Physics-informed deep learning: A promising technique for system reliability assessment

On the basis of a brief introduction about the recent progress of reliability assessment of complex systems, combined methods and state methods are introduced. In view of the shortcomings and deficiencies of these two kinds of methods, a reliability assessment method is put forward which based on fault tree analysis (FTA) and Bayesian network (BN). By applying the results of FTA of a complex system and adopting the transforming method of FTA to BN, BN model for reliability assessment of the complex system is established. A missile engine is taken for example and the effectiveness and the priority of the model in reliability assessment of system is verified.

Assessing the structural health of operational wind turbines is crucial, given their exposure to harsh environments and the resultant impact on longevity and performance. However, this is hindered by the lack of data in commercial machines and accurate models based on manufacturers’ proprietary design data. To overcome these challenges, this study focuses on using Gaussian Process Regression (GPR) to evaluate the loads in wind turbines using a hybrid approach. The methodology involves constructing a hybrid database of aero-servo-elastic simulations, integrating publicly available wind turbine models, tools and Supervisory Control and Data Acquisition (SCADA) measurement data. Then, constructing GPR models with hybrid data, the prediction is validated against the hybrid and SCADA measurements. The results, derived from a year of SCADA data, demonstrate the GPR model’s effectiveness in interpreting and predicting turbine performance metrics. The findings of this study underscore the potential of GPR for the health and reliability assessment and management of wind turbine systems.

This paper deals with the reliability assessment of electrical power supply systems under redundancy. For doing so, three probabilistic safety analysis (PSA) methods are proposed and developed namely fault tree analysis (FTA), reliability block diagram (RBD), and Monte Carlo Simulation (MC) method. The FTA method is performed using RiskSpectrumPSA® software and the RBD and MC methods are coded using Matlab software. The electrical power supply system reliability is analyzed and enhanced considering various cases of redundancy. The FTA method is based on qualitative and quantitative reliability assessment. The quantitative assessment calculates the failure probability of the top event, and the qualitative approach determines the minimal cut sets (MCS). However, the RBD and MC methods are proposed to compute the system failure probability. The considered methods are applied on three case studies and the obtained results are compared to each other. The obtained results show that the three methods provide almost the same results. Furthermore, the reliability of the electrical power supply is clearly enhanced with the number of redundant components.

Bayesian system reliability assessment method with maximum entropy as prior distribution is developed in the present analysis. The system reliability assessment model is established, and the method and process of system reliability assessment are put forward. The method is based on the theory of information entropy. The probability density function and distribution function of system reliability parameters are given by using maximum entropy method. Prior distribution of Bayesian analysis method is the distribution function. Likelihood function is given by a small amount of field test data. According to Bayesian methods, posterior distribution is determined by using prior distribution and likelihood function. The point estimation and interval estimation of system reliability parameters are calculated in order to evaluate the system reliability. The method solves better the problem of system reliability assessment for small sample cases, and is an optimum method for reliability assessment of large and complex systems.

Efficient reliability assessment of structural dynamic systems with unequal weighted quasi-Monte Carlo Simulation

Since a typical distribution system accounts for 40% of the cost to deliver power and 80% of customer reliability problems, distribution system design and operation is critical for financial success and customer satisfaction. To address this situation, Commonwealth Edison is integrating reliability assessment techniques into distribution planning and engineering. Just as equipment loading and voltage regulation are treated with analytical rigor with power flow models, interruptions and outages can be treated with analytical rigor with reliability models. This is made possible through the use of predictive reliability assessment tools that are able to predict customer reliability characteristics based on system topology and component reliability data. Reliability software predicts expected interruption frequencies and duration in a manner analogous to current and voltage evaluations in traditional power flow models. This article describes the distribution system's reliability assessment model jointly developed by Commonwealth Edison and ABB. This effort modeled, calibrated, and assessed the reliability of more than 3300 feeders. After completing the model, an intelligent system was used to automatically identify potential reliability problems and recommend reliability improvement projects based on expected benefits and costs.

Research on adequacy assessments of systems with integrated renewable generation sources has been carried out for decades, with increasing attention recently on stand-alone systems. In this paper, a reliability assessment of a stand-alone system that runs on wind turbine generators (WTGs), photovoltaic (PV) modules, batteries for storage, and backup micro gas turbines (MGTs) with different configurations is carried out using the Sequential Monte Carlo Simulation (SMCS) method. Some atmospheric data on wind speed, solar irradiation, and ambient air temperature are needed to obtain the power outputs of the corresponding sources. Historical data were fitted into a Weibull distribution, and the extracted scale and shape parameters were used to re-simulate the important factors of renewable generation power output models. The two-state reliability model was employed to obtain the adequacy model of the MGTs. The two-state reliability model was combined with the output power models to build the generation model of WTGs, PVs and batteries. The generation model was then tested on the IEEE Roy Billinton Test System (RBTS) as a load model.

The concept of Big Data of Software Reliability (BDSWR) is discussed. According with publication overview it's concluded that BDSWR are not systematically analyzed and used for the reliability assessment of the software systems (SWS). It restricts the potential opportunities of the increase of the SWS reliability assessment accuracy. We offer to use big experimental data about the actual reliability of different SWS which are similar to assessed system. The approach is implemented by way of: 1) the search of BDSWR of the earlier developed similar systems; 2) data collection about of reliability indices of similar systems; 3) use of similar systems indices for reliability assessment and prediction of the evaluated system. Within the approach we offer the technique of big data and similarity-based reliability assessment. The offered technique is based on four principles of the SWS similarity and includes nine stages. The input data of the technique are artefacts of the evaluated system. The output data of the technique are reliability assessments of the evaluated system. The software tools of implementation of the technique are described such as: the agent for search and assessment of data size; the agent of similarity assessment; the agent of reliability assessment; the agent of the verification efficiency improving.

Objective. Monte Carlo (MC) simulations are the benchmark for accurate radiotherapy dose calculations, notably in patient-specific high dose rate brachytherapy (HDR BT), in cases where considering tissue heterogeneities is critical. However, the lengthy computational time limits the practical application of MC simulations. Prior research used deep learning (DL) for dose prediction as an alternative to MC simulations. While accurate dose predictions akin to MC were attained, graphics processing unit limitations constrained these predictions to large voxels of 3 mm × 3 mm × 3 mm. This study aimed to enable dose predictions as accurate as MC simulations in 1 mm × 1 mm × 1 mm voxels within a clinically acceptable timeframe. Approach. Computed tomography scans of 98 breast cancer patients treated with Iridium-192-based HDR BT were used: 70 for training, 14 for validation, and 14 for testing. A new cropping strategy based on the distance to the seed was devised to reduce the volume size, enabling efficient training of 3D DL models using 1 mm × 1 mm × 1 mm dose grids. Additionally, novel DL architecture with layer-level fusion were proposed to predict MC simulated dose to medium-in-medium (D m,m ). These architectures fuse information from TG-43 dose to water-in-water (D w,w ) with patient tissue composition at the layer-level. Different inputs describing patient body composition were investigated. Main results. The proposed approach demonstrated state-of-the-art performance, on par with the MC D m,m maps, but 300 times faster. The mean absolute percent error for dosimetric indices between the MC and DL-predicted complete treatment plans was 0.17% ± 0.15% for the planning target volume V 100, 0.30% ± 0.32% for the skin D 2cc , 0.82% ± 0.79% for the lung D 2cc , 0.34% ± 0.29% for the chest wall D 2cc and 1.08% ± 0.98% for the heart D 2cc . Significance. Unlike the time-consuming MC simulations, the proposed novel strategy efficiently converts TG-43 D w,w maps into precise D m,m maps at high resolution, enabling clinical integration.

Reliability assessment of generic geared wind turbines by GTST-MLD model and Monte Carlo simulation

Traffic state estimation (TSE) refers to the inference of traffic state information from partially observed traffic data and some prior knowledge of the traffic dynamics. TSE plays a key role in traffic management as traffic control relies on an accurate estimation of the traffic states. Macroscopic traffic models describe traffic dynamics with aggregated values such as traffic density, velocity, and flow and are often employed for TSE of a freeway road segment. This paper integrates Partial Differential Equation (PDE) observer design and deep learning paradigm to estimate spatial-temporal traffic states from boundary sensing. With the PDE observer providing an rigorous guarantee for state estimates, we propose Observer-Informed Deep Learning (OIDL) paradigm which is a data-driven solution to TSE that leverages the PDE observer design. An Observer-Informed Neural Network (OINN) is constructed by training NN to generate state estimates and use the boundary observer for regularization. The OINN forms a novel class of data-efficient function approximators that encode PDE observer as theoretical guarantee and improves the accuracy and convergence speed. Experiments using NGSIM data-set demonstrate that the proposed OIDL reduces the estimation error compared to either the model-based observer or either the data-driven neural network. We also compare OIDL with the existing Physics-informed Deep Learning (PIDL) approach for TSE.

With the increasing dependencies on Information and Communication Technologies (ICTs), the smart distribution system has similar characteristics to Cyber-Physical System (CPS) and is seen as the so-called “cyber-physical power system” (CPPS). However, the impact of upstream substation is becoming one of the main factors impacting on distribution reliability. To avoid the too optimistic conclusion, the Cyber-Induced Dependent Failures (CDFs) of upstream substation are important to be considered. Thus the cyber security risk brought by both the malfunction of ICT components and the cyber-attack should be taken into account. This paper provided a methodology to incorporate both the two sides of CDFs in upstream substation into cyber-enabled distribution system reliability assessment. In the sequential Monte Carlo simulation (MCS), the game process between attackers and defenders can be modelled as the semi-Markov Chain (SMP), and indices are calculated by the minimal path method. The effects of CDFs are quantified and converted onto the joint feeders, based on the Breadth-First-Search (BFS), as well. Some conclusions for the distribution and substation automation planning have been made and verified by the cyber-extended Roy Billinton Test System (RBTS).

Considering characteristics of nuclear power plant digital instrumentation and control system, challenges of reliability assessment of digital instrumentation and control system are pointed out. As for its dynamic behavior and multi-states failure mode, limitation of traditional methods, research status and applicability of dynamic flowgraph methodology and other advance dynamic reliability assessment approach were introduced and compered. It's concluded that Dynamic flowgraph methodology maybe the dominant one in future after simplify its model procedure and reduction computational cost of its deductive analysis.

Wind power has the characteristics of randomness and intermittent, with the increasing unit capacity of wind turbines and wind farm scale, it is urgent to study the impact on generation and transmission system reliability. In addition, there are different degrees of correlation among regional multiple wind farms because of their geographic proximity. When it comes to research on the influence on the power system reliability, the analysis result will be larger error without considering the correlation, which is optimistic for reliability assessment. Therefore, it is meaningful to take the correlation into account in wind power system reliability assessment. In this paper, IEEE-RTS 79 generation and transmission reliability test system is adopted to verify the reliability assessment of power systems with wind farms considering correlations. The experimental results show that adding wind turbines in the original power system, system reliability has been greatly improved; and the system has higher reliability with given the considering the correlation between wind farms.

Accurate simulation of pipe transient wave is crucial for the design, reliability assessment, and defect detection of water supply systems. However, this is a difficult task since the wave propagation media in pipeline systems are often highly complex and involve a range of uncertainties. Experimental data show that key parameters such as valve closure time, wave speed, and pipe material coefficients are strongly uncertain and stochastic that may vary in a wide range. To analyze the transient behavior in such complex environments, surrogate models are built to realize the fast computation of uncertainty propagation. Three surrogate modeling methods, i.e., the polynomial‐chaos‐expansion, ordinary‐Kriging, and polynomial‐chaos‐Kriging, are tested and compared in terms of transient wave forecast accuracy. The global sensitivity is then analyzed via the Sobol index method, by which the most influential uncertainty sources on various transient signal features are identified.