Landslides are serious and complex geological and natural disasters that threaten the safety of people’s health and wealth worldwide. To face this challenge, a landslide displacement prediction model based on time series analysis and modified long short-term memory (LSTM) model is proposed in this paper. Considering that data from different time periods have different time values, the weighted moving average (WMA) method is adopted to decompose the cumulative landslide displacement into the displacement trend and periodic displacement. To predict the displacement trend, we combined the displacement trend of landslides in the early stage with an LSTM model. Considering the repeatability and periodicity of rainfall and reservoir water level in every cycle, a long short-term memory fully connected (LSTM-FC) model was constructed by adding a fully connected layer to the traditional LSTM model to predict periodic displacement. The two predicted displacements were added to obtain the final landslide predicted displacement. In this paper, under the same conditions, we used a polynomial function algorithm to compare and predict the displacement trend with the LSTM model and used the LSTM-FC model to compare and predict the displacement trend with eight other commonly used algorithms. Two prediction results indicate that the modified prediction model is able to effectively predict landslide displacement.

/pdf/landslide-displacement-prediction-model-using-time-series-25hu1n3w.pdf

Landslide Displacement Prediction Model Using Time Series Analysis Method and Modified LSTM Model

In landslide displacement prediction, random factors that would affect the performance of prediction are usually ignored by using a time series analysis method. In order to solve this problem, in this paper, a landslide displacement prediction model, the local mean decomposition-bidirectional long short-term memory (LMD-BiLSTM), is proposed based on the time-frequency analysis method. The model uses the local mean decomposition (LMD) algorithm to decompose landslide displacement and obtains several subsequences of landslide displacement with different frequencies. This paper analyzes the internal relationship between the landslide displacement and rainfall, reservoir water level, and landslide state. The maximum information coefficient (MIC) algorithm is used to calculate the intrinsic correlation between each subsequence of landslide displacement and rainfall, reservoir water level, and landslide state. Subsequences of influential factors with high correlation are selected as input variables of the bidirectional long short-term memory (BiLSTM) model to predict each subsequence. Finally, the predicted results of each of the subsequences are added to obtain the final predicted displacement. The proposed LMD-BiLSTM model effectiveness is verified based on the Baishuihe landslide. The prediction results and evaluation indexes show that the model can accurately predict landslide displacement.

/pdf/landslide-displacement-prediction-based-on-time-frequency-1nap9v3i.pdf

Landslide Displacement Prediction Based on Time-Frequency Analysis and LMD-BiLSTM Model

Landslide disasters cause serious property losses and casualties every year. Landslide displacement prediction is fundamental for mitigating landslide disasters. Several approaches have been used to predict landslide displacement, yet a more accurate and reliable displacement prediction still has a poor understanding of landslide early warning systems for landslide mitigation, due to limited data and mutational displacements. To boost the robustness and accuracy of landslide displacement prediction, this paper assembled a new hybrid model containing the local mean decomposition (LMD), innovations state space models for exponential smoothing (ETS), and the temporal convolutional network (TCN). The proposed model, which is based on over 10 years of long-term time series monitoring GPS data, was tested on the selected case—stepwise Baijiabao landslide in the Three Gorges Reservoir area of China (TGRA) was tested by the proposed model. The results presented that the LMD–ETS–TCN model has the best performance in comparison with other benchmark models. Compared with autoregressive integrated moving average (ARIMA), support vector regression (SVR), and long short-term memory neural network (LSTM), the accuracy was noticeably improved by an average of 40.9%, 46.2%, and 22.1%, respectively. The robustness and effectiveness of the presented approach are attested, and it has discernible improvements for landslide displacement prediction.

/pdf/a-novel-hybrid-lmd-ets-tcn-approach-for-predicting-landslide-3dkaltuc.pdf

A Novel Hybrid LMD-ETS-TCN Approach for Predicting Landslide Displacement Based on GPS Time Series Analysis

The high-pressure heater system is an important part of the return heat system of thermal power units, which can significantly reduce the boiler fuel consumption and is of great significance to the safe and economic operation of the units. Taking into consideration the issues that the high-pressure heater system data has strong non-linear characteristics and the fault diagnosis accuracy is low, this paper proposes a hybrid model-based fault monitoring and diagnosis method for a high-pressure heater system. Firstly, an improved particle swarm optimization algorithm (IEDPSO) is proposed. A differential evolution operation is introduced to enhance particle diversity, and the inertia weight coefficients and learning factor parameters are improved to optimize the particle position and velocity update process. The problem that PSO tends to fall into local optimum at the late stage of iterative optimization search is solved. Numerical simulation experiments demonstrate that IDEPSO has high convergence speed and accuracy in the optimization process of the test function. Secondly, a fault monitoring and diagnosis method based on a hybrid kernel principal component analysis (KPCA)-IDEPSO-probabilistic neural network (PNN) model is proposed. Non-linear features are extracted using KPCA for fault monitoring. The IDEPSO algorithm is used to iteratively find the best PNN to improve the fault diagnosis accuracy. Simulation experiments prove that compared with the traditional PNN model, the fault diagnosis accuracy of the KPCA-IDEPSO-PNN model is improved by 4.9% and the number of fault misclassifications is reduced by 34, effectively improving the fault diagnosis accuracy of the high-pressure heater system and ensuring the safe and stable operation of thermal power units.

Fault monitoring and diagnosis of high-pressure heater system based on improved particle swarm optimization and probabilistic neural network

In recent years, low-cost single-frequency GNSS receivers have been widely used in many fields such as mass navigation and deformation monitoring; however, due to the poor signal quality of low-cost patch antennae, it is difficult for carrier phase real-time kinematic (RTK) technology to fix the integer ambiguity. Differential GNSS (DGNSS) positioning with pseudorange can effectively meet the high robustness and reliability requirements for the submeter to the meter level positioning accuracy of UVA/vehicle/aerospace users. To improve the DGNSS positioning accuracy and reliability of low-cost single-frequency GNSS receivers in complex environments, we propose a differential barometric altimetry (DBA)-assisted DGNSS positioning algorithm, which solves the DGNSS observation equations jointly and rigorously with the Earth ellipsoidal constraint equations constructed by the DBA altitude. The DBA altitude accuracy at different baseline lengths was evaluated in detail, and the DGNSS positioning performance of the single-frequency low-cost u-blox receiver NEO-M8T with a patch antenna and DGNSS/DBA combined positioning performance with the BMP280 barometer was analyzed by several sets of static and dynamic experiments under different environments. The results show that the single-frequency NEO-M8T receiver with patch antenna DGNSS positioning accuracy is submeter level in the static environment and drops to meter level in the dynamic environment. GPS+BDS dual system has higher positioning accuracy than single GPS or single BDS. DGNSS/DBA combination has higher positioning accuracy than DGNSS, especially the root mean square error (RMSE) can be improved by 30% to 80% in the U direction and slightly improved in the N and E directions. This study can provide an effective solution reference for various applications of low-cost sensor fusion positioning in the mass consumer market.

/pdf/low-cost-single-frequency-dgnss-dba-combined-positioning-29mqlpxr.pdf

Low-Cost Single-Frequency DGNSS/DBA Combined Positioning Research and Performance Evaluation

In recent years, machine learning models facilitated notable performance improvement in landslide displacement prediction. However, most existing prediction models which ignore landslide data at each time can provide a different value and meaning. To analyze and predict landslide displacement better, we propose a dynamic landslide displacement prediction model based on time series analysis and a double-bidirectional long short term memory (Double-BiLSTM) model. First, the cumulative landslide displacement is decomposed into trend and periodic displacement components according to time series analysis via the exponentially weighted moving average (EWMA) method. We consider that trend displacement is mainly influenced by landslide factors, and we apply a BiLSTM model to predict landslide trend displacement. This paper analyzes the internal relationship between rainfall, reservoir level and landslide periodic displacement. We adopt the maximum information coefficient (MIC) method to calculate the correlation between influencing factors and periodic displacement. We employ the BiLSTM model for periodic displacement prediction. Finally, the model is validated against data pertaining to the Baishuihe landslide in the Three Gorges, China. The experimental results and evaluation indicators demonstrate that this method achieves a better prediction performance than the classical prediction methods, and landslide displacement can be effectively predicted.

/pdf/landslide-displacement-prediction-based-on-time-series-3ok1ir12.pdf

Landslide Displacement Prediction Based on Time Series Analysis and Double-BiLSTM Model

Due to the characteristics of measurement datum distribution, the altitude positioning error of GPS (Global Positioning System) is relatively large. This study discusses a differential barometric altimetry (DBA) system from systematic implementation level and provides a key technical solution for the continuous and precise altitude measurement from the DBA systematic implementation aspect, which can provide a good supplement and enhancement for GPS. The structure of signal message is designed; the power spectrum of spread spectrum ranging signal is given, and the fusion of navigation and communication is realised. The two dimensional curved surface fitting function of pressure and temperature (P, T) is constructed, and the fitting parameters are calculated by the least-square solution, which can realise calibration compensation and measurement correction of sensor. The influence of the system noise variance matrix Q k and the measurement noise variance matrix R k on the filtering effect in Kalman filter is analysed. An algorithm of multi-base station (BS) correction is proposed, which can realise the height information correction calculation in measuring station (MS) via seamless handover between multi-BSs. The experimental results show that the key technology solutions in DBA system can achieve a well systematic performance..

https://ietresearch.onlinelibrary.wiley.com/doi/pdf/10.1049/iet-smt.2018.5316

Applications of differential barometric altimeter in ground cellular communication positioning network

The optimal search algorithm based on double-difference carrier phase is one of the methods to solve the attitude angle. An adaptive weighted particle swarm optimization (AWPSO) algorithm based on the Chi-square test is proposed to solve the attitude angle of ultra-short baseline. We establish the fitness function by introducing the relationship between attitude angle and baseline vector into the observation equations of double-difference carrier phase. Searching the attitude according to the fitness model instead of calculating the angle directly, which avoids solving the integer ambiguity. Using adaptive inertial weights and synchronous adaptive learning factors to speed up the convergence of attitude search. Constructing the candidate solution sequence to save the search result and calculating the Mahalanobis distance of the candidate solution. The local optimal solution is eliminated by the Chi-square test, and the attitude angle is gained by weighting the filtered sequence. Various static experimental results show that the new algorithm performs better than the direct solution method, the least square estimation method, and the PSO-based attitude solution method. The root mean square error of yaw angle and pitch angle is 0.245° and 0.236° directly at 0.575 m baseline.

Yuanfa Ji

Papers

Landslide Displacement Prediction Based on Time Series Analysis and Double-BiLSTM Model

Landslide Displacement Prediction Model Using Time Series Analysis Method and Modified LSTM Model

Landslide Displacement Prediction Based on Time-Frequency Analysis and LMD-BiLSTM Model

Applications of differential barometric altimeter in ground cellular communication positioning network

Adaptive weighted particle swarm optimization (AWPSO) attitude determination algorithm based on Chi-square test