Assessment of the Electric Demand Management Potential of Educational Buildings’ Mechanical Ventilation Systems

Abstract

Demand management is expected to reduce emissions from energy systems and support the utilization of renewable energy sources. In this paper, the focus is on the viability of educational buildings’ mechanical ventilation systems’ participation in electric demand management. The results suggest that when load shedding lasts for a short duration, the ventilation machine load seems more promising than expected for electric demand management, as even 60% of its electric power could be granted to such markets. Prolonging the load-shedding duration increases the risk of the indoor carbon dioxide (CO₂) concentration exceeding the limit for good indoor air quality. This paper contributes to the academic community by providing information for the assessment of the demand management potential of buildings and eventually their significance in decarbonizing the electric energy system and filling research gaps concerning the impact of implementing demand management that involves a reduction in ventilation rate.

1. Introduction

1.1. Maintaining the Energy Systems’ Balance

In energy systems, the balance between production and consumption must be constantly preserved. This balance is maintained by regulating production according to demand, supported by demand management implemented by the consuming party [1]. When the production side increases the exploitation of renewable energy sources and seeks to reduce the use of polluting peak power plants, the significance of the adjustability of consumption increases. Since buildings consume a considerable amount of energy, technically and economically viable methods are explored to engage their demand management potential.

However, many factors limit the buildings’ potential to adapt their energy consumption due to external events. Demand management should neither hinder the principal function of buildings—as safe and productive sites for working, studying or living—nor compromise the performance of the buildings’ technical systems or their structures. Energy consumers in non-residential buildings that could participate in demand management include heating, ventilation and air conditioning (HVAC) systems, chillers, boilers, pumps and heat exchangers [2]. In each consumer category, compliance with specific rules and technical implementation are also requirements for participation in demand management.

Demand management activities also depend on the energy system involved. In Nordic non-residential urban buildings, the energy carriers are generally the electric grid and district heating networks. District heating companies can be considered natural monopolies, so a solution developed for a company in one city may be difficult to copy in other cities. The benefits of demand management are also challenged by the considerable internal inertia of district heating networks and the changing consumption profiles of different customer segments [3,4]. In contrast, for electric demand management, marketplaces are available whose rules are identical across larger areas, including nationwide.

1.2. The Demand Management Potential of Buildings

Electric demand management can be implemented for various objectives, including cutting peak energy consumption or participating in energy or balancing power markets [5]. In the district-heated building stock, a prominent load type involved in electric demand management is the ventilation machine load [6]. Ventilation machines consume a significant amount of electricity, their power is often steplessly adjustable, and they are interconnected with building automation, making implementation of demand management possible without major investments in new equipment [7].

However, ventilation is primarily responsible for maintaining healthy and comfortable indoor conditions for the occupants of a building. If ventilation is reduced too much, the concentration of indoor air pollutants will increase, and indoor climate ratings will not be achieved. More information is needed to understand which indoor air pollutant is the most critical when implementing demand management in ventilation power and how much of that power could be available for demand management. This leads to our research questions:

Research question 1. “How does mechanical ventilation demand management affect the concentrations of the most common pollutants measured from indoor air?”

Research question 2. “What share of mechanical ventilation power in educational buildings may be designated for electric demand management while preserving their indoor air quality based on the set rates?”

This research was carried out through field tests in an educational building. When the building was occupied during school hours, temporary power reduction commands, imitating the outcomes of demand management, were given to ventilation machines, and their effects on indoor air quality were monitored. Because the quality of indoor climate is a factor that is also influenced by how occupants perceive it, occupant surveys were conducted as well. This allowed us to analyze how different indoor air determinants were developed during the activation of demand management, as well as to estimate the demand management potential of mechanical ventilation, which is essential in evaluating its economic profitability.

1.3. Motivation for the Research

An educational building was chosen as the subject for the experiments because a substantial number of such buildings exist; for example, the City of Helsinki alone is responsible for 235 educational buildings. Thus, it can be speculated that they comprise a distinct group of buildings with a considerable combined demand management potential. The potential estimate obtained in this way can also be used as a starting point for evaluating the demand management potential of mechanical ventilation in other types of buildings. This paper contributes to the academic community by providing information on the assessment of the demand management potential of buildings and their significance in decarbonizing the electric energy system, and also by filling research gaps concerning the impact of implementing demand management that involves a reduction in the ventilation rate.

2. Electric Demand Management and Air Quality of Buildings

With electric demand management, a building may contribute to balancing the electric grid by adjusting its consumption according to the grid’s fluctuating needs. The longer the demand alteration lasts, the greater its impact is on the building’s processes. Basically, two marketplaces exist: market demand response (DR), namely the energy market, and physical demand response, also known as the ancillary market [8]. In the energy market, a building may reduce its energy costs by shifting its heavy consumption times from the expensive to the inexpensive periods. For Finnish consumers, the hourly changing energy price is set on the Nordpool exchange. In the case of the ancillary market, a building could join the balancing power markets of the national transmission system operator Fingrid, which requires a maximum of a 30 min activation time in its FCR-N and FCR-D markets, or the intraday markets of Nordpool, with future 15 min settlement periods (Figure 1). Some ancillary marketplaces are asymmetric, where the capability to reduce power, depending on the need of the grid, is traded; the others are symmetric, where both the increase and the decrease in consumption are traded [9,10].

“Good indoor air quality (IAQ) is achieved by providing air in occupied spaces in which there are no known or expected contaminants at concentrations likely to be harmful (…) and air with which virtually no occupants express dissatisfaction” [12]. Air quality can be monitored by measuring various factors affecting indoor air quality, such as temperature, humidity, carbon dioxide (CO₂), particles and volatile organic compounds (VOCs). Occupants are the primary source of CO₂ in occupied indoor spaces. TVOC refers to the sum of the concentration of the volatile organic compounds [13]. Examples of materials with high emission rates are paints and coatings, adhesives, carpeting, etc. TVOC represents an indication of contamination load in the indoor environment and of the sustainability of the ventilation rate supplied [14]. The particles (PM) found in the indoor air are mainly of outdoor origin. Different particle sizes travel differently in ventilated rooms [15].

The Finnish Ministry of Social Affairs has determined various limit values for acceptable indoor air quality in its regulation [16]. They include the indoor temperature limits for educational buildings (20–26 °C; 32 °C outside the heating season) and the limit for CO₂ concentration in indoor air (1150 ppm higher than that in outdoor air). However, the City of Helsinki (together with many other public building owners) has decided to comply with the Indoor Air Association’s guidelines for good indoor air quality, which set stricter limit values: the lower indoor temperature limits are 20.5–21 °C, and the upper limits are 23–26 °C, depending on the outdoor temperature, and the limit for CO₂ concentration in indoor air is 550 ppm higher than that in outdoor air [17,18]. Building occupants also tend to report different comfort levels, even when the conditions they experience are similar. Perceived thermal comfort and sensitivity to perceptible indoor air pollutants are individual characteristics. To arrive at a definite conclusion about indoor air quality, occupant surveys are therefore needed in addition to direct measurements [19].

In

Table 1. Summary of the literature.

Publication	Building Classification	Markets	Indoor Air Metrics
Aduba et al. (2016) [20]	Office	Market DR	Thermal comfort, CO2, relative humidity
Beil et al. (2016) [24]	Office	Ancillary market	Thermal comfort
Hao et al. (2014) [6]	Educational	Ancillary market	Thermal comfort
Kjærgaard et al. (2016) [5]	Office	Ancillary market	Thermal comfort, CO2
Lin et al. (2015) [22]	Educational	Ancillary market	Thermal comfort
Lin et al. (2017) [25]	Educational	Ancillary market	Thermal comfort
Olsen et al. (2014) [21]	Commercial	Ancillary market and market DR	Not stated
Park (2020) [2]	Residential	Ancillary market	Thermal comfort
Wang et al. (2018) [26]	Residential	Market DR	Thermal comfort, CO2
Watson et al. (2012) [23]	Various	Ancillary market	Thermal comfort
Yadav et al. (2020) [27]	Educational	Ancillary market	Thermal comfort
Zhao et al. (2013) [28]	Office	Ancillary market	Thermal comfort

, we have listed publications that discuss the effect that demand management implemented on mechanical ventilation and/or HVAC systems has on indoor air quality. A few of them state numerical estimates for the ventilation power demand management potential. According to Aduba et al. [20], up to a 60% reduction in ventilation power for a maximum of 120 min was achieved without compromising indoor air quality. The experiments were run in an office building where the most densely populated room contained eight persons. After a number of experiments, Hao et al. [6] have estimated that it is feasible to use 15% of the total fan power as a flexible demand without noticeable effects on the building’s indoor environment and its occupants’ comfort. However, in their experiments, the air handling units were used for air cooling, and room temperature was the sole indoor air metric considered. Olsen et al. [21] assessed the demand management potential of the building stock of the western USA. In their assessment, acceptable shedability of ventilation power varied between 0 and 77%, depending on the building occupancy and the time of day. According to Lin et al. [22], 40% of the fan power of a lecture hall could be provided for symmetrical ancillary markets. Watson et al. [23] stated that HVAC loads of schools could offer 20 to 40% shed level, including turning off compressors. Generally, the number of publications that discuss the impact of reduced ventilation rates by demand management actions and also consider indoor air quality metrics other than temperature seems low.

3. Implementation of the Research

The majority of educational buildings in the Nordics have mechanical supply and exhaust ventilation systems. The supply air fans transfer air from outside to inside, and the exhaust air fans transfer from inside to outside. In the case of earlier systems, ventilation power was maintained in a fixed way during operational hours. Modern systems are demand-controlled, where ventilation for each room is continuously adapted by air dampers based on CO₂ feedback, and variable speed drives (VSD) control fans to maintain a consistent air duct pressure (Figure 2). The supply and exhaust fans are also controlled synchronously to minimize the pressure difference above the building envelope. The supply air is preheated with the exhaust air through heat recovery and with district heat. The supply air heating setpoint is commonly set at approximately 2 °C below the room heating setpoint to ensure that fresh air is mixed with indoor air while avoiding discomfort caused by air drag. The eventual heating of room air is the responsibility of the buildings’ heating systems. Typically, room air cooling is non-existent [29].

This study’s experiments were carried out in a physics classroom of a Helsinki upper secondary school. The fan power for the ventilation service area was regulated, and the resulting electric energy consumption levels, together with airflows, were monitored through the building automation supervisory system. Via the same automation, the data on indoor temperature and CO₂ concentration, measured by wall-mounted sensors inside classrooms, were collected. Internet of things (IoT) sensors were also installed, allowing us to monitor additional air quality parameters, including relative humidity (RH), particulate matter (PM10, PM2.5, and PM1.0) and total VOCs (TVOCs). Such sensing devices that are also technically suitable as sensors in building automation systems were used. Occupant feedback on the experienced indoor air quality was collected on a feedback terminal. Prior to the actual tests, the ventilation power curve was assessed to determine the relation between the set values for the fans’ speed and the fans’ electric power consumption.

Two load-shedding durations were chosen, 15 min and 1 h, mimicking the shortest and the longest activation durations of the available marketplaces, and test times were chosen according to the classroom timetable so that the classroom definitely was in use during the tests. In the first test, a 15% power consumption reduction was selected, based on the estimate reported in the literature [11]. The following tests were planned to double, triple and eventually quadruple the power reduction (30%, 45% and 60%, respectively), under the condition that the indoor air quality limits set by the Ministry of Social Affairs had not been violated during the previous tests. Some test periods were intended for control purposes, where fan power was not altered, but indoor air quality was recorded. Based on these preconditions, we prepared the test schedule (

Table 2. The planned test schedule.

Test Number	Test Date and Time	Fan Power Alteration
1	3.10. 10:00–10:15	−15% 15 min
2	3.10. 15:15–15:30	−30% 15 min 1
3	4.10. 10:00–11:00	Control
4	4.10. 12:00–13:00	−15% 1 h
5	5.10. 12:00–13:00	−30% 1 h 1
6	6.10. 10:00–11:00	Control
7	6.10. 12:00–13:00	−45% 1 h 1
8	7.10. 12:15–12:30	−45% 15 min 1
9	10.10. 10:00–10:15	−60% 15 min 1
10	10.10. 15:00–16:00	Control
11	11.10. 12:00–13:00	−60% 1 h 1
12	12.10. 12:00–13:00	Control

¹ Test will be run according to the schedule only if the previous test of the same duration did not violate indoor air quality limits set by the Ministry of Social Affairs.

). The classroom occupants were generally aware that tests would be performed, but they did not know the test schedule.

The duration of each test was two hours. The start time of each test coincided with the start time of the lessons. After 15 min, load shedding was started by switching the building management system (BMS) from automatic to manual mode and assigning new set values, determined during the power curve test (

Table 3. The power curve test results. All runtime values (the first row) were observed from the BMS running in automatic mode. Next the BMS was switched to manual mode and such fan speed set values were experimented that resulted in the intended reduction in total fan power while preserving the supply and exhaust air volume ratio.

Power Level	Supply Air Fan Set Value %	Exhaust Air Fan Set Value %	Supply Air Fan Power kW	Exhaust Air Fan Power kW	Fan Power Total kW	Supply Air Volume m3/s	Exhaust Air Volume m3/s
Runtime	75.0	76.0	4.25	3.90	8.15	5.2	4.3
−15%	71.2	71.2	3.75	3.19	6.94	4.9	4.0
−30%	66.2	66.2	3.12	2.59	5.71	4.5	3.7
−45%	60.7	60.7	2.44	2.02	4.46	4.1	3.4
−60%	54.1	54.1	1.79	1.47	3.26	3.7	3.0

), to the fan speed drives. The durations of the load sheddings were 15 or 60 min, depending on the test, after which the building management system was returned to automatic mode. The resulting data obtained from the energy meters and the air quality sensors were recorded.

4. Results

4.1. Power Curve

To assess the power curve of the ventilation system, we first observed the set values for the actual fan speed controller, resulting from running building automation programs, and the consequent power consumption. Next, we experimented with such fan speed set values that reduced their combined power consumption by 15, 30, 45 and 60% of their runtime consumption. Supply and exhaust air volumes were observed to ensure that their ratio remained constant (Table 3).

4.2. Demand Management Tests

The fan power shedding tests were run from 3 to 12 October 2022, according to the test plan (Table 2). During the tests, the classroom was occupied by 29 to 31 students and their teacher. The teaching lessons were each 75 min in duration, and load shedding was started 15 min after the scheduled start of the lessons. Test numbers 9 and 11 failed due to a system malfunction and a human error, respectively, but by replacing them with test numbers 10 and 12 (controls), we successfully completed all the planned power shedding experiments. The results are shown in Figure 3, Figure 4, Figure 5 and Figure 6.

When comparing the test results with the indoor air quality ratings, we observed that the CO₂ concentration in the classroom tended to be close to the limit values prior to our experiments. In the 15 min tests, the magnitude of the power reduction did not seem to be substantially correlated with the actual CO₂ concentration during the teaching lessons. However, in the 60 min tests, the indoor CO₂ concentration significantly exceeded the limit values, indicating insufficient ventilation. The room temperature increased during the tests but stayed within the rated limit values. In the other measurements that we performed (RH, PM and TVOCs), the periodic reduction in the ventilation power was not found to have effect on the results, and the level of air pollutants was generally low. However, a few highly irregular TVOC and PM readings can be observed. Such readings can be explained by occupant activity in the classroom during the tests. For example, someone may have used perfume in close proximity to a sensor.

Occupant feedback was collected on a feedback terminal. The answering process was planned to allow each student to rapidly provide feedback on individually experienced indoor air quality when leaving the classroom. The evaluation scale is shown in

Table 4. Feedback terminal evaluation scale and symbols.

Air Quality Rating	Corresponding Symbol on the Terminal	Written Criteria
1		Fresh, odorless, optimal temperature
2		Not quite fresh or odorless, or not quite optimal temperature
3		Somewhat stuffy or some odors, or somewhat uncomfortable temperature
4		Somewhat stuffy and/or some odors, and/or somewhat uncomfortable temperature
5		Stuffy and/or odors and/or uncomfortable temperature

, and the results are shown in

Table 5. Occupant feedback average for each test.

Test	Occupant Feedback on Air Quality
15 min control	3.24
15 min −15%	2.12
15 min −30%	2.68
15 min −45%	2.40
15 min −60%	3.68
60 min control	2.16
60 min −15%	1.85
60 min −30%	2.36
60 min −45%	2.48
60 min −60%	2.24

.

5. Discussion and Conclusions

In the experiments, demand management was simulated by temporarily reducing the ventilation power in a classroom during teaching periods. The effect of the reduction on the indoor climate was monitored with air quality sensors. This gave us an understanding of the percentage of ventilation power that may be allocated to downward balancing or time shifting without notable effects on the occupants’ wellbeing. A classroom is a lively environment, where the main task of a building management system is to maintain healthy indoor conditions with rule-based algorithms and information available from the installed sensors. Introducing demand management adds complexity to these processes.

When students enter a classroom, its indoor air quality begins to worsen until a new equilibrium is reached. According to our results, a 15 min power reduction has hardly any observable effects on the build-up of CO₂ concentration. However, in the 60 min tests, the equilibrium was reached during the load shedding, resulting in decreased indoor air quality even in the case of the most careful power reduction. The increase in the indoor air temperature stayed within the rated limit values during all tests. Changes in the TVOC concentration are usually caused by occupant activity indoors, while particles (PM) originate from outside the building. No trend changes in these metrics were observed during the experiments. There also seems to be no correlation between the occupant feedback and our demand management tests. This could indicate that the occupants did not notice when the experiments were occurring. However, the feedback showed a wide range of air quality ratings. More accurate methods, such as interviews, would be needed for confirmation.

Based on the experiments, we conclude that educational buildings’ ventilation machines seem promising for electric demand management when load shedding has a short duration, that is, in ancillary markets and peak-power cutting. Even 60% of runtime ventilation power could be granted to such markets. In contrast, when load shedding has a long duration, there is a considerable risk of indoor air CO₂ concentration rising beyond the limit values for good indoor air quality, even with the most cautious power reduction. The objective of shifting consumption times in accordance with the hourly energy market is therefore questionable in the case of ventilation machines in educational buildings.

According to Abuda et al. [20], up to 60% reduction in ventilation power was possible. However, their experiments were run in an office building where occupant density compared to a classroom was low. Hao et al. [6] estimated the reduction potential to be 15% without a time limit. Our results seem to be consistent with the previously published literature.

Limitations and Future Work

When the target market is an ancillary market requiring symmetrical adjustability, maximum fan power may further limit the demand management potential. Actions towards energy efficiency, such as time-based programs shutting down ventilation at night, on weekends and during holiday periods, also need to be taken into account when assessing the continual demand management potential of a building.

Although the test site was trusted to represent a regular classroom in a typical school building, all buildings have individual characteristics, which should be considered when assessing the replicability of the results. Furthermore, a classroom tends to have a large number of occupants, given its room floor area and volume; therefore, it may be speculated that in other kinds of buildings, such as offices, hotels and residential apartments, the flexibility potential of mechanical ventilation may be greater.

Further experiments are recommended in order to obtain a comprehensive understanding of the electric demand management potential of mechanical ventilation systems. Additionally, business models need to be developed to exploit the demand management potential of the building stock in a way that is profitable for the owners of the buildings, the energy systems, and the solution providers who interconnect them.