Information Technology for Energy and Maintenance Management

We describe the design of a tenant interface for energy and maintenance systems (TIEMS) in commercial buildings. TIEMS is designed for use by occupants (tenants) of commercial buildings. Our hypothesis is that by allowing tenants access to information from the energy and maintenance systems and by giving them some control over these systems, energy and maintenance performance can be improved. We used interviews with potential users and existing energy and maintenance databases to guide the design. Results of a field trial demonstrate the utility of TIEMS. We also describe the design of a maintenance and operations recommender (MORE). MORE uses information from computerized maintenance management systems (CMMS) and energy management and control systems (EMCS) to recommend what maintenance personnel should do in response to a maintenance service request. MORE integrates text descriptions of problems with sensor information related to the problem. After work orders are closed, MORE uses the information about what was actually done to solve the problem to learn how to improve its recommendations. INTRODUCTION Modern buildings use computerized maintenance management systems (CMMS) and computer-based energy management and control systems (EMCS). These systems contain large databases of information about historical building maintenance operations. Additionally EMCS systems contain real-time information about various subsystems important to building operations, including heating, ventilating, and air-conditioning (HVAC) systems, life-safety systems, lighting systems, and power systems. One of the most advanced energy and maintenance systems developed to date is GEMnet (Piette, 2002). GEMnet is an integrated information technology infrastructure for energy and maintenance management. It uses a common database system for all components and an open communications protocol called BACnet (ASHRAE, 2001). GEMnet uses modern web-based technology for its user interface. However, the intended users of GEMnet only include maintenance personnel. The way that building occupants interact with GEMnet is by making a telephone call to someone in the maintenance department to request service or report a problem. The maintenance personnel manually enter the service request into GEMnet, sometimes long after the phone call has been made. Our work on the interaction between building occupants and energy and maintenance systems (Federspiel 1998, 2001) suggests that building occupants should also be considered users of energy and maintenance systems. Until recently this has been considered an unwise, even radical, idea among facility management professionals. However, providing tenants with a user interface to energy and maintenance systems should improve thermal comfort, improve the performance of energy management strategies, eliminate some redundant service requests, and an improve the quality of data in maintenance databases. A well-designed user interface should also improve the satisfaction of the occupants with the services provided to them by maintenance personnel. CMMS systems are used to monitor the frequency of maintenance activities and the time required to perform them. This capability is particularly useful when maintenance services are provided by third parties. EMCS systems are used to monitor energy-intensive equipment such as HVAC equipment. They monitor key system variables such as temperatures, flow rates, and pressures, and derived performance metrics such as chiller efficiency so that when alarms or problems are reported the maintenance personnel can look at these variables to diagnose the problems. There have been some efforts at integrating CMMS operations and EMCS operations. For example, some energy and maintenance systems will automatically initiate a work order in a CMMS in response to an alarm in an EMCS. In this paper we describe the design of a Tenant Interface for Energy and Maintenance Systems (TIEMS). TIEMS is designed to operate as a component of GEMnet. The next section covers the methods we used for design. The following section includes the results of our pre-design investigations, and a description of the design itself. We conclude with a section on expected benefits of TIEMS. We also propose integrating CMMS data with EMCS data for the purpose of recommending to building engineers what they should do in response to a service request or problem report from an occupant. Recommender systems are commonly used to retrieve useful documents from large databases and from the internet (Resnick and Varian, 1997). In the document retrieval application, the recommender system recommends documents that match a weighted query. The recommender system either uses feedback from the user or watches the user’s search patterns to determine the best set of weights for the queries. Hayes and Pepper (1989) describe a system that provides maintenance recommendations based on a decision-tree approach and a database of faults. The system provides sequential recommendations of tests that the maintenance technician should perform in order to diagnose a problem. I. TENANT INTERFACE FOR ENERGY AND MAINTENANCE SYSTEMS – (TIEMS) METHODS We used four sources of information to guide the design of TIEMS: interviews with tenants, meetings with maintenance personnel, historical maintenance records, and tenant scenarios. We interviewed eight tenants from three different agencies in a U.S. federal building before and after deploying the interface to determine their needs. We asked questions designed to provide information about how the complaint reporting process currently works, their satisfaction with the current process, their needs for changes, and whether or not they would be receptive to our proposed design ideas. After deploying the interface we interviewed tenants to learn about their perception of the interface and whether it met their needs. We held meetings with the building energy and maintenance staff to discuss our concepts for the design of the user interface. We described our previous research results, proposed our design for TIEMS, and asked for their feedback. We analyzed data from several maintenance databases. We looked for sensor data recorded in these databases, and we studied the problem descriptions and descriptions of actions taken. We used the statistical test described in Fleiss (1981) for comparing two Poisson parameters to test whether or not there is evidence that the service request rate after deploying the user interface is different than the service request rate prior to deploying it. The null hypothesis for this test is that the rates are equal. The test statistic is as follows: