Research Publications

A.    Introduction

As energy prices rise and more attention is focused on human environmental impacts, organizations of all types are interested in reducing their energy consumption in order to lower costs, curtail greenhouse gas (GHG) emissions, and improve the overall sustainability of their operations. In response to that demand, a number of companies are developing technologies to help organizations to better manage their energy use. This paper explores the potential for one such technology, Cisco Systems’ EnergyWise software, to monitor and manage the energy consumption of network-connected IT devices on a university campus.  Examining a pilot implementation of this technology at two schools on the University of Michigan campus, the paper describes the organizational and technical challenges that arose, discusses the reductions in energy use that were achieved, and presents the project team’s recommendations and conclusions.

B.    An Evolution in Commercial Energy Management

For this study, the project team chose to use The American Council for an Energy-Efficient Economy definition of energy management: the “systematic tracking and planning of energy use and can be applied to equipment, buildings, industrial processes, industrial or institutional facilities, or entire corporations. A thorough energy management program consists of metering and monitoring energy consumption, identifying and implementing energy saving measures, and verifying savings with proper measurements.”(ACEEE 2011)
This paper focuses on energy management in buildings. Many new technologies are emerging that improve monitoring and control of major building systems including Heating, Ventilation and Cooling (HVAC), lighting and Information Technology.  Advancing codes and standards like the American Society of Heating, Refrigerating, and Air Conditioning Engineers (ASHRAE) Standard 90.1 for commercial buildings are pushing the development and adoption of these energy management technologies.
Sophisticated network technologies have been developed recently for better tracking of energy consumption in buildings, enabling higher efficiency. Since society’s prioritization of energy efficiency will directly impact the rates of technology development and adoption, it’s important to understand why energy is consequential to large organizations and estimate its level of importance. An outreach effort was performed to gauge the importance of energy management in various industries. The outreach included interviews, conversations and a survey targeted to top-level energy or sustainability managers within leading companies in healthcare, building controls, higher education, Energy Service Companies, metals, Consumer Package Goods firms, and others.
Research indicated that energy is moderately important within the organizations surveyed. There are many reasons, though most interest is driven by financial considerations. Buildings, which are composed of complex independent systems operated by unique and independent controls, are a central concern for increasing commercial energy efficiency. In order for decision makers to set more efficient operational parameters for building components, feedback data needs to be as close to real-time as possible. However, many building operators and energy managers do not have access to granular energy information for individual devices whose, but only back-dated utility meter data for the entire facility.
Energy Information Systems provide a hardware and software solution for monitoring energy consumption in buildings. Inputs to these systems include data acquisition hardware such as submeters and utility meters, formal Energy Management Systems, exported energy data from Building Management Systems, or other energy data points. This data can be normalized against weather, user behavior and other variables to more clearly indicate consumption trends and the impacts of efficiency measures.
In most organizations, Energy Information Systems (EIS) bridge a gap between operations and IT, requiring a collaborative approach between departments to properly set up data inputs to an EIS and confirm that information is being properly stored. Implementation of the EIS will likely require more collaboration between these departments, which may involve overcoming interdepartmental challenges.
Government policies and programs, regional carbon market activity, reporting requirements, and other drivers will heavily influence adoption rates for information technology energy management tools in commercial and industrial settings. Since government requirements and incentives can propel markets, technology developers should stay abreast of current and expected legislation. Other factors that may affect adoption include Energy Service Company behavior, utility-driven activity such as rebate programs, and the development of a smart grid.

C.    Energy Management in the University Environment

Universities across the country are moving toward more rigorous energy and resource management through independent campus initiatives as well as larger peer movements, such as the American College and University Presidents’ Climate Commitment (ACUPCC). The cultural aspects of the campus environment play a large role in determining the energy use patterns and considerations that must be made when assessing and implementing energy management at a university.
 The stakeholders who inhabit universities drive their institutions approach to adopting new concepts. Their motivations, needs, and characteristics vary considerably, and therefore demand a wide range of support and expertise from the university. Among these stakeholders are a few core groups: students, faculty, staff, institutes & organizations, and alumni. Each of these players strives to achieve the university’s thought and action leadership mission, but in manners that result in different behaviors, needs and, ultimately, energy use. 
A number of functional roles exist within the university to provide the environment and tools necessary to deliver the best educational and research experience. These groups interact in a number of ways, but also act relatively independently in pursuit of achieving their own organizational missions. As a result of these independent pursuits, there are a number of misaligned incentives around campus energy management that present operational and financial opportunities throughout the university. 

D.    Case Study: IT Energy Management at the University of Michigan

The central component of this project was an implementation case study of a network-based energy management solution at the University of Michigan.  As part of the study, it was necessary to survey existing energy management and sustainability initiatives at the University.  Three organizations coordinate related activity. At the highest level, the Office of Campus Sustainability coordinates the University’s overarching approach to sustainability. At the operational level, Planet Blue optimizes building and facility efficiency. Finally, the Graham Institute coordinates the University’s academic efforts around sustainability.
It was unrealistic to attempt a full building integration of an energy management solution, so the project focused on energy management of IT systems as an introduction to the University environment.  Cisco Systems, the project sponsor, donated the software and expertise necessary for the project team to deploy an energy monitoring and control package, Cisco’s EnergyWise (EW) and Orchestrator. EW is the underlying communication protocol for command, control and monitoring of devices that resides on Cisco brand network switches.  Orchestrator is a Graphical User Interface that gives administrators the ability to manage EW enabled devices.
The software was installed at two pilot locations. Those locations include the computer labs and an administrative staff office at the School of Natural Resources & Environment’s (SNRE) Dana Building, and IP phones and Wireless Access Points in the Executive Residence (ER) at the Ross School of Business.  These locations were chosen in order to compare the differences between an older, although recently renovated, building, and a newly constructed building. 
After installing the software, energy consumption was monitored for two weeks to create a baseline reference.  This information was later used to assess energy savings.  Following the baseline, the team enacted policies on computers in the labs and administrative office. 

E.    Results

Power management using EW and Orchestrator did not result in significant savings on the computers in the administrative office because the machines were already being power managed and the new policies were only slightly more aggressive than what was previously enforced.  However, using Orchestrator to manage the computers in the computer labs resulted in a marked decrease in energy consumption.  Policy implementation on lab computers caused average Sleep state (low power state) time to increase from zero to approximately 7-12 hours per day depending on user activity levels.  Average Sleep state time during a seven-day week after policy implementation was approximately 30% of the day.  The policies implemented in the labs were not aggressive in terms of idle time before sleep. Transitioning computers to a Sleep state after a shorter period of time could have resulted significant additional reductions in computer idle time and energy usage.
Analysis of computer energy state and user activity data for 12,500 of the university’s 48,000 computers revealed that approximately one-third of the computers average 21-24 hours/day in the On state. For this group, average Idle time (time in a high power state with no user activity) was over 20 hours per day, presenting a significant opportunity for management and energy savings. The individual department with the greatest number of computers that were on 21-24 hours/day deliberately prevented its computer from entering the Sleep state in order to create a good experience for users.  The department found that the login experience after its computers from the Sleep state was highly inconsistent, and this problem sometimes disrupted classes. This situation highlights an important finding: Energy management must take user experience into account in order to be accepted.
Further examination computer energy state and user activity data showed wide variation across departments in the number of hours per day that computers spend in the Idle state. This indicates that the level of computer energy management differs significantly from department to department.  Coordination of computer energy management programs could improve the overall results at the university level.
Further analysis of data showed that the opportunity for energy and cost savings is fragmented across departments. This means that the economic incentive for individual departments to manage their computer energy use is much smaller than for the university as a whole. A centralized university initiative for improved management is most likely to be effective to realize the potential for reduced computer energy use.  Based study results, the project team estimates that such an initiative could produce computer energy cost savings on the order of hundreds of thousands of dollars.

F.    Conclusions and Recommendations

Conclusions
Current limitations in energy measurement impede efforts to incentivize energy efficiency.
The University’s organizational complexity impedes implementation of energy management solutions.
A centralized Energy Information System containing university-wide data would allow energy managers to identify opportunities in complex environments.
A complete energy management solution needs to consider end-user experience.
A complete energy management solution must satisfy the needs of both IT professionals and facilities managers.
Coordination of computer energy management programs between departments could significantly improve overall results at the university level.
A centralized initiative to drive improved computer energy management is more likely to be effective in realizing the potential for reduced computer energy use.
Central PC energy management can potentially save the University several hundred thousand dollars.
Convergence of building systems would provide opportunities for even greater savings.

Recommendations
Align energy decisions and costs at the University. In particular, the project team believes that it is important to:
Investigate finance and accounting structures that better incentivize future energy management.
Map energy decision making at U of M.
Conduct a study of factors affecting user experience and their relative importance.
Implement a more robust Energy Information System at the University.
Conduct a pilot of converged building energy management.