TME Online Articles

A Foundation for Integrating Facilities Data

A new holistic software framework being developed by the U.S. Army Engineer Research & Development Center provides a platform to enable real-time awareness and effective decision-making for military installation commands.

By Louis Bartels, P.E., PMP, M.SAME, Michael Grussing, Ph.D., P.E., M.SAME, and Matthew Walters, PMP

From real property records, floor plans, and building information modeling, to condition assessments, work orders, and meter readings, vast amounts of data are collected daily about the facilities on military installations.

Despite this information-rich environment, it can be difficult to extract meaningful intelligence about the performance state of these assets to support knowledgeable decisions. Too often the data sits in stove-piped systems with little interoperability. To understand the facility portfolio’s condition, operational profile, capability to support the mission, utilization capacity, and energy demand, disparate data sources need to be connected, patterns analyzed, and results presented in a clear and straightforward way.

Supporting this particularly unique military need, the U.S. Army Engineer Research & Development Center is researching facility data integration points to develop a holistic analysis framework— the Virtual Testbed for Installation Mission Effectiveness (VTIME)—that provides a foundational platform to enable real-time awareness and effective decision-making.

REQUIREMENTS & CHALLENGES

The Department of Defense has legislative mandates and policy directives in the areas of energy, water, waste, sustainability, resilience, land and species management, and greenhouse gas emissions. Despite the record growth of these responsibilities, military installations have been faced with an unprecedented reduction in resources, particularly, personnel.

New information systems have been deployed to automate processes. The result, however, is numerous disjointed data sets that increase data entry work rather than reduce it. Furthermore, because this information is disjointed, installation managers are unable to receive a coherent operating picture and assemble the relevant information needed in forming optimal solutions. In short, managers are drowning in data while starving for information that will assist them in making decisions that maximize the effectiveness of their installations in supporting the mission.

The objective of the VTIME project is to provide actionable insights about current condition, potential threats, resiliency gaps, space/occupancy and scheduling issues, faulty equipment, cost reductions and energy savings, water usage, and transportation issues. Researchers, planners, and managers will be able to measure and test future installation management scenarios systematically with repeatable results as it relates to mission, technology, resources, resilience, and human capital risks. This is a significant capability as it allows installation stakeholders to make risk-informed decisions that optimize scarce or constrained resources, assets, and processes across multiple business lines.

Even when a base has the right built capacity, that infrastructure must be designed and configured to support the mission

Inputs into the framework include an array of sources such as facility condition, work orders, security threats, energy consumption, emergency services, training operations, base services, equipment sensor data, space planning, and real property data. The format and delivery of these data sets can be unstructured (through sensors and Internet of Things devices) or structured (through web services, relational databases, or other cloud storage sources).

Raw data is stored in the VTIME data factory where extraction, modeling, data mining, deep learning, data fusion, pattern recognition, and other machine learning techniques can be performed. The combined VTIME data is then ready for output in a data warehouse environment.

OUTPUTS & VISUALIZATION

Analysis outputs can be visualized through hierarchical dashboards that provide insights about current condition, potential threats, resiliency gaps, space/occupancy and scheduling issues, faulty equipment, possible cost reductions and energy savings, water usage, and transportation issues.

Data reporting flows up from the facility (asset) level, to facility group, installation, and finally, to a national command level. Different views can be designed for each stakeholder. And users can save customized dashboards to their profile for quick access.

Capacity. If a base does not have the right amount of a certain asset type, it can put mission at risk as personnel then have to work in overcrowded buildings, or make use of facilities not designed for their purpose. Even installations that are at capacity can quickly have missions affected when a disruptive event occurs that reduces it further. As a result, some level of redundancy or overcapacity may be important to help absorb adverse effects from a disruptive event.

Mission Alignment. Even when a base has the right built capacity, that infrastructure must be designed and configured to support the mission. Aging facilities can be less effective in meeting mission requirements due to obsolescence, changes in user requirements, and code/compliance gaps and evolving design standards. As technology and weapons systems advance and training and doctrine evolve, facilities must adapt to support those forces. This usually happens through evolving design standards for new construction; but the functional capability of the existing portfolio can be measured and reflected across the portfolio as a whole, or broken down by asset type to pinpoint investment opportunities. Furthermore, not all facilities are equally important to the mission, so being able to filter down to just mission-critical or mission-dependent is important.

Condition/Quality. A base that has the right capacity of asset classes, and those facilities are generally aligned with mission requirements, still must contend with the physical deterioration of systems and components, manifested in condition loss, which affects the performance and overall mission utility. This deterioration can further decrease capacity and capability. Existing facilities may not be able to provide the full range of services. In addition, facilities in a degraded condition may be more susceptible to a disruptive event as it takes more time and resources to recover. VTIME dashboards provide the visibility to quickly anticipate potential weaknesses in the asset portfolio that could amplify the effects of a disruptive event.

Operations. To provide reliable operations and mitigate risk from disruptive events, Public Works Directorates need to know where their mission-critical and condition-critical assets are that require the most attention. These high-risk systems and components are not always obvious to decision-makers, especially with such a large number of assets to track. Knowing what and where these assets are helps in prioritizing preventive and corrective work needs. Advanced knowledge of the root cause of potential disruptions also can speed up response and recovery time.

As the research continues, the VTIME Analytics Engine will utilize advanced data mining, deep learning, data fusion, pattern recognition, and other machine learning capabilities to support installations.

THE WAY AHEAD

To date, the VTIME research has been focused on inputs from the BUILDER Sustainment Management System (facility condition), quarterly extracts from the Army’s Headquarters Installation Information System (real property), and the System Master Planner/Net Zero Planner Tool (energy planning). This data was readily available for use and able to be quickly pulled together in a local SQL database environment and joined together at the facility level. Basic analytics were performed and several dashboards were designed to display data at the facility, facility systems, facility groups, site, installation, and national command levels.

As the research continues, the VTIME Analytics Engine will utilize advanced data mining, deep learning, data fusion, pattern recognition, and other machine learning capabilities to support installations.

The VTIME data factory will soon move to a secure cloud environment and various data input methods are being designed and constructed to gather all the raw inputs in an automated environment.