Due to the insufficient feature learning ability and the bloated network structure, the gear fault diagnosis methods based on traditional deep neural networks always suffer from poor diagnosis accuracy and low diagnosis efficiency. Therefore, a small channel convolutional neural network under the multiscale fusion attention mechanism (MSFAM-SCCNN) is proposed in this paper. First, a small channel convolutional neural network (SCCNN) model is constructed based on the framework of the traditional AlexNet model in order to lightweight the network structure and improve the learning efficiency. Then, a novel multiscale fusion attention mechanism (MSFAM) is embedded into the SCCNN model, which utilizes multiscale striped convolutional windows to extract key features from three dimensions, including temporal, spatial, and channel-wise, resulting in more precise feature mining. Finally, the performance of the MSFAM- SCCNN model is verified using the vibration data of tooth-broken gears obtained by a self-designed experimental bench of an ammunition supply and delivery system.

This paper proposes an improved FMEA method that integrates component importance and Fuzzy-TOPSIS. The aim is to overcome the limitations of traditional FMEA methods, such as the significant subjective influence in determining risk factors O, S, and D, and the lack of consideration for the weights among risk factors. Notably, this is the first instance where component importance is incorporated into the FMEA method, thereby enhancing the precision of system failure mode ranking. The proposed method is applied to a water supply system, presenting the analysis process and results. The effectiveness and accuracy of the method are demonstrated through comparisons with other approaches. This method offers fresh perspectives for FMEA analysis, elevating the scientific rigor and precision of system FMEA analysis.

The electric propulsion system of a ship represents a quintessential electromechanical system, the reliability of which is crucial to assess. This paper introduces a novel approach for modeling the reliability of electric propulsion systems using the Modelica language, designed to overcome the shortcomings of traditional reliability modeling methods that often fail to account for the heterogeneity, dynamicity, and interactivity inherent in electric propulsion systems. This approach utilizes multi-domain modeling to address system heterogeneity, employs parametric modeling to capture environmental dynamics, and facilitates device interactions through the connectors provided by the Modelica language. Additionally, modeling efficiency is enhanced through the reuse of the device model packages. An electric propulsion system from a specific vessel is used as a case study to demonstrate the modeling process and simulation results, validating the effectiveness and adaptability of the proposed method. This approach introduces a novel tool for the reliability modeling of complex electromechanical systems, thereby enhancing the precision and efficiency of system reliability assessments.

The kriging-based adaptive structural reliability analysis method has become widely used, and various learning functions have been proposed. In this study, a fast convergence strategy (FCS) for adaptive structural reliability analysis is proposed based on the Kriging Believer criterion and importance sampling. FCS considers the improvement in the accuracy of the failure probability estimation instead of overemphasizing the approximation accuracy of the limit state function. Contribution of samples to the accuracy of failure probability estimation is quantified based on the Kriging Believer criterion. FCS can implement sequence and parallel additions. The optimal importance sampling function is constructed to further improve the efficiency of the FCS. Several examples are used to demonstrate that FCS can efficiently and accurately handle complex limit state function and the engineering problem of implicit functions.

One of the important challenges in structural reliability is using only a few function calls to obtain an accurate failure probability. To solve this problem, a combination of the adaptive Kriging method and Monte Carlo simulation (i.e., AK-MCS) has been proposed. However, AK-MCS may not be the most efficient method to estimate small failure probabilities because of the large candidate sample pool. This study proposes a new method that combines the reliability analysis method based on importance sampling and k-medoids clustering (RBIK) with subset simulation (SS) to estimate small failure probabilities. The proposed method replaces the MCS sample pool in RBIK with a smaller SS population to overcome the limitations of computer memory. Considering the influence of epistemic uncertainty on SS conditional samples, the conditional quasi-optimal importance distribution is derived. Then, a partial convergence condition (PCC) is introduced to control the estimation accuracy of failure probability. To accelerate the convergence of PCC, the parallel addition strategy based on the partial optimal importance sampling function and k-medoids method is presented. Finally, four examples (i.e., two numerical and two engineering examples) are studied. The results illustrate that RBIK-SS can solve rare failure events with satisfactory accuracy and efficiency.

Phased Mission System (PMS) is widely used in large complex equipment such as spacecraft and satellites. Due to sophisticated structure and complex missions, the components and systems of the large complex equipment have multiple failure modes which are correlated. On the one hand, taking multiple correlated failure modes in to consideration can describe the multi-state characteristics of the system more accurately and decouple the correlation between different phases of the system more clearly. On the other hand, its complexity makes the reliability evaluation of PMS more difficult. To solve this problem, this paper discusses the modeling method of multiple failure modes and their correlation, and proposes an improved combination method of Multiple Failure Mode Tree (MFMT) and Multi-state and Multi-valued Decision Diagram (MMDD) model, completes the reliability evaluation of PMS in cross-level (failure mode-component-system) and cross-scale (single phase-multiple phase). Finally, the MFMT-MMDD model of the satellite attitude orbit control system (AOCS) is established and Monte Carlo simulation is used to verify the feasibility and accuracy of this modeling method.

In structural reliability analysis, Kriging-based adaptive analysis approaches considerably improve the analysis's efficiency by utilizing proper learning strategies. However, the time cost of building a Kriging model may become unacceptable when input spaces have high dimensions. This study proposes a novel method to tackle this difficulty. The basic idea of the proposed method is to implement the adaptive analysis process in the lowdimensional space that activity scores have identified. To determine the activity scores of variables, the active subspace method is implemented. Then, the proposed adaptive analysis method based on the interval squeezing method (ISM) is applied to the suggested low-dimensional space. ISM is designed to improve the accuracy of failure probability with cognitive uncertainty by squeezing its interval. Considering the difficulties associated with applying ISM directly, two variants of ISM are developed: the sequential interval squeezing method (sISM) and parallel interval squeezing method (pISM). Finally, five typical high-dimensional examples (i.e., three numerical and two engineering examples) are investigated to verify the performance of the proposed method. The results indicate that the proposed method can maintain satisfactory accuracy and efficiency.

Due to the insufficient feature learning ability and the bloated network structure, the gear fault diagnosis methods based on traditional deep neural networks always suffer from poor diagnosis accuracy and low diagnosis efficiency. Therefore, a small channel convolutional neural network model under the convolutional block attention module (CBAM-SCCNN) is proposed in this paper. Firstly, a small channel convolutional neural network (SCCNN) model is constructed based on the framework of the traditional AlexNet model in order to lightweight the network structure and improve the learning efficiency. Then, the convolutional block attention module (CBAM) is embedded into the SCCNN model to extract the "what" and "where" information of important features from the channel and spatial dimensions respectively to enhance feature recognition ability and improve the accuracy of the model. Finally, the diagnostic performance of the CBAM-SCCNN model is verified using the vibration data of tooth-broken gears, which were obtained by an own-designed experimental bench of the swing arm gear from a naval gun loader. The findings demonstrate that the proposed model improves fault diagnosis accuracy while significantly reducing diagnostic time.

Dynamic Distributed Cooperative Systems ( DDCSs ) are complex repairable systems with the characteristics of distribution, dynamism, and cooperation, which makes it difficult to evaluate their mission reliability. Traditional reliability analysis methods either simplify the system or face the problems such as dimensional curses and space explosions. To address this issue, the paper proposed a multi-agent based modeling and simulation method. The method can first construct multiple agents to decompose the "distributed" characteristics of systems, and secondly define the state transfer of agents to describe the "dynamic" characteristics during operation, and then determine the communication rules between agents to simulate the "cooperative " characteristics during mission process and achieve the mission reliability evaluation. At last, an aviation security system of an aircraft carrier is taken as a case study, the feasibility, flexibility, and strong solving ability of the proposed method are demonstrated by evaluating its mission reliability.

In gear fault diagnosis, most current intelligent fault diagnosis methods show good classification performance for fault pattern recognition. However, when detecting fault severity, the difficulty of diagnosis is increased due to the high similarity between the monitoring signals, which requires improving the sensitivity, stability, and accuracy of diagnosis methods. To address this issue, a parameter-optimized deep belief network (DBN) based on sparrow search algorithm (SSA) is proposed for gear fault severity detection. Firstly, the initial DBN is trained by the labeled gear fault signals in different severities. Secondly, SSA is introduced to optimize the learning rate and the batch size of the initial DBN, so as to avoid the interference caused by selecting network parameters by subjective experience. Finally, the detection method of gear fault severity based on the improved DBN with the optimal parameter combination is constructed. The performance of the proposed method is evaluated by analyzing the gear datasets under five degrees of tooth-breaking fault, the results show that the average detection accuracy reaches over 96% with a standard deviation of 1.46%. Compared with other methods, it is proved that the proposed method has better feature extraction ability, stability, and accuracy for gear fault severity detection.

Reliability of motorized spindles has a great effect on the performance and productivity of computer numerical control (CNC) machine tools for intelligent manufacturing. Condition-based maintenance (CBM) is an efficient method to prevent serious failures, to improve system reliability, and to reduce management costs for motorized spindles. However, owing to various degradation features acquired during condition monitoring, the challenge is to propose an appropriate feature to evaluate the reliability level of motorized spindles and to set up optimal CBM policies. Based on the motivation, a three-stage approach is proposed in this paper. In the first stage, proportional hazard model (PHM) is developed to describe the reliability considering failure events together with multiple degradation features. Next, statistical process control (SPC) charts are constructed for condition monitoring and anomaly detection in order to achieve early detection of potential failures. At last, a CBM schedule is modeled in consideration of maintenance cost minimization; the maintenance plan is optimized by determining the optimal control limits of SPC charts.

A new compiling method based on cutting force model is proposed for an eight-block program-loading spectrum of a motorized spindle. In the proposed method, all loads exerted on the motorized spindle are dependent, which is accurate for loading in reliability tests. A cutting force measuring system is set up to obtain cutting forces accurately. Then, two concluded cutting force models are compared to choose the better one through fitting the measured cutting force signals under different machining process parameters. In accordance with the characteristic of the milling load, program-loading spectrum for the amplitude of radial force is compiled while neglecting other factors. Thereafter, the compiled spectrum is converted into a complete program-loading spectrum where torque and loads’ correlation are calculated by the chosen cutting force model. Finally, the accelerated factor of the program-loading spectrum relative to actual machining conditions is solved.

The reliability analysis of complex repairable systems is an important but computationally intensive task that requires fitting the failure data of whole life cycles well. Because the existing reliability models are mainly based on the assumptions that systems are either unrepairable or will go through an “as good as new”type of repair which does not describe actual situations effectively and precisely, a two-segment failure intensity model based on sectional non-homogeneous Poisson process is developed. This model is capable of analyzing repairable systems with bathtub-shaped failure intensity. It considers minimal maintenance activities and preserves the time series of failures based on the whole life cycle. The advantages of this model lie in its flexibility to describe monotonic, non-monotonic failure intensities and its practicality to determine the burn-in or replacement time for repairable systems. Three real lifetime failure data sets are applied to illustrate the developed model. The results show that the model performs well regarding the Akaike information criterion value, mean squared errors, and Cramér-von Mises values.

To improve the deficiency of previous reliability evaluation methods of CNC machine tools, an evaluation method based on function monotonicity is proposed to evaluate the reliability of CNC machine tools. First, residual correlation method is used to eliminate abnormal data. Second, an empirical distribution function of CNC machine tools is established through mean rank method. Furthermore, reliability models of CNC machine tools are developed, and model parameters are estimated through maximum likelihood method. A comprehensive evaluation method is adopted to select the optimal distribution function, which is the model used in our study. Finally, the interval of mean time between failures is estimated on the basis of function monotonicity. The proposed method is also used to evaluate the reliability of a certain CNC lathe. Results verify the feasibility and effectiveness of the proposed method.

After the random failure period, CNC machine tools will enter the wear-out failure period as a result of abrasion, fatigue and aging. Therefore, it is extremely important to monitor the changing trend of operation state and thus to construct rational maintenance policy or determine when to scrap it. In this paper, the mean time between failures (MTBF) is chosen to be the characteristic variable of reliability degradation and is estimated from a Weibull process model. A statistical process control (SPC) chart is then developed using the Metropolis-Hastings (MH) algorithm for condition monitoring. The availability and sensitivity of the proposed method are illustrated through analyzing the field data of a CNC machining center.

The purpose of this paper is to evaluate remaining useful life (RUL) of machine tools, which is a key indicator for equipment maintenance and health management. The proposed method takes the factors of reliability, maintainability and economy into account and constructs the evaluation index system based on analytical hierarchy process (AHP) method. Euclid approach degree is then calculated through the fuzzy matter-element model in order to assess the health condition and determine RUL. An application example of a heavy-duty machine tool is provided to demonstrate the feasibility in engineering practice, results that are compared with fuzzy comprehensive evaluation and grey relational analysis illustrate the reliability and accuracy of the method.