•The power system state estimation model-based Gaussian process is first proposed to simulate non-Gaussian noise and improve estimation accuracy.•The power system state forecast method-based ...long-short term memory network is proposed to realize the power system forecasting-aided state estimation (FASE).•The abnormal measurement detection method-based Gaussian mixture model is first proposed to realize the robust FASE.
This paper presents a real-time robust power system forecasting-aided state estimation method based on the Bayesian framework, deep learning, and Gaussian mixture model to dynamically estimate the non-Gaussian measurement noise in the real-time power system. The approach is data driven and model independent. A non-linear mapping function between measurement and state is formulated based on the historical operating data of the power system and the Gaussian process. Then combine the anomaly detection technology in machine learning and the Gaussian mixture model to accurately judge and delete the abnormal data in measurement information. Thus, a power system state forecasting model based on long-short term memory neural network is established, which can solve the problem of missing data combining power flow calculation. Numerical simulations carried out on the IEEE 118-bus and IEEE 300-bus test system reveal that the proposed method has high accuracy and robustness.
In engineering optimization, surrogate model (SM) is widely used to replace the involved time expensive model, due to the expensive model is complex and high precise requirement caused a long ...calculation cycle. In traditional process of engineering optimization, the separation of the surrogate model static construction stage and dynamic optimization stage depresses the optimization accuracy and efficiency. Moreover, in order to ensure the accuracy of the surrogate model, expensive model had to be intensively invoked to get enough representative samples in the design space for the SM training. In this paper, a surrogate model adjoint refine based global optimization method combining with the multi-stage fuzzy clustering space reduction strategy (MFCPR-SGO) is proposed to improve the optimization accuracy and efficiency. Firstly, the optimal Latin hypercube design method (OLHD) is used to sample in design space to assure the initial sample set with strong space filling property. Then, the design space is subdivided into three tiered subspaces by using the space reduction strategy of multi-stage fuzzy clustering, which has the ability of space focusing, space reduction and jumping out of local optimum. On this basis, the hierarchical optimization method with ADAM gradient descent is proposed to quickly and accurately search the local minimum value of the objective function in each subspaces. At the same time, combined with the extremum sampling and the gaussian process sampling, a dynamic sampling algorithm is given to realize the synchronization of optimization and surrogate model update. Finally, the benchmark test problems in 12 different dimensions are used to verify the proposed method. The results show that the optimization accuracy can be improved by 21.3% and expensive model invoking times are reduced by 31.5% compared with other three heuristic optimization methods and the three recent surrogate-based optimization (SGO) algorithms. It indicated that the optimization precision and efficiency can be greatly improved by synchronizing the dynamic updating of the surrogate model with the engineering optimization search.
•Proposes a surrogate-adjoint refine based global optimization method to synchronize the optimization and surrogate model update.•Subdivides the design space into three tiered subspaces invoking the space reduction strategy of multi-stage fuzzy clustering.•Invokes the hierarchical optimization method with ADAM gradient descent to search the local minimum in each subspace.•Provides a combined sampling strategy with extremum sampling and gaussian process to update the surrogate model.•In solving 12 benchmark problems, optimization accuracy is improved by 21.3% and expensive model invoking times are reduced by 31.5%.
A Gauss process state-space model trained in a laboratory cannot accurately simulate a nonlinear system in a non-laboratory environment. To solve this problem, a novel Gauss process state-space model ...optimization algorithm is proposed by combining the expectation–maximization algorithm with the Gauss process Rauch–Tung–Striebel smoother algorithm, that is, the EM-GP-RTSS algorithm. First, a theoretical formulation of the Gauss process state-space model is proposed, which is not found in previous references. Second, a Gauss process state-space model optimization framework with the expectation–maximization algorithm is proposed. In the expectation–maximization algorithm, the unknown system state is considered as the lost data, and the maximization of measurement likelihood function is transformed into that of a conditional expectation function. Then, the Gauss process–assumed density filter algorithm and the Gauss process Rauch–Tung–Striebel smoother algorithm are proposed with the Gauss process state-space model defined in this article, in order to calculate the smoothed distribution in the conditional expectation function. Finally, the Monte Carlo numerical integral method is used to obtain the approximate expression of the conditional expectation function. The simulation results demonstrate that the Gauss process state-space model optimized by the EM-GP-RTSS can simulate the system in the non-laboratory environment better than the Gauss process state-space model trained in the laboratory, and can reach or exceed the estimation accuracy of the traditional state-space model.
Three-dimensional graph of standard deviation, leakage rate and pressure.
Display omitted
•The experimental setup has been improved.•The experiments of the valve leakage detection are performed.•A ...multivariate mathematical model is established.•The probability distribution of leakage rate is obtained.•The modeling results of various methods are compared.
As a non-destructive testing method, acoustic emission (AE) technology can be used for internal leakage detection of valve on-line. In order to enhance the practicability of valve leakage detection by AE technology and realize valve leakage detection under various pressure conditions, the Gaussian process regression (GPR) is used to establish a multivariate mathematical model for describing the relationship between the characteristic of AE signal and the pressure, leakage rate. Based on the mathematical model built by GPR, we can not only determine leakage rate, but also obtain the probability distribution of leakage rate according to the predicted mean and standard deviation. In order to verify the modeling effect of GPR, the analysis results are compared with those of least squares linear regression, polynomial regression and support vector machine regression. For the HTS50-17 valve, the root mean square error of GPR is the smallest, which is 22.6948 ml/min. Moreover, the fitting degree of the GPR is greatest, which is 0.9653. On this basis, in order to improve the universality of valve leakage AE detection, the mixed experimental data of different flow coefficient of the same type valve are analyzed, and the root mean square error of GPR is the smallest, which is 23.0646 ml/min. It is indicated that GPR can achieve better results for the AE signal of valve leakage.
In this paper, supervised machine learning is applied to the parameter estimation for optimal asteroid transfer trajectories. Efficient models for the estimation of important trajectory parameters ...are developed based on the Gaussian Process Regression (GPR) technique. The essence of constructing the GPR-based model is to learn the correlation between the trajectory parameters and the selected features. The asteroid orbital elements are considered as an original feature set due to their decisive influence on transfer trajectories. Two strategies are introduced to enhance the prediction performance of GPR-based models. The first one, the grouping strategy, is able to improve the prediction accuracy by dividing the candidate asteroids into several groups. The second one is that two new compound features are constructed based on the idea of feature extraction, whose function is to provide more crucial information for the inference of transfer time. The efficiency of the proposed models is substantiated by evaluating the global optimal two-impulse transfers to inner main-belt asteroids. This paper provides a basic framework for evaluating the interplanetary trajectories by using supervised machine learning. The proposed approach can be easily extended to solve other trajectory optimization and analysis problems.
The presence of unburned carbon in coal-burning systems undoubtedly causes a loss in the amount of energy that can be obtained from the system, and also reveals an inadequacy in terms of the ...usability of the ashes. The expensiveness of existing unburned carbon prediction methods is one of the reasons why these technologies cannot be used. This situation requires working on alternative non-combustible carbon technologies. In this paper, a new approach is presented for estimating unburned carbon in a small-scale coal burner system using the Gaussian regression model and CCD camera-acquired flame image. The proposed approach evaluates brightness, fluctuation amplitude, area, and radiation signal properties of the flame image. The proposed non-combustible carbon estimation technique does not require prior knowledge of CCD camera features. In the feature acquisition phase, results were obtained for each natural component of the flame image in RGB colour space separately, in pairs, all together and for three artificial colour channels (grey image). With the proposed method, the unburned carbon estimation was obtained with an accuracy of R = 0.9664 when all colour channels of the RGB image were used together. This result shows that unburned carbon can be estimated from the instantaneous flame images obtained with the CCD camera.
With the variance of preload and vibration in working conditions, wind turbine bolt loosening is difficult to predict accurately. To address the problem, wind turbine bolts are employed as the study ...object, and the loosening mechanism of bolts as well as the prediction of preload variation are investigated by means of finite element analysis. The result shows that, under the action of transverse vibration load, the magnitude of vibration load is the main factor affecting the loosening, and the larger the load magnitude, the more likely the loosening occurs. Besides, a bolt loosening prediction model based on Gaussian process regression is developed to obtain confidence intervals for the variation of the preload in a probabilistic sense. This study provides a theoretical basis for solving the problem of bolt loosening and preload relaxation in wind power under vibration conditions, and improves the safety and reliability of wind turbine operation.
Owing to the problem that it is difficult to measure State of Health (SOH) of Li-ion battery online, a method to estimate SOH of Li-ion battery using the charge performance under different SOH is put ...forward. By Grey Incidence Analysis (GIA), the correlation degree between the charge performance and SOH is measured. Based on Multi-island Genetic Algorithm (MIGA) and Gauss Process Regression (GPR), with charge performance as input, SOH is estimates and variance of estimation is also calculated. The result shows that the above-mentioned method can evaluate SOH well.
Effective indoor localization largely relies on the fingerprint database (model) of Received Signal Strength (RSS) in connection with Radio Frequency sources, such as the most widely used Bluetooth ...Low Energy (BLE) iBeacons. RSSs exhibit significant random variations in both the spatial and temporal domains. It is a notoriously onerous and challenging task to construct the fingerprint database for accurate localization, as the BLE RSSs must be captured via a full space scan from one point to another every few meters in a certain period of time. In order to tackle this problem, this study proposes an approach to fast fingerprints construction that only requires a sparse sampling of RSS of the space. First, a smartphone records the time series of RSS over a designated path, and a radio map for the path is then generated by a spatio-temporal mapping method using the Pedestrian Dead Reckoning algorithm. Second, the radio map of the entire space can be obtained via Gauss Process Regression (GPR), with outliers reduced to improve the reliability of the fingerprint database. Experiments have been performed in an underground carpark (38 m
×
14 m), and the experimental results indicate that the proposed approach can construct the fingerprint database 300% faster than the conventional approach does. The localization accuracy of both approaches is quite similar (80% error in 2.8 m). The proposed approach offers potential for the construction of a large-scale fingerprint database for a wide-area Location Based Service (LBS) of Smart City indoor and outdoor integration, where big RSS data processing is a must.