Project Overview

In this effort, PredictionProbe, Inc. (PPI) was contracted by the National Aeronautics and Space Administration (NASA) Glenn Research Center (GRC) to demonstrate capability and sample results for a combined probabilistic and Bayesian analysis reliability assessment of the NASA Space Communication and Navigation (SCaN) assets. The existing NASA SCaN network consists of many ground-based and space-based assets. The ground-based assets include many communication antennas of various diameters (up to 70 meters or about 230 feet), command centers, networks, controllers, data storage and data processing facilities. The space-based assets include a fleet of earth-orbiting NASA communications satellites. Beyond this, hundreds of national and international, commercial, and academic customers use the SCaN assets to transfer data to earth from their in-space and remote assets. As shown in Figure 1, the NASA SCaN assets are organized under three divisions: i) the Deep Space Network (DSN, see Figure 2), ii) the Near Earth Network (NEN, see Figure 3) and iii) the Space Network (SN, see Figure 4).

Problem Definition

This effort focuses on the development of a user-friendly, quickly executing tool that can be used to enable rapid, risk-informed decisions by non-engineering users, without the assistance of software specialists. Ideally, the tool can be used by management within the setting of meeting in which decision alternatives need to be quickly evaluated. The tool must provide the specific capabilities of uncertainty/risk analysis, quantification, propagation, decomposition, and management, robust/reliability design methods, and extensions of these capabilities into decision analysis methods.

Such decision support tools must provide whatever are the required capabilities across a potentially wide variety of traditional specialty areas (disciplines) such as structures, controls, and electronics. Decision support tools should also perform analyses which span across, and are distinct from, these specialty areas to support the process of making decisions based upon computational simulations. For example, systems integration, multidisciplinary optimization, mission and trade space analyses, life cycle cost analyses, uncertainty/risk analysis and management, robust and reliability design methods, technology assessments, research portfolio analyses, and “system of systems” architecture analyses all fall into this category of capabilities. For convenience, this group of disciplines will simply be referred to collectively herein as Decision Support (DS) methods and tools.

These DS methods and tools are closely related to the processes of systems analysis (SA) and/or systems engineering (SE). The NASA Procedural Requirements, NPR 7123.1, Systems Engineering Procedural Requirements, defines systems engineering (SE) as “a logical systems approach performed by multidisciplinary teams to engineer and integrate NASA’s systems to ensure NASA products meet customer’s needs.” The same document also defines systems approach as the application of a systematic, disciplined engineering approach that is quantifiable, recursive, iterative, and repeatable for the development, operation, and maintenance of systems integrated into a whole throughout the life cycle of a project or program.”

This systems approach includes 17 common technical processes grouped under the headings of System Design Processes, Product Realization Processes and Technical Management Processes. It is expected that, at least, some of the processes should and must take place concurrently. Furthermore, it is expected that numerous passes through each of the 17 processes may occur, with iteration cycles and re-entry points established as appropriate. However, simply utilizing a good SA or SE process does not ensure that all the customer’s requirements can be satisfied within cost, schedule, or safety constraints, or with the tools and methods available. Likewise, satisfying the customer’s requirements does not mean that the results were obtained by a systematic, disciplined engineering approach that is quantifiable, recursive, iterative, and repeatable for the development, operation, maintenance, and disposal of systems. The two aspects of this problem, customer satisfaction and good SA/SE process, are mutually independent, though correlated. A good SA or SE process should include a negotiation between the developer/provider and the customer, early in project lifetime and throughout, to ensure that a reasonable chance exists to satisfy the customer’s requirements within cost, schedule, and safety constraints. The architecture of a real-time decision support tool must be very flexible and easily reconfigurable to enable all the aspects of DS, SA and SE to be implemented in a user-friendly environment.

As noted above, the SCaN Network consists of many assets such as ground assets, in-space assets and even remote assets (e.g., mars rover), all organized under three divisions of SCaN: i) the Deep Space Network (DSN), ii) the Near Earth Network (NEN) and iii) the Space Network (SN). Each asset is comprised of one or more systems (antennae, power generation, power storage, networking, etc.) and numerous sub-systems (for example sub-systems within the antennae system include gimbal, ground structure, dish, etc.). Likewise, each sub-system may contain numerous components such as amplifiers, polarizers, and diplexers. The reliability of each component and sub-system affects the reliability of the system-level assets in which they are contained, and the reliability of each asset, in turn, influences the reliability of the SCaN network. While these assets, systems, sub-systems and components have each been designed to meet certain reliability levels, in practice however, the actual reliability of these systems and assets are uncertain due to numerous uncertainties such as physical damage, power outages or interruptions, environmental conditions, unforeseen usage scenarios, and so on.

The objective of this effort was to demonstrate how PPI’s Technologies can be used to assess and quantify the reliability of the NASA’s SCaN Network and to create a decision support capability to: 1) use Bayesian technology to update input variable statistical models when additional data becomes available; 2) predict service reliability and its variation; 3) determine the impact of aging; 4) perform parametric studies if data is not available; 5) evaluate reliability of various configurations; 6) perform repair planning; 7) determine sensitivity of service reliability to its components; 8) perform resource allocation using relevant data; and 9) determine the cost of uncertainty. The work enables decision support capabilities to: 1) evaluate new concepts, 2) modify the existing network, 3) evaluate current and possible future optimal maintenance procedures, 4) develop reliability assessment and improvement strategies, and 5) develop a plan to obtain and use data to guide maintenance and acquisitions.

PredictionProbe’s Solution

Using a set of Enabling Technologies, PPI proposed to provide to NASA GRC SCaN:

• Custom, highly flexible analysis, and design framework for SCaN Reliability Assessment Software (SRAS)
• Fully integrated SCaN Compatibility Environment for Networks and Integrated Communications (SCENIC) Emulation Lab within SRAS environment
• Fully reconfigurable and updateable system-wide SCaN model
• Combined probabilistic and Bayesian analysis technologies
• Static and dynamic reliability assessments (over time, or with changing services and/or components)
• Current, alternative, and/or future SCaN system reliabilities
• Specialized assessments for specific customers / services
• Real-time decision support for SCaN Management
• Actionable investment guidance for resource allocation process
• Actionable guidance for maintenance planning process
• Design capability for future systems
• Sensitivities, Probabilities of Failure, Most Probable Point (MPP) of Failure, Response Distributions
• Reliability assessments for services and elements (assets)

Over the course of the project, several different demonstration problems were defined, analyzed, and delivered to the customer: 1) probabilistic reliability analysis (PRA) of a network with two assets; 2) Bayesian analysis of a network with three assets; 3) PRA of a Tracking and Data Relay Satellite; 4) PRA of two Goldstone Complex 34-meter antennas within the DSN; and 5) Bayesian updating of a selected input variable statistical model of the DSN demo problem. Numerous conclusions and recommendations from each demonstration were also provided in detail to the customer. These examples are described in a related Case Study. The same decision support technology has also been applied to the spread of a deadly virus.

The approach adopted by PPI in developing the reliability assessment of the existing SCaN network was to envision many layers of capabilities that could be used to enhance this existing SCaN network. The enhancements were based upon PPI’s understanding of the system and several detailed conversations with the several NASA Teams charged with managing various aspects of the system. Some of the possible enhancements to existing SCaN network included the use of optical transmissions instead of radio frequency communications, on-the fly reconfigurable communication paths, and increased redundancy in various aspects of the SCaN network. Also, the NASA Team was interested in integrating their NASA GRC SCaN Compatibility Environment for Networks and Integrated Communications (SCENIC) Emulation Lab within the PPI SPISE environment. PPI then embedded most of these capabilities (not SCENIC integration) within a very easily reconfigurable SPISE workflow. Thus, the system to perform the reliability assessment of the existing SCaN network could easily and quickly be reconfigured to analyze a wide variety of possible future systems.

Results & Conclusions

Figures 6 and 7 illustrate two different versions of an easily reconfigurable sample communications network, similar to elements within the SCaN network. In Figure 6, one specific communication path is highlighted in red. In Figure 7, a different specific communication path is highlighted in red. Likewise, the use of optical transmission instead of radio frequency communications, and increased redundancy in various aspects of the SCaN network could similarly be modeled in much the same way.