Presented at National Commissioning Conference, 1999.

Use of Billing Simulation Tool for Commissioning

David Robison, Howard Reichmuth

Stellar Processes, Inc.

Synopsis

A spreadsheet tool has been developed that allows quick adjustment of a simplified engineering model to match actual utility bills. The tool utilizes billing analysis of commercial facilities to:

Diagnose energy patterns and end use consumption;
Calibrate savings estimates to agree with actual usage;
Verify vendor claims for energy products and services;
Generate performance targets and compare against actual energy bills. This application represents a low-cost, simplified commissioning check.

The tool is designed to operate with only simple information about the facility and to focus on the HVAC system. It represents one quick approach to treating the facility as an integrated whole. Case examples illustrate how the tool is useful in diagnosing energy problems, guiding on-site audits, establishing predicted targets for O&M tracking and performance verification.

About the Authors

David Robison and Howard Reichmuth have extensive experience in verifying energy savings in commercial facilities. They participated in end-use monitoring of the Energy Edge project and in end-use planning for Demand Side programs. They assisted in the development of commissioning procedures for PacifiCorp's programs, including field procedures for use of short-term dataloggers. Their interest has been to provide techniques that lead to confirmation of annual energy savings, as well as the verification of current operations.

Introduction and Background

The process of commissioning relies on functional performance tests to verify that equipment is operating correctly. Such tests may involve review of control settings, one-time measurements, short-term monitoring or specific equipment tests. Skilled engineering consultants are required to conduct the test and interpret the results. As a result, the process of commissioning is an expense difficult to justify for small projects.

The commissioning process is silent on another important issue -- do the expected energy savings actually occur? Since the commissioning takes place at one point in time, commissioning agents are careful to point out that they cannot quantify savings that occur during a different season or under different operating conditions. All the commissioning can do is to point out that the equipment is operating as designed -- one must then assume that the overall energy savings will occur. "Building squirm" or change in operating conditions is one possibility that interferes with assuming actual savings will match expectations. A commissioning verification of the whole building performance has been difficult to do, short of a complex re-modeling exercise. Yet, verification of the actual amount of savings may be very important to certain customers that need to justify the financial investment or that have performance-based contracts.

From the customer's standpoint, utility bills are where the "rubber hits the road". Yet commissioning agents use billing data in only a cursory manner for two reasons. First, if commissioning is done shortly after the installation, there is no post-retrofit billing data available. Second, both analysts and customers are skeptical of simply comparing energy bills. Bills are sensitive to weather and the length of the metering interval -- they rarely match predictions exactly. While energy accounting programs normalize for weather variations, the methodology is not transparent to the user. Normalizing corrections are primarily statistical and without relationship to building physics. For a commissioning of conservation measures, billing analysis needs to be tied to the engineering and control parameters of a building.

The Billing Simulation Approach

As one approach to resolve these problems, we developed a "billing simulation" tool. This tool is a spreadsheet that ties together whole-building level billing data and a simplified engineering simulation model of a commercial facility. The tool is designed to quickly "tune" or calibrate the engineering model to match the bills, using actual site weather. The tuning process often provides diagnostic insight toward identifiable operation problems. Of course, the tuned model can provide calibrated estimates of conservation savings on an integrated basis. More importantly, the tuned model can be used to predict future billings, taking into account the actual, local weather and operations. The predictions represent performance targets. Comparing the post-retrofit bills to the targets provides a first-order commissioning check at low-cost. For small projects, this check may be the only affordable methodology. The comparison may facilitate performance-based contracting by providing answers in a format the customer can understand. Or it can be used for on-going quality assurance -- to make sure that measures persist over time.

The key to this process is the "tuning graph" shown in Figure A. This graph shows an example of how the billing data may identify operational problems. The data points are billed consumption normalized for differences in duration of the billing period and for building size. They are plotted on the y-axis as average energy usage versus average temperature on the x-axis. We refer to this visual as a "tuning" graph because it also facilitates quick adjustment of the engineering model to match actual bills. This example shows a case where the economizers were locked in the full open position during the heating season. After the problem was fixed, consumption fell to the lower line.

Figure A. Example of Operational or "Tuning" Plot

This visual image presents the building energy use (electric, gas, or key enduses) in a picture of average operations versus mean temperature. Billing data typically form a reliable pattern in this analysis space. These patterns are intuitively comprehensible and they bear an engineering relationship to enduse interactions of building energy use. Review of this plot provides the first level of an operations check -- is the building performing as expected? Under all temperature conditions? Using this type of plot, one can often identify operational errors for specific equipment problems, even though the data are collected at the whole-building level. Since these utility data are readily available, they provide a low-cost checking mechanism.

Once the tuning graph has been properly matched to be accurately modeling the facility, the model can be used to generate consumption estimates for any desired conditions. One such application is provides the calendar presentation of energy use as demonstrated in Figure B. The calendar presentation can be used to compare directly to monthly bills. In this example, the black line represents what the baseline building would have used under the same weather and occupancy as actually occurred post-retrofit. The bars show respectively the predicted and the actual electric bills.

Figure B shows an example of computing performance targets for a community college campus. The college utilized an ESCO contractor to install a series of efficiency improvements. Even though, these measures were all carefully commissioned, the college balked at purchasing a second round of measures. The college wanted to see that the first round actually produced savings before investing in a second round. The ESCO contractor had monitoring logs and commissioning tests, but these were not understandable to the customer. The billing analysis provided a result they could understand. In part, because the tuning plot was transparent to the customer for any adjustment due to weather or operations. Also the analysis related directly to what the customer perceived as the result -- their own monthly bills. This graph convinced the customer that the project was on-track for savings and they then entered into discussions with the ESCO for a second round of measures.

Figure B. Example of Performance Verification or "Commissioning" Plot

This example shows that an initial level of commissioning can be provided based on the utility bills instead of functional tests. If the building is on-target for savings, the measures must be operating correctly. We refer to this visual as a "commissioning" graph. Together the tuning plot and the commissioning graph provide a crosscheck on both the physical (temperature) and time varying energy use of a building. The methodology supports a surprisingly detailed functional understanding of building energy use.

The simulation model has been shown to provide results consistent with DOE-2, from which it was derived. Figure C shows results from a benchmark comparison of the two modeling techniques. Since the size of these projects spanned a wide range, results are presented as the Realization Rate or the ratio of actual, "tuned" savings estimates to the initial design estimates. In some cases, extended operating hours or increased floor area in the "as-built" case provided realization rates of greater than 100%.

Use of an engineering model provides certain advantages. Starting with little more than the utility bills, the model provides an estimate of energy end uses within the facility. Inclusion of sophisticated HVAC options ensures that the model offers the following:

A comprehensive analysis approach, including interaction effects between end uses;
Explicit inputs allowing changes for operations or equipment efficiency;
Results based on the actual local weather, not average weather;
Graphic outputs that are readily understood by the customer.

Figure C. Benchmark Comparison

Verification Cases

The following examples illustrate some of the ways the tool can be useful.

Large Office

Figure D. Large Office

This retrofit project was extensively commissioned including functional performance tests of equipment as installed, review of trend logs and short-term monitoring. The monitoring revealed that some initial modeling assumptions were incorrect. Specifically, plug loads and night fan usage were higher than assumed. It was, however, not feasible to redo the expensive DOE2 model for such small changes. Despite the detailed information, the service company was not able to provide the customer with a concise statement of exactly what monthly savings were accomplished.

The simplified model in Figure D was corrected for the changes revealed by monitoring but otherwise matches the DOE-2 model. Results show that actual savings are about 33% rather than the predicted 41%, with the difference explained by the monitored changes. The simplified model is better able to show the comparison because it provides results based on the actual weather compared to the actual post-retrofit bills.

Supermarket

Figure E. Initial Supermarket Billings

This example shows the commissioning graph for a supermarket that conducted lighting retrofit. At the same time, they also added a number of additional energy-efficient refrigeration cases. The customer notes that his bills have not changed and wonders if the efficiency measures have been effective. The results in Figure E are ambiguous. Any decrease in the monthly bill is small due to the added equipment and the variability of operations.

Using the model, we are able to estimate what the old store would have used with the old lighting and the old type of refrigeration for the new cooler cases. This "hypothetical" baseline provides a better representation of what the customer's bills would have been for purposes of estimating savings. In Figure F, the difference between the hypothetical basecase and the actual bills is more apparent. Based on this graph, the efficiency measures appear to be effective.

Figure F. Revised Supermarket

It must be noted that the customer may be skeptical of introducing a hypothetical baseline. The key to this approach lies in first demonstrating with the operational plot or tuning graph, that the modeler has accurately and fairly represented the building's performance. It is important that the methodology be transparent to the customer so that the extrapolating to a revised baseline will appear fair to both parties.

Retail Store

Figure G. Retail Store Tuning Graph

The retail store in this example conducted a successful lighting retrofit and dramatically reduced electric bills. They took advantage of the savings to restore air conditioning, which they had earlier chose to minimize. As Figure G shows, operations in the store have changed so that one can no longer directly compare pre- and post-retrofit bills in order to compute savings. However, the model can be used to estimate what bills would have been had air conditioning been in use during the pre-retrofit period.

Small Office

Figure H. Small Office Pre and Post Weatherization

Figure H shows a "commissioning" graph of fuel use in government facility. Weatherization occurred during the summer. Thus, the energy usage does not match predictions during spring but matches well during fall. The facility manager was pleased with this graph because he had never before received confirmation that weatherization efforts were successful. Continued production of this graph on a routine basis provides a check to assure that O&M measure savings persist over time. This example demonstrates how the billing analysis may be used verify installations for projects that would otherwise be too expensive to permit detailed commissioning.

School

Figure I shows post-retrofit bills for a junior high school that conducted a lighting retrofit. As is apparent in the plot, the post-retrofit bills have not decreased as much as expected. Upon investigation, it turned out that the school operations changed with the addition of a community basketball camp during evening hours. For an accurate picture of the savings, operating hours need to be extended.

Figure I. School Operations Plot

Recommissioning

The tool is useful in the context of "prospecting for savings" or identifying potential existing buildings that could benefit from recommissioning. Older buildings may exhibit high energy usage that cannot be explained by lighting or space conditioning alone. This is a clue that the HVAC system is inefficient and a good candidate for commissioning.

Figure J. Hospital

Figure J shows billing data from a small-town hospital. It is apparent that a gas boiler runs during all seasons, perhaps to supply hot water -- an obvious opportunity for more efficient fuel use. Similarly, space heating is high, indicating the need for boiler tune-up or weatherization measures. However, electric consumption is low, suggesting partial occupancy or low utilization. This facility would not be a good candidate for acquiring electric savings. In this case, review of the billing data before conducting a site visit may save program expenses by helping to eliminate poor candidates.

Figure K. Existing Office

Figure K shows the billing data from an existing office building that has never been commissioned. Usage is much higher than can be explained by lighting alone. The tuning process suggests an inefficient HVAC system with leaky dampers and excessive terminal reheat. This facility would be a good candidate for recommissioning.

Conclusions

A simplified modeling tool has been developed to link utility bills and engineering simulation modeling. The tool has been demonstrated to provide similar results to DOE-2, but with greatly reduced data requirements. The primary advantage is the ability to quickly match the model to actual bills, providing a tuned, as-built model. The process of tuning often reveals opportunities for operational improvements. The completed model facilitates extrapolation from current operations to estimates of annual energy consumption. Modeling can also provide performance targets to be compared against post-retrofit utility bills. This check represents a simple-level form of commissioning that can be conducted at low cost. It also provides a mechanism for on-going quality assurance to verify that operational improvements persist over time.

Download Paper 3. Word 2000 file. Return to Home Page

Return to Library