctuation, are formed as multi-factor. The change characteristics of carbon price time series data and other associated factors are extracted in the carbon price prediction. The MEEMD is used to decompose data which is taken as potential input variables into LSTM neural network for prediction and the machine reasoning system based on production rules can automatically search and optimize the parameters of LSTM to further improve the prediction results. The experimental results demonstrate that the proposed method has better prediction effect, robustness and adaptability than the LSTM model without MEEMD decomposition, the single factor MEEMD-LSTM method and other benchmark models. Overall it seems that the proposed method is an advanced approach for predicting the non-stationary and non-linear carbon price time series.

A novel hybrid model integrating modified ensemble empirical mode decomposition and LSTM neural network for multi-step precious metal prices prediction

Accurately predicting precious metal prices is of extreme significance as they possess an essential position in both financial and industrial fields. To achieve higher prediction accuracy, in this paper, a modified ensemble empirical mode decomposition (MEEMD) method integrated with long short-term memory neural network (LSTM) is used for precious metal price prediction. The multi-scale permutation entropy (MPE) analysis demonstrated that MEEMD has better decomposition effect than ensemble empirical mode decomposition (EEMD). Then input each intrinsic mode function (IMF) obtained by MEEMD into LSTM for prediction. Finally add each IMF forecasting value to get the final prediction results. Compared with traditional multilayer perceptron neural network (MLP), support vector regression (SVR) and a combination forecasting model super learner (SL), MEEMD-LSTM improves the prediction performance in one-step ahead prediction and multi-step ahead prediction. The multi-horizon model confidence set (MCS) test are adopted to comprehensively and statistically testify the best prediction performance of MEEMD-LSTM. Furthermore, this research indicates that the model still shows better prediction accuracy either under different proportions of the training set to the test set or in different periods of business cycle. This competitive precious metal price forecasting model is a promising technique for government agencies, investors and related enterprises.

Marine SINS Self-Alignment Method in Complex Interference Environment Based on MEEMD Algorithm

It is difficult to rapidly and accurately achieve alignment under complicated interferential environmental conditions with a marine strapdown inertial navigation system (SINS) subject to angular swaying interference and linear motion interference. To solve this problem, an SINS self-alignment algorithm in a complex interference environment is presented in this article. For the angular motion disturbance problem, we used the attitude-updating algorithm in the coagulation inertial coordinate system to reflect the carrier’s attitude change in real time under the swaying interference so as to eliminate the influence of angular motion disturbance. To solve the problem of linear motion interference, we first use Fourier transform analysis (FFT) to analyze the frequency characteristics of linear motion interference signals. We then adaptively decompose the original signal into a series of signal components with different time scales and local characteristics according to the amplitude and frequency using the modified ensemble empirical mode decomposition (MEEMD) algorithm. Finally, we remove and reconstruct the disturbing signal components according to the frequency characteristics of the disturbing signal to filter the line motion disturbing signal. We use the MEEMD algorithm to conduct a ship alignment test. After pre-filtering by MEEMD, high accuracy can be achieved after 100 s of alignment, and the mean square error of the heading angle is 2.85′. The test results show that the MEEMD method has a good noise removal effect. The anti-jamming initial alignment algorithm based on the MEEMD algorithm has the ability to suppress the interference of angular motion and linear motion and achieve high-precision anti-jamming initial alignment under the complicated interferential environmental conditions with angular swaying interference and linear motion interference.

A Data-Driven Fault Diagnosis Method Using Modified Health Index and Deep Neural Networks of a Rolling Bearing

AbstractTo improve fault diagnosis accuracy, a data-driven fault diagnosis model based on the adjustment Mahalanobis–Taguchi system (AMTS) was proposed. This model can analyze and identify the characteristics of vibration signals by using degradation monitoring as the classifier to capture and recognize the faults of the product more accurately. To achieve this goal, we first used the modified ensemble empirical mode decomposition (MEEMD) scalar index to capture the bearing condition; then, by using the key intrinsic mode function (IMF) extracted by AMTS as the input of classifier, the optimized properties of bearing is decomposed and extracted effectively. Next, to improve the accuracy of the fault diagnosis, we tested different modes, employing the modified health index (MHI), which is designed to overcome the shortcomings of the proposed health index as a classifier in a single fault mode and the deep neural networks (DNNs) as a classifier in a multifault mode. To evaluate the effectiveness of our model, the Case Western Reserve University (CWRU) bearing data were used for verification. Results indicated a strong robustness with 99.16% and 1.09s, 99.86% and 6.61s fault diagnosis accuracy in different data modes. Furthermore, we argue that this data-driven fault diagnosis obviously lowers the maintenance cost of complex systems by significantly reducing the inspection frequency and improves future safety and reliability.

A Combined Method of Improved Grey BP Neural Network and MEEMD-ARIMA for Day-Ahead Wave Energy Forecast

Since wave fluctuates continuously, the forecast of the wave energy is very important for the operation of power systems integrated with large-scale wave energy generation. A combined model of day-ahead wave energy forecast based on improved grey BP neural network (BPNN) and modified ensemble empirical mode decomposition (MEEMD) -autoregressive integrated moving average (ARIMA) is proposed in this paper. Firstly, the wave is decomposed into wind waves and swells by wave theories. Secondly, the correlation between wind wave and wind speed is analyzed with improved grey BPNN, and the average height of wind waves can be forecasted based on the historical wind speed data. Thirdly, the MEEMD-ARIMA model is utilized to forecast the average wave height of swells. Thus, combining the wind wave and the swell, the average wave height of the integrated wave can be obtained. Finally, a conversion model from wave elements to wave energy for Archimedes wave swing (AWS) is introduced. A case study using the measured wind and wave data from a real ocean is illustrated, and the effectiveness of the proposed wave energy forecast model is validated.

Separation of long-period components of ground motion and its impact on seismic response of long-period diagrid structures

The improved modified ensemble empirical mode decomposition (MEEMD) method is used to decompose the long-period ground motion, and the introduction of permutation entropy is used to eliminate the abnormal signal, which can avoid the phenomenon of mode aliasing in the traditional empirical mode decomposition method, so that the components of the long-period ground motion can be accurately separated. Taking the separated and reconstructed ground motions as input action, the influence of long-period components on the seismic response of diagrid structure of high-rise building is studied. The results demonstrate that the MEEMD method can accurately extract long-period ground motion components from long-period ground motion. Long-period ground motion contains abundant short-term energy with low-frequency, which is the main reason for the large seismic response of long-period structure. The seismic response of diagrid structure is closely related to the natural period of the structure and the spectrum characteristics of the ground motion.

A Hydraulic Pump Fault Diagnosis Method Based on the Modified Ensemble Empirical Mode Decomposition and Wavelet Kernel Extreme Learning Machine Methods.

To address the problem that the faults in axial piston pumps are complex and difficult to effectively diagnose, an integrated hydraulic pump fault diagnosis method based on the modified ensemble empirical mode decomposition (MEEMD), autoregressive (AR) spectrum energy, and wavelet kernel extreme learning machine (WKELM) methods is presented in this paper. First, the non-linear and non-stationary hydraulic pump vibration signals are decomposed into several intrinsic mode function (IMF) components by the MEEMD method. Next, AR spectrum analysis is performed for each IMF component, in order to extract the AR spectrum energy of each component as fault characteristics. Then, a hydraulic pump fault diagnosis model based on WKELM is built, in order to extract the features and diagnose faults of hydraulic pump vibration signals, for which the recognition accuracy reached 100%. Finally, the fault diagnosis effect of the hydraulic pump fault diagnosis method proposed in this paper is compared with BP neural network, support vector machine (SVM), and extreme learning machine (ELM) methods. The hydraulic pump fault diagnosis method presented in this paper can diagnose faults of single slipper wear, single slipper loosing and center spring wear type with 100% accuracy, and the fault diagnosis time is only 0.002 s. The results demonstrate that the integrated hydraulic pump fault diagnosis method based on MEEMD, AR spectrum, and WKELM methods has higher fault recognition accuracy and faster speed than existing alternatives.

Fault diagnosis of mine asynchronous motor based on MEEMD energy entropy and ANN

The stator current signals of mining asynchronous motor are often non-stationary, making it challenging to extract fault features in the time domain. Therefore, this paper proposes a rotor fault diagnosis method based on the combination of Modified Ensemble Empirical Mode Decomposition (MEEMD) energy entropy and Artificial Neural Network (ANN). Firstly, the stator current signals are decomposed into a series of Intrinsic Mode Function (IMF) components by the MEEMD. Secondly, the IMF components with the most abundant information are selected by the cross-correlation criterion, and their energy entropy is calculated to construct feature vectors. Finally, the feature vectors are input into the ANN for training and state recognition. The faulty motor is modeled by ANSYS Maxwell software to obtain the simulated data. It is verified that the MEEMD-ANN method is feasible for fault diagnosis of mine motors, which can accurately identify the different status of motors, including normal state, broken rotor bars, and air gap eccentricity, the recognition rate can reach 99%. The MEEMD-ANN improves the accuracy by 2% compared with the EEMD-ANN, improves the accuracy by 3.75% compared with the MEEMD-SVM.

Remaining Useful Life Estimation Combining Two-Step Maximal Information Coefficient and Temporal Convolutional Network With Attention Mechanism

Remaining useful life (RUL) estimation has received extensive attention in many fields, which provides improved decision-making for condition-based maintenance (CBM) and enhances the stability of engineered systems. However, the effective construction of the feature set and design of the high-accuracy prediction model are the main challenges in RUL estimation. Consequently, a novel RUL estimation method is proposed in this paper. Initially, to extract the deep characteristics of the raw sensor data, the modified ensemble empirical mode decomposition (MEEMD), combined with the correlation coefficient threshold to select the sensitive intrinsic mode function (IMF) components, is proposed to reconstruct more representative features. Additionally, maximal information coefficient-based two-step feature selection (TSMIC) method is used to select the optimal feature subset. Ultimately, temporal convolutional network with attention mechanism (TCNA) is proposed to capture long-term time-series information and achieves more precise prediction results. The proposed framework is verified through a case study on aircraft turbofan engines, and comparisons with other state-of-the-art methods are presented. The experimental results of the case study show the effectiveness and superiority of the proposed approach.

Open-Circuit Fault Diagnosis of Three-Phase PWM Rectifier Using Beetle Antennae Search Algorithm Optimized Deep Belief Network

Effective open-circuit fault diagnosis for a two-level three-phase pulse-width modulating (PWM) rectifier can reduce the failure rate and prevent unscheduled shutdown. Nevertheless, traditional signal-based feature extraction methods show poor distinguishability for insufficient fault features. Shallow learning diagnosis models are prone to fall into local extremum, slow convergence speed, and overfitting. In this paper, a novel fault diagnosis strategy based on modified ensemble empirical mode decomposition (MEEMD) and the beetle antennae search (BAS) algorithm optimized deep belief network (DBN) is proposed to cope with these problems. Initially, MEEMD is applied to extract useful fault features from each intrinsic mode function (IMF) component. Meanwhile, to remove features with redundancy and interference, fault features are selected by calculating the importance of each feature based on the extremely randomized trees (ERT) algorithm, and the dimension of fault feature vectors is reduced by principal component analysis. Additionally, the DBN stacked with two layers of a restricted Boltzmann machine (RBM) is selected as the classifier, and the BAS algorithm is used as the optimizer to determine the optimal number of units in the hidden layers of the DBN. The proposed method combined with feature extraction, feature selection, optimization, and fault classification algorithms significantly improves the diagnosis accuracy.

Fault Diagnosis of Spindle Device in Hoist Using Variational Mode Decomposition and Statistical Features

By analyzing nonlinear and nonstationary vibration signals from the spindle device of the mine hoist, it is a challenge to overcome the difficulty of fault feature extraction and accurately identify the fault of rotor-bearing system. In response to this problem, this paper proposes a new approach based on variational mode decomposition (VMD), SVM, and statistical characteristics such as variance contribution rate (VCR), energy entropy (EE), and permutation entropy (PE). Comparisons have gone to evaluate the performance of rolling bearing defect by using EMD (Empirical Mode Decomposition), MEEMD (Modified Ensemble EMD), BP (Back Propagation) network, single or multiple statistical characteristics, and different motor loads. The experiment was carried out on the mechanical failure simulator of the main shaft device of the hoist, which verified the reliability and effectiveness of the method. The results show that the diagnosis method is suitable for feature extraction of bearing fault signals, with the highest diagnosis accuracy. It can provide a good practical reference for the fault diagnosis of mechanical equipment of the hoist spindle device and has certain practical value.

Detection of Synchrophasor False Data Injection Attack Using Feature Interactive Network

The synchrophasor data recorded by Phasor Measurement Units (PMUs) plays an increasingly critical role in the regulation and situational awareness of power systems. However, the widely installed PMUs are vulnerable to multiple malicious attacks from cyber hackers during data transmission and storage. To address this problem, a Modified Ensemble Empirical Mode Decomposition (MEEMD) is proposed first to extract the intrinsic mode functions of each Synchrophasor Data Attacks (SDA). The frequency-based adaptive screening criterion embedded in MEEMD is used to eliminate the false intrinsic mode functions. Next, a Multivariate Convolutional Neural Network (MCNN) is proposed to identify multiple SDA by utilizing the extracted intrinsic mode functions and original SDA as input vectors. A fusion block as the main structure of MCNN is also leveraged to increase the diversity of features and compress the model parameters. Integrating MEEMD and MCNN, a framework with automatic feature extraction and multi-source information fusion capability, referred to as Feature Interactive Network (FIN), is proposed to detect multiple SDA. Based on the proposed FIN framework, six types of SDA are explored for the first time using actual synchrophasor data in FNET/Grideye that was collected from different locations in the U.S. Eastern Interconnection. Finally, a large quantity of experiments with different attack strengths are used to evaluate the adaptability and classification performance of the proposed FIN.

Noise Source Identification Method for a Warp Machine Based on MEEMD_AIC

In order to recognise the noise source of a warp knitting machine, a method based on Modified Ensemble Empirical Mode Decomposition (MEEMD) and Akaike Information Criterion (AIC) is proposed. The MEEMD_AIC method is applied to measure the noise signal of a warp knitting machine and analyse every single effective component selected. Noise source identification is realised by combining the vibration signal characteristics of the main parts of the warp knitting machine. Firstly, MEEMD is used to decompose the measured noise signal of the warp knitting machine into a finite number of intrinsic mode function (IMF) components. Then, singular value decomposition (SVD) is performed on the covariance matrix of the component matrix to get the eigen value of the matrix. Next, the number of effective components is estimated based on the AIC criterion, and the effective components are selected by combining the energy characteristic index and the Pearson correlation coefficient method. The results show that the noise signal of the warp knitting machine is a mixture of multiple noise source signals. The main noise sources of the warp knitting machine, including the vibration of the pulling roller, the main shaft of the loop forming mechanism and the push rod of the guide bar traverse the mechanism, provide theoretical support for recognition of the active noise reduction of the warp knitting machine using the MEEMD_AIC method.

Nonlinear Flight Mode Identification by Applying Modified Ensemble Empirical Mode Decomposition and Hilbert–Huang Transformation

Identification of the “aircraft dynamic modes” is one of the researchers’ favorite subjects in the flight mechanics field to study the behavior of airplane in maneuvers and nonlinear regime. Technically, by measuring the flight test parameters, the flight dynamic modes could be estimated. Although several approaches calculate precisely the linear flight dynamic modes, there is not a known standard procedure to calculate the nonlinear flight dynamic modes. In this paper, identification of the flight dynamic modes in highly nonlinear maneuver conditions and particularly in the coupled dynamic spin maneuvers is investigated. A new approach based on Hilbert–Huang Transform (HTT) for identification of the flight dynamic modes is presented in this paper. To overcome the mode mixing problem during decomposition which leads to imprecise results in the estimation of the flight dynamic modes, a Modified Ensemble Empirical Mode Decomposition (MEEMD) algorithm for processing of the complex signals that originate from FDR is presented in this research. This new algorithm is able to make a precise reconstruction of the original signal and it performs the task of signal decomposition with fewer iterations and so with less complexity order rather than the competitor approaches. An innovative algorithm for calculating flight dynamic modes natural frequencies and the damping ratios which uses the IMFs produced by MEEMD is presented. The obtained approach was applied to flight data, which belong to spin maneuvers. The results show the correct performance of the presented technique in the nonlinear flight regime.

Carbon price forecasting based on modified ensemble empirical mode decomposition and long short-term memory optimized by improved whale optimization algorithm.

The accurate prediction of carbon prices poses a tremendous challenge to relevant industry practitioners and governments. This paper proposes a novel hybrid model incorporating modified ensemble empirical mode decomposition (MEEMD) and long short-term memory (LSTM) optimized by the improved whale optimization algorithm (IWOA). This model is based on the nonlinear and non-stationary characteristics of carbon price. The original carbon price is first decomposed into nine intrinsic mode functions (IMFs) and a residual using the MEEMD model. Then, the random forest method is applied to determine the input variables of each IMF and the residual, in the LSTM neural network. The carbon price is then predicted by the LSTM model optimized by the IWOA. The proposed hybrid model is applied to predict the carbon prices of Beijing, Fujian, and Shanghai to assess its effectiveness. The results reveal that the model achieved higher prediction performance than 11 other benchmark models. Our observations indicate that decomposition of carbon price can effectively improve the accuracy of prediction. Moreover, the improved LSTM model is more suitable for time series prediction. The proposed model provides a novel and effective carbon price forecasting tool for governments and enterprises.

Load State Identification Method for Wet Ball Mills Based on the MEEMD Singular Value Entropy and PNN Classification

To overcome the difficulty of accurately judging the load state of a wet ball mill during the grinding process, a method of mill load identification based on the singular value entropy of the modified ensemble empirical mode decomposition (MEEMD) and a probabilistic neural network (PNN) classifier is proposed. First, the MEEMD algorithm is used to decompose the vibration signals recorded under different load states to obtain the intrinsic mode components, and a correlation coefficient threshold is used to select the sensitive mode components that characterize the state of the original signal. Second, singular value decomposition is used to obtain the singular value entropy. Finally, the load state of the wet ball mill is judged based on the magnitude of the singular value entropy. A characteristic mill load vector is constructed from the singular value entropies of the cylinder vibration signals recorded under different load conditions and is used as the input to a PNN, which then outputs the predicted ball mill load state; in this way, a load state identification model is established. Grinding experiments are presented to verify the effectiveness of the proposed method, showing that the method can accurately identify the load state of a wet ball mill.