Rasa Remenyte-Prescott

An efficient phased mission reliability analysis for autonomous vehicles

Autonomous systems are becoming more commonly used, especially in hazardous situations. Such systems are expected to make their own decisions about future actions when some capabilities degrade due to failures of their subsystems. Such decisions are made without human input, therefore they need to be well-informed in a short time when the situation is analysed and future consequences of the failure are estimated. The future planning of the mission should take account of the likelihood of mission failure. The reliability analysis for autonomous systems can be performed using the methodologies developed for phased mission analysis, where the causes of failure for each phase in the mission can be expressed by fault trees. Unmanned Autonomous Vehicles (UAVs) are of a particular interest in the aeronautical industry, where it is a long term ambition to operate them routinely in civil airspace. Safety is the main requirement for the UAV operation and the calculation of failure probability of each phase and the overall mission is the topic of this paper. When components or sub-systems fail or environmental conditions throughout the mission change, these changes can affect the future mission. The new proposed methodology takes into account the available diagnostics data and is used to predict future capabilities of the UAV in real-time. Since this methodology is based on the efficient BDD method, the quickly provided advice can be used in making decisions. When failures occur appropriate actions are required in order to preserve safety of the autonomous vehicle. The overall decision making strategy for autonomous vehicles is explained in this paper. Some limitations of the methodology are discussed and further improvements are presented based on experimental results.

An enhanced component connection method for conversion of fault trees to binary decision diagrams

Fault Tree Analysis (FTA) is widely applied to assess the failure probability of industrial systems. Many computer packages are available which are based on conventional Kinetic Tree Theory methods. When dealing with large (possibly non-coherent) fault trees, the limitations of the technique in terms of accuracy of the solutions and the efficiency of the processing time becomes apparent. Over recent years the Binary Decision Diagram (BDD) method has been developed that solves fault trees and overcomes the disadvantages of the conventional FTA approach. First of all, a fault tree for a particular system failure mode is constructed and then converted to a BDD for analysis. This paper analyses alternative methods for the fault tree to BDD conversion process. For most fault tree to BDD conversion approaches the basic events of the fault tree are placed in an ordering. This can dramatically affect the size of the final BDD and the success of qualitative and quantitative analyses of the system. A set of rules are then applied to each gate in the fault tree to generate the BDD. An alternative approach can also be used, where BDD constructs for each of the gate types are first built and then merged to represent a parent gate. A powerful and efficient property, sub-node sharing, is also incorporated in the enhanced method proposed in this paper. Finally a combined approach is developed taking the best features of the alternative methods. The efficiency of the techniques is analysed and discussed.

A reliability analysis method using binary decision diagrams in phased mission planning

The use of autonomous systems is becoming increasingly common in many fields. A significant example of this is the ambition to deploy unmanned aerial vehicles (UAVs) for both civil and military applications. In order for autonomous systems such as these to operate effectively, they must be capable of making decisions regarding the appropriate future course of their mission responding to changes in circumstance in as short a time as possible. The systems will typically perform phased missions and, owing to the uncertain nature of the environments in which the systems operate, the mission objectives may be subject to change at short notice. The ability to evaluate the different possible mission configurations is crucial in making the right decision about the mission tasks that should be performed in order to give the highest possible probability of mission success. Because binary decision diagrams (BDDs) may be quickly and accurately quantified to give measures of the system reliability it is anticipated that they are the most appropriate analysis tools to form the basis of a reliability-based prognostics methodology. The current paper presents a new BDD-based approach for phased mission analysis, which seeks to take advantage of the proven fast analysis characteristics of the BDD and enhance it in ways that are suited to the demands of a decision-making capability for autonomous systems. The BDD approach presented allows BDDs representing the failure causes in the different phases of a mission to be constructed quickly by treating component failures in different phases of the mission as separate variables. This allows flexibility when building mission phase failure BDDs because a global variable ordering scheme is not required. An alternative representation of component states in time intervals allows the dependencies to be efficiently dealt with during the quantification process. Nodes in the BDD can represent components with any number of failure modes or factors external to the system that could affect its behaviour, such as the weather. Path simplification rules and quantification rules are developed that allow the calculation of phase failure probabilities for this new BDD approach. The proposed method provides a phased mission analysis technique that allows the rapid construction of reliability models for phased missions and, with the use of BDDs, rapid quantification.

An efficient real-time method of analysis for non-coherent fault trees

Fault tree analysis is commonly used to assess the reliability of potentially hazardous industrial systems. The type of logic is usually restricted to AND and OR gates which makes the fault tree structure coherent. In non-coherent structures not only components’ failures but also components’ working states contribute to the failure of the system. The qualitative and quantitative analyses of such fault trees can present additional difficulties when compared to the coherent versions. It is shown that the Binary Decision Diagram (BDD) method can overcome some of the difficulties in the analysis of non-coherent fault trees. This paper presents the conversion process of non-coherent fault trees to BDDs. A fault tree is converted to a BDD that represents the system structure function (SFBDD). A SFBDD can then be used to quantify the system failure parameters but is not suitable for the qualitative analysis. Established methods, such as the meta-products BDD method, the zero-suppressed BDD (ZBDD) method and the labelled BDD (L-BDD) method, require an additional BDD that contains all prime implicant sets. The process using some of the methods can be time consuming and not very efficient. In addition, in real time applications the conversion process is less important and the requirement is to provide an efficient analysis. Recent uses of the BDD method are for real time system prognosis. In such situations as events happen, or failures occur the prediction of mission success is updated and used in the decision making process. Both qualitative and quantitative assessment are required for the decision making. Under these conditions fast processing and small storage requirements are essential. Fast processing is a feature of the BDD method. It would be advantageous if a single BDD structure could be used for both the qualitative and quantitative analyses. Therefore, a new method, the ternary decision diagram (TDD) method, is presented in this paper, where a fault tree is converted to a TDD that allows both qualitative and quantitative analyses and no additional BDDs are required. The efficiency of the four methods is compared using an example fault tree library.

A fault detection method for railway point systems

Failures of railway point systems (RPSs) often lead to service delays or hazardous situations. A condition monitoring system can be used by railway infrastructure operators to detect the early signs of the deteriorated condition of RPSs and thereby prevent failures. This paper presents a methodology for early detection of the changes in the measurement of the current drawn by the motor of the point operating equipment (POE) of an RPS, which can be used to warn about a possible failure in the system. The proposed methodology uses the one-class support vector machine classification method with the similarity measure of edit distance with real penalties. The technique has been developed taking into account specific features of the data of infield RPSs and therefore is able to detect the changes in the measurements of the current of the POE with greater accuracy compared with the commonly used threshold-based technique. The data from infield RPSs, which relate to incipient failures of RPSs, were used after the deficiencies in the data labelling were removed using expert knowledge. In addition, possible improvements in the proposed methodology were identified in order for it to be used as an automatic online condition monitoring system.

A systems reliability approach to decision making in autonomous multi-platform systems operating a phased mission

This paper presents a decision making strategy for autonomous multi-platform systems, wherein a number of platforms perform phased missions in order to achieve an overall mission objective. Phased missions are defined for both single and multi-platform systems and a decision making strategy is outlined for such systems. The requirements for a tool performing such a strategy are discussed and methods and techniques, traditionally used for system reliability assessment, are identified to fulfill these requirements. Two examples are presented in order to demonstrate how a decision making tool would be employed in practice. Finally, a brief discussion of the efficient implementation of such a strategy is presented.

Analysis of non-coherent fault trees using ternary decision diagrams

Risk and safety assessments performed on potentially hazardous industrial systems commonly utilize fault tree analysis (FTA) to forecast the probability of system failure. The type of logic for the top event is usually limited to AND and OR gates, which leads to a coherent fault tree structure. In non-coherent fault trees, the working states of components as well as their failures contribute to the failure of the system. The qualitative and quantitative analyses of non-coherent fault trees can introduce further difficulties over and above those seen in the coherent case. It is shown that the binary decision diagram (BDD) method can be used for this type of assessment. The BDD approach can improve the accuracy and efficiency of the quantitative analysis of non-coherent fault trees. This article demonstrates the value of the ternary decision diagram (TDD) method for the qualitative analysis of non-coherent fault trees. Such analysis can be used to provide information to a decision-making process for future actions of an autonomous system, and therefore it must be performed in real time. In these circumstances, fast processing and small storage requirements are very important. The TDD method provides a fast processing capability, and small storage is achieved when a single structure is used for both qualitative and quantitative analyses. The efficiency of the TDD method is discussed and compared with the performance of the established methods for analysis of non-coherent fault trees.

Quantitative risk prognostics framework based on Petri Net and Bow-Tie models

A simulation framework based on the Petri Net model is proposed in this paper used for performing quantitative risk prognosis through extending the Bow-Tie model. A Petri Net model is built to include features, specific to assets, such as the condition of the asset, the projected operational usage, inspection and maintenance policies and degradation process, so that the future condition of the asset over time can be estimated. Several new Petri Net modelling features which advance the traditional Bow-Tie approach are proposed, such as asset usage generating and usage dependent transitions, and the possibility of entering evidence about the actual condition of the asset through the use of truncated distributions. Monte Carlo simulation method is used to simulate the developed Petri Net model over a selected time frame, in order to obtain statistics necessary to perform risk assessment using the Bow-Tie model. The paper reports on the overall proposed methodology and then focusses on the development of the Petri Net model. The methodology is applied in risk prognostics of operating an underground passenger lift. In particular, the combination of the Petri Net and the Bow-Tie models is illustrated to predict the likelihood and the consequences of an event when a lift gets stuck in a shaft between landings.

Reliability analysis of missions with cooperating platforms

A phased mission analysis approach for cooperating platforms is outlined in this paper, when a common mission goal is achieved through collaboration of a number of individual platforms. Phase failure probabilities for individual platform mission and the overall mission are used to measure the mission success. Since any platform capability can degrade during the mission, the mission failure probability can exceed an acceptable level and mission reconfiguration is then considered. Phase failure models are built, updated and analyzed using fault trees converted to binary decision diagrams (BDD), due to the efficiency and accuracy of the BDD method required to support decisions on the operation of complex systems. The application of the methodology in the decision making strategy is illustrated using a multiplatform phased mission example from the military arena.

A Petri Net Approach to Fault Verification in Phased Mission Systems using the Standard Deviation Technique

Health management systems are now standard aspects of complex systems. They monitor the behaviour of components and sub-systems and in the event of unexpected system behaviour diagnose faults that have occurred. Although this process should reduce system downtime it is known that health management systems can generate false faults that do not represent the actual state of the system and cause resources to be wasted. The authors propose a process to address this issue in which Petri nets are used to model complex systems. Faults reported on the system are simulated in the Petri net model to predict the resultant system behaviour. This behaviour is then compared to that from the actual system. Using the standard deviation technique the similarity of the system variables is assessed and the validity of the fault determined. The process has been automated and is tested through application to an experimental rig representing an aircraft fuel system. The success of the process to verify genuine faults and identify false faults in a multi-phase mission is demonstrated. A technique is also presented that is specific to tank leaks where depending on the location and size of the leak, the resulting symptoms will vary.