Bayesian Belief Networks (BBN)
A Bayesian belief network (BBN) or Bayesian network (BN), also known as a Bayesian net, is a graphical modeling tool for specifying probability distributions that utilizes directed graphs together with associated set of probability tables. These graphical structures are used to represent knowledge about a system. In particular, each node in the graph represents a random variable, while the edges between the nodes represent probabilistic dependencies among the corresponding random variables. These conditional dependencies in the graph are often estimated by using known statistical and computational methods.
A BBN is a compact representation of the joint probability distribution of the system variables. Formally, it is known as an acyclic directed graph (DAG) with nodes connected by arcs. The nodes are random variables whose values represent the observed or unobserved system variables. The arcs represent the causal relationship between variables. They are quantified by the conditional probabilities that a child node would reach to a certain value, given values of all its parent nodes.
BBN can also be used to represent the generic knowledge of a domain expert, and to function as a computational architecture for storing factual knowledge and manipulating the flow of knowledge in the network structure. The graph structure in the network significantly reduces the storage required for the joint probability distribution, and the computational burden associated with the inference process.
The audience will become familiar with the basics of Bayesian belief networks, their operation, their applications for solving real world problems and available commercial tools that can be used for BBN modeling and analysis.
Bayesian Sensor Placement Optimization (BSPOTM)
KimiaPower team is the proud developer of the new and innovative Bayesian Sensor Placement Optimization (BSPOTM) methodology. One of the strongest motivations for developing these techniques was to find a methodology for optimum sensor placement based on logical or functional placement. The Bayesian methodology developed for optimally locating the sensors throughout a system is aimed at finding the optimum sensor placement scenario for extracting the most amount of reliability information from the measured data. The approach takes into account all uncertainties within the probabilistic framework and combines the different sources of information using the rules of probability.
The developed BSPOTM algorithms utilize Bayesian networks for modeling, updating and reasoning the causal relationships and uncertainties as well as for updating the state of knowledge for unknowns of interest. Information metrics are used to assess the potential information gain for each sensor placement scenario and the results are used to select the sensor placement scenario with the highest amount of reliability information.
The audience will learn about the basics of this new methodology and how it will change the concept of sensor placement optimization across all industries.
Bayesian System Health Monitoring (BaySHMTM)
System health monitoring is defined as a set of activities implemented on a system to assure its operable condition. Monitoring activities are often limited to the observation of current state of operations. In other cases, systems are monitored solely for the prediction of future operation status and predictive diagnosis of future failure states. KimiaPower’s revolutionary Bayesian System Health Monitoring (BaySHMTM) techniques are designed to monitor the health and reliability of virtually any system from power equipment and infrastructures to trains and bridges and complex machines or even human body.
System health monitoring and sensor placement are areas of great technical and scientific interest. Prognostics and health management of a complex system require multiple sensors to extract required information from the sensed environment, because no single sensor can obtain all the required information reliably at all times. The increasing costs of aging systems and infrastructures have become a major concern and system health monitoring techniques can ensure increased safety and reliability of these systems. Similar concerns exist for newly designed systems as well. System reliability monitoring assesses the state of systems health and, through appropriate data processing and interpretation, can predict the remaining life of the system.
One way to minimize both maintenance and repair costs as well as the probability of failure is through continuous health assessment of the system and prediction of future failures based on current health and maintenance history. Therefore, one of the goals of implementing system health monitoring is to alleviate the growing concern over the maintenance of legacy equipment by replacing scheduled maintenance with as-needed maintenance and saving the cost of unnecessary maintenance, as well as preventing unscheduled maintenance activities. In other words, system health monitoring allows condition-based maintenance (CBM) inspection instead of schedule-driven inspections. Similarly, when new systems are developed and designed, appropriate power system health monitoring can be embedded in a design which is expected to reduce the life-cycle operating cost.
Another method of utilizing reliability estimates of a system is to formulate a cost-effective maintenance plan, which is often called reliability centered maintenance (RCM). The goal of RCM methodologies is to find logical ways to identify what equipment is required to be maintained on a preventive maintenance basis rather than let it run to failure.
The audience will learn about system health monitoring techniques and specifically about the philosophy behind the new and innovative KimiaPower’s Bayesian System Health Monitoring (BaySHMTM) and how it will improve the reliability of systems by predicting the systems’ health to minimize unscheduled downtime.
Military Handbook 217 (MIL-HDBK-217)
MIL-HDBK-217 was the original reliability prediction tool developed by Rome Laboratory and was published by the US Department of Defense (DoD) in 1961. According to MIL-HDBK-217, Reliability Prediction provides the quantitative baseline needed to assess progress in reliability engineering. A prediction made for a proposed design may be used in several ways:
- As a guide to improve the design
- Comparing the reliability of existing and modified systems
- As a measure for resilience under environmental effects
- Evaluate the effect of complexity on the probability of mission success
- Evaluate the significance of reported failures
- Establish goals for reliability tests
The purpose of this handbook is to establish and maintain consistent and uniform methods for estimating the inherent reliability of military electronic equipment and systems. It provides a common basis for reliability predictions during acquisition programs for military electronic systems and equipment. It also establishes a common basis for comparing and evaluating reliability predictions of related or competitive designs. The handbook is intended to be used as a tool to increase the reliability of the equipment being designed.
MIL-HDBK-217 contains two methods of reliability predictions: “Part Stress Analysis” and “Parts Count”. These methods vary in degree of information needed to apply them.
The Part Stress Analysis Method requires a greater amount of detailed information and is applicable during the later design phase when actual hardware and circuits are being designed.
The Parts Count Method requires less information, limited to part quantities, quality level, and the application environment. This method is applicable during the early design phase and during proposal formulation. The Parts Count Method will usually result in a more conservative estimate (i.e., higher failure rate) of system reliability than the Parts Stress Method.
Audience will become familiar with the history of MIL-HDBK-217, methodologies used in this standard for reliability prediction, and available commercial tools developed bases on MIL-HDBK-217.
Reliability Block Diagram (RBD)
Reliability block diagram is a success-oriented network describing the function of the system. A reliability block diagram is a graphical depiction of the system’s components and its interconnections which can be used to determine the overall system reliability. Reliability block diagram modeling is a method of representing the functional relationships among the components and subsystems and indicates which ones must operate successfully for the system to accomplish its intended function. The blocks in RBD represent the system components and the lines show the interconnections among the components and subsystems. Reliability block diagrams are commonly used in reliability and availability analyses and for safety assessment.
The audience will learn about basic definitions in reliability engineering, basics of Boolean algebra, selective distribution models, calculating reliability for exponential distribution models, series systems, parallel systems, complex systems comprised of various series and parallel combinations, definition of redundancy, and many more. In addition several examples will be discussed and finally the attendees learn the name of few commercially available tools that can be used for RBD modeling.