Energy management often focuses on technologies that increase the energy efficiency of key energy-consuming processes and equipment. This certainly is a key ingredient to increased energy efficiency, but often such steps do not result in consistent and sustained savings.

Uneven results can also handicap decisions toward undertaking new energy management initiatives. For project proposals that compete for funding and involve significant capital investment, risk assessments may be required that include projected savings. However, as energy consumption is often strongly linked with variable external factors like outdoor temperature and production volume, past and future savings are often difficult to attribute.

Executives may also insist that energy efficiency projects follow the management philosophy outlined in modern management practices such as ISO 9000 and Six Sigma, which highlight the importance of measuring baseline performance, setting goals, and tracking performance against those goals. These steps are included in some recently developed best energy practices, including the International Performance Measurement and Verification Protocol and Georgia Institute of Technology's MSE 2000.

Similar to quality management programs, information systems can be built to support energy management. Enterprise energy management (EEM) systems can effectively collect relevant data, as well as compensate for external factors and provide the information required to monitor performance and keep projects on track. Unfortunately, many of these systems overwhelm users with the volume of data generated. Applications with hundreds of monitored points can quickly become unusable; a "catch everything" approach will make finding valuable information either time consuming or impossible.

A well-designed EEM system considers the "nuggets" of information required to support the key goals of an energy management plan. Modern business management practice refers to these as key performance indicators (KPI), and they will make any energy efficiency initiative easier to manage and more effective.

Defining Your KPIs

Start planning your EEM system by considering how it will support your energy management goals. Then convert those goals into key performance indicators that can be measured and tracked. As an example, consider this goal statement from Executive Order 13123, Greening the Government Through Efficient Energy Management, targeted at U.S. federal industrial and research facilities:

Through life-cycle cost-effective measures, each agency shall reduce energy consumption per square foot, per unit of production, or per other unit as applicable by 20 percent by 2005 and 25 percent by 2010 relative to 1990.

Based on this, the following sample KPI definition could be used for a particular facility:

1. 1990 baseline energy consumption determined from electric and gas utility bills.


2. 1990 baseline production volume determined from manufacturing resource planning (MRP) system.


3. 1990 baseline measurement = energy consumption per unit of production, reduced by 20% to set the 2005 target, and by 25% to set the 2010 target.


4. Energy consumption data (gathered via metering on all service points), summed monthly and annually for reporting against targets.


5. Production volume (from MRP system) summed monthly and annually for reporting against targets.
   
6. At end of each year, consumption and production are combined to generate an energy performance metric, combined with others and presented to the executive team.

This definition provides the foundation for determining what data to collect, how often to collect it and how to present it. Assumptions being made are stated so that everyone involved understands exactly what is being measured.

Figure 1 shows how this definition could be further expanded to allow an energy manager to keep things on track by "drilling down" to richer details regarding what behaviors might be causing deviations from the plan. Electricity and gas consumption can be broken down to each major consuming category (e.g. motors, heaters, other) and by shorter daily or even hourly intervals. By first understanding the underlying drivers of the performance metric, managers will be able to determine which measurement details they need to help correct deviations from target goals.

Data Collection

The next step is to determine how the required data will be collected. Using a KPI approach, only the total amount of data required to accomplish the primary goals is captured, which can be an order of magnitude less than a typical energy information system. This data tends to fall into one of two main categories:

• Static -- e.g. facility floor space and equipment ratings, typically collected during an initial audit and used to normalize measurements for benchmark comparisons.


• Dynamic -- e.g. energy consumption, external temperature, production volume. This data tends to be more expensive to manage because continuous effort is involved in acquiring and processing it, and because it takes up the vast majority of storage space. For this reason, selecting which data to collect should be done with care.

Once the measurement parameters required have been selected, consider potential data sources. For energy consumption measurements, data can be manually keyed in from utility bills, or read from a "shadow" meter installed at the utility service point. For more detailed breakdowns, submeters can be installed on major loads or other points within an energy distribution system. Some building and industrial automation systems equipped with simple metering may be able to provide basic energy consumption data.

Temperature data is available from government weather office websites, or can be purchased through live online services that provide current and forecasted temperatures. You can also take your own temperature measurements, with equipment choices ranging from simple to sophisticated. As mentioned above, production volume data can typically be sourced from an existing manufacturing planning or automation system.

Basic Modeling

The next important step is to build basic models that highlight the relationship between energy consumption and the primary driver of that consumption. Proper modeling ensures the KPIs accurately reflect the impact of actions taken to manage energy.

For buildings, there is normally a direct relationship between the energy consumed by a building and outdoor temperature. For production processes where energy use is largely determined by the physics of the process (such as heat-based and chemical processes), there is normally a direct relationship between the energy consumed and production volume.

The three-step process of building models, illustrated in Figure 2, involves selecting a baseline period, creating and testing a baseline model, and creating one or more target models to track performance. Note that the following method is quite basic and will not generate robust energy consumption models in all circumstances. More sophisticated techniques are available, for example those described in ASHRAE RP-1050, Inverse Modeling Toolkit: Numerical Algorithms.

The baseline period is created by choosing a data set selected over a defined length of time to represent the energy-consuming behavior of some load -- for example, a building or process -- before an energy management plan is implemented. Data normally consists of energy consumption and the associated primary driver over a common time interval, such as daily. To be accurate, the period should encompass the time need for the load to cycle through its entire operating range. In the case of a commercial building, you would choose one year, to capture the behavior across all seasons.

The baseline model is created using a visual scatter plot showing a "best fit" line, highlighting an often strong relationship between energy consumption and the primary driver. Figure 3 shows an example of energy consumption plotted against production volume over a baseline period, which can be easily created using Microsoft Excel or other such spreadsheet applications. In many cases there may be a strong linear relationship between the two variables, but where this is not the case, technical resources are available to help interpret results and develop appropriate baselines.

The next step, target modeling, provides the "yardsticks" by which the success of energy management activities will be measured. Target models are constructed by applying the key goals embedded in the performance metrics against the baseline model to generate the reference model that ongoing measurements will be compared against.

Figure 3 illustrates two target models based on the Executive Order 13123 example used previously. For the first goal, reducing the slope and intercept constants of the baseline model by 20% creates the straight-line equation for the 20% energy reduction model. Alternatively, if the goal is to reduce consumption to "best practice levels," the normalized best-practice units for a particular industry type or application can be used to create a simple straight-line target model.

Tracking Performance

The final step is to build information displays using the data collected and models created. These displays should include high-level KPI views to give a general indication of energy management performance, as well as detailed drill-downs to help understand why a plan might be starting to go off track. The best choice of format for each will depend on what information is being conveyed and how it will be used.

Considering again the example of Executive Order 13123, a simple table is often the best way to organize and display the high-level target numbers that support a particular KPI. In this case, a table would present a list of target reductions over each of five years to bring energy consumption to the final 20% goal. A bar chart like Figure 4 can then be used to compare current and past performance against target goals, year-by-year, and further broken out month-by-month.

For more detail, a time-series chart can provide a view of the data behind the high-level KPI. For example, after identifying that measured energy consumption exceeded the target reduction in February, the more detailed chart in Figure 5 reveals that much of the deviation from the goal occurred on the third and fourth days of that month.

By reviewing high-level KPIs first and drilling down into details only when there are deviations from target goals, an energy manager can avoid searching though thousands of data points to find the few that are of interest. Note that detailed data captured while KPIs are on track are still of value; this data can be used for a variety of other tasks, including the development of operating "profiles" for monitored equipment.

There will typically be different audiences within an organization for the different information displays described above. All stakeholders will be interested in the high-level KPIs, while the core energy management team will be the primary audience for detailed drill-down views. An energy manager will likely make use of both the high-level KPIs as well as some selected detailed views when presenting updates to executives.

Enterprise energy management systems are becoming a key part of modern energy management practice. This is especially true as advances in technology make components more widely available and the cost to acquire an increasing number of measurements steadily decreases. This, in turn, can increase the "total cost of ownership" on the data management and processing side of the equation. EEM systems that support a performance management approach to energy can address this challenge by helping organizations focus only on the key indicators and supporting data that affect results. Further, the quality of insight such systems deliver can help realize more consistent and sustainable savings beyond those resulting from energy-efficient equipment upgrades alone.