Skip to main content
eScholarship
Open Access Publications from the University of California
Cover page of Predicting older people's thermal sensation in building environment through a machine learning approach: Modelling, interpretation, and application

Predicting older people's thermal sensation in building environment through a machine learning approach: Modelling, interpretation, and application

(2019)

There is insufficient knowledge on how environmental and physiological factors affect older people's thermal perceptions. In this paper, we present two data-driven models (a field study model and a lab study model) using the algorithm of random forests to predict older people's thermal sensation. These two models were developed from a field study dataset and a lab study dataset separately. The field study dataset was collected from 1040 old subjects (70 + years) who lived in 19 aged-care homes, which contains multi-dimension factors such as environmental parameters, subjects' demographic information, health condition, acclimatization degrees, living habits and thermal perceptions' votes. The lab study dataset was collected from a lab study and contains 18 old subjects' (65 + years) eight local skin temperatures and thermal perceptions' votes under five thermal environments (21/23/26/29/32 °C). After the procedure of feature selection, the field study model was developed with four environmental variables (air temperature, velocity, CO2 concentration, illuminance) plus two human-related variables (health condition and living time in aged-care homes) as inputs. It produced an overall accuracy of 56.6%, which was 24.9% higher than that of the PMV model. The lab study model was built on five local skin temperatures including head, lower arm, upper leg, chest and back temperatures, which demonstrated an overall accuracy of 76.7%, 30.1% higher than UC Berkeley thermal sensation model's accuracy. We then interpreted how these inputs distinguish thermal sensations by applying a partial dependence analysis. Finally, we proposed two applications of the above models and present older people's seasonally neutral indoor temperature zones.

Cover page of Comparison of mean radiant and air temperatures in mechanically-conditioned commercial buildings from over 200,000 field and laboratory measurements

Comparison of mean radiant and air temperatures in mechanically-conditioned commercial buildings from over 200,000 field and laboratory measurements

(2019)

We assessed the difference between mean radiant temperature ((t_r ) ̅) and air temperature (t_a) in conditioned office buildings to provide guidance on whether practitioners should separately measure (t_r ) ̅ or operative temperature to control heating and cooling systems. We used measurements from 48 office buildings in the ASHRAE Global Thermal Comfort Database, five field studies in radiant and all-air buildings, and five test conditions from a laboratory experiment, including both radiant and all-air spaces. Considering only the ASHRAE Global Thermal Comfort Database because it is the largest and most representative dataset, under typical office conditions, the median absolute difference (e.g., disregarding direction of the difference) between (t_r ) ̅ and t_a was 0.4 ℃ (with interquartile range = 0.4 ℃), and more specifically, the median difference shows that (t_r ) ̅ was 0.4 ℃ (with interquartile range = 0.4 °C) warmer than t_a. In the radiant cooled laboratory tests, (t_r ) ̅ was significantly (p<0.05) cooler than t_a (average difference -0.1 ℃) while in the all-air cooled laboratory tests (t_r ) ̅ was significantly (p<0.05) warmer than t_a (average difference +0.3 ℃). While these observations are significant, the effect sizes are negligible to small based on Cohen’s d and Spearman’s rho. These observations indicate that (t_r ) ̅ and t_a are typically closer in radiantly cooled spaces than in all-air cooled spaces. The results suggest that t_a measurements are sufficient to estimate (t_r ) ̅ under typical office conditions, and that separately measuring (t_r ) ̅ or operative temperature is not likely necessary to improve thermal comfort, especially in buildings with radiant systems. Furthermore, spatial and temporal variations in t_a can be greater than the difference between (t_r ) ̅ and t_a at any one location in a thermal zone, thus we expect that such variations have a greater impact on occupant thermal comfort than the differences between (t_r ) ̅ and t_a.

Cover page of Measuring air speed with a low-power MEMS ultrasonic anemometer via adaptive phase tracking

Measuring air speed with a low-power MEMS ultrasonic anemometer via adaptive phase tracking

(2019)

Indoor air movement affects many functions of buildings, including ventilation and air quality, comfort and health of occupants, fire safety, and building energy use. Accurately measuring air movement has been difficult and expensive over extended periods of time, especially for velocities below 1 m/s. A new type of high frequency ultrasonic transceiver provides high sensitivity measurements and low cost through microelectromechanical systems (MEMS) manufacturing. However, at high frequencies, conventional ultrasonic signal processing algorithms function only over small ranges of ambient temperature and velocity. In this paper, we describe three algorithms that use the complex phase angle of an ultrasonic pulse to measure velocity and temperature over extended ranges of temperature and velocity. They employ heuristics to track the vibration cycle of the measured phase angle. These methods are applied in a pulse-based anemometer whose 176kHz MEMS transceivers both transmit and receive. In wind tunnel tests between 0-4 m/s, the tracking algorithm with a low-pass filter measured air speed with high sensitivity and accuracy (0.026 m/s mean absolute error). The ability to monitor to this accuracy with such low power draw and low cost is currently unprecedented in the industry.

Cover page of Field evaluation of occupant satisfaction and energy performance in eight LEED-certified buildings using radiant systems

Field evaluation of occupant satisfaction and energy performance in eight LEED-certified buildings using radiant systems

(2019)

In this study, we present the results of a post-occupancy assessment on thermal comfort, indoor air quality, and acoustical quality from 568 occupant surveys in eight LEED-certified buildings with radiant heating and cooling systems, and trends in low-energy consuming buildings based on building characteristics, radiant design, and building operator interviews. This study follows-up on a quantitative assessment of 60 office buildings that found radiant and all-air buildings have equal satisfaction with indoor environmental quality, with a tendency for increased thermal satisfaction in radiant buildings. Our objective was to investigate reasons of comfort and discomfort in the radiant subset, and to relate these to building characteristics and operations strategies. Our analysis revealed that the primary sources of temperature dissatisfaction are lack of control over the thermal environment (both temperature and air movement) and slow system response, both of which were seen to be alleviated with fast-response adaptive opportunities such as operable windows and personal fans. There was no optimal radiant design or operation that maximized thermal comfort, and building operators were pleased with reduced repair and maintenance associated with radiant systems compared to all-air systems. Occupants reported low satisfaction with acoustics. This was primarily due to sound privacy issues in open offices which may be exacerbated by highly reflective surfaces common in radiant spaces. Indoor air quality satisfaction appears to be aligned with thermal comfort more than ventilation strategy, and buildings with low annual energy consumption take advantage of free cooling and avoid heating and cooling in the same day or same season.

Cover page of A thermal comfort environmental chamber study of older and younger people

A thermal comfort environmental chamber study of older and younger people

(2019)

We investigated whether or not, when exposed to the same conditions, older people (those aged 65 and over) had different thermal sensations, comfort, acceptability and preferences from their younger counterparts. The study was conducted in a thermal comfort environmental chamber, involving 22 older (average 69.7 years old) and 20 younger (29.6 years old) subjects, exposed to four test conditions between slightly cool and slightly warm. Subjective thermal comfort perceptions for local body parts and whole-body were surveyed. Skin temperatures were measured at four body locations: neck, right scapula, left hand, and right shin. We also investigated the correlation between the frailty level of the subjects and their thermal comfort levels. The study found no significant difference between the thermal sensation, comfort and acceptability of older and younger subjects. We also found no correlation between subjects’ frailty level and their thermal sensation, comfort, acceptability and preference but we did not have many frail subjects. In both older and younger subjects, the hand’s skin temperature had a significant correlation with the local and overall thermal sensation.

Cover page of Analysis of the accuracy on PMV – PPD model using the ASHRAE Global Thermal Comfort Database II

Analysis of the accuracy on PMV – PPD model using the ASHRAE Global Thermal Comfort Database II

(2019)

The predicted mean vote (PMV) and predicted percentage of dissatisfied (PPD) are the most widely used thermal comfort indices. Yet, their performance remains a contested topic. The ASHRAE Global Thermal Comfort Database II, the largest of its kind, was used to evaluate the prediction accuracy of the PMV/PPD model. We focused on: (i) the accuracy of PMV in predicting both observed thermal sensation (OTS) or observed mean vote (OMV) and (ii) comparing the PMV-PPD relationship with binned OTS – observed percentage of unacceptability (OPU). The accuracy of PMV in predicting OTS was only 34%, meaning that the thermal sensation is incorrectly predicted two out of three times. PMV had a mean absolute error of one unit on the thermal sensation scale and its accuracy decreased towards the ends of the thermal sensation scale. The accuracy of PMV was similarly low for air-conditioned, naturally ventilated and mixed-mode buildings. In addition, the PPD was not able to predict the dissatisfaction rate. If the PMV model would perfectly predict thermal sensation, then PPD accuracy is higher close to neutrality but it would overestimate dissatisfaction by approximately 15-25% outside of it. Furthermore, PMV-PPD accuracy varied strongly between ventilation strategies, building types and climate groups. These findings demonstrate the low prediction accuracy of the PMV–PPD model, indicating the need to develop high prediction accuracy thermal comfort models. For demonstration, we developed a simple thermal prediction model just based on air temperature and its accuracy, for this database, was higher than PMV.

Cover page of Thermal comfort under radiant asymmetries of floor cooling system in 2 h and 8 h exposure durations

Thermal comfort under radiant asymmetries of floor cooling system in 2 h and 8 h exposure durations

(2019)

Radiant heating and cooling systems inherently exhibit radiant asymmetries. Although many researchers have investigated the thermal comfort effects of asymmetric radiant environments, the exposure duration has not been emphasized, especially under floor heating and cooling scenarios. In this study, we conducted a series of tests in a climate chamber with floor cooling radiant asymmetries with human participants to investigate their thermal comfort effects from short-term (2 h) and long-term (8 h) exposure perspectives. The 2 h exposure test indicates that the floor cooling systems cause discomfort complaints more easily than other radiant systems such as ceiling heating/cooling because of its stronger cooling effects on the lower body parts. The cold floor resulted in significantly colder local thermal sensations and lower local skin temperatures in the foot, calf, and thigh areas. The comparison between the 2 h and 8 h exposures suggests that exposure duration affects both the subjective and physiological thermal comfort responses significantly. Further, 2.5~4 hours are required for the foot and calf temperatures to stabilize in radiant floor cooling asymmetry cases. In accordance with these laboratory tests, we proposed two radiant asymmetry-satisfaction curves and equations for the floor cooling system with consideration of exposure duration. The calculated temperature limits for typical floor cooling room are >18.5 oC at a 2 h exposure and >20.5 oC at an 8 h exposure. These curves and temperature limits can serve as a reference for future guidelines for floor cooling system design and operation.

Cover page of Measurement of airflow pattern induced by ceiling fan with quad-view colour sequence particle streak velocimetry

Measurement of airflow pattern induced by ceiling fan with quad-view colour sequence particle streak velocimetry

(2019)

Ceiling fans have been widely used for a long time as an effective cooling equipment to create sustainable indoor environment. However, it is rather difficult for the current measuring techniques to capture such a complicated airflow field in a whole-room scale. In this study, a novel large-scale airflow measurement technology, quad-view colour sequence particle streak velocimetry (CSPSV), is developed and applied to measure the airflow induced by a ceiling fan in a 4 m × 2.5 m × 3 m chamber. Four cameras were used in the new method, two at the higher position measuring airflow near the ceiling while the other two at the lower position measuring airflow near the floor, to capture the room-scale flow field. After reconstructing the vectors from each camera pair, the airflow vectors are merged to fill the blind zone near the ceiling and floor. Based on the three-dimensional three-component vector field measurement data, the averaged velocity vector, turbulence intensity, and vorticity were calculated and the airflow patterns were analyzed. The results indicate that the quad-view CSPSV method provides a more comprehensive measurement in room-size complex air movements such as ceiling fan airflow. Six pattern zones can be identified for a typical ceiling-fan-induced airflow. The flow under ceiling fan swirls along its path with the same rotation direction of fan blades with the core shrinking gradually and becoming diluted by the surrounding air. This study provides a new velocimetry method for room-sized complex airflow and a better understanding of ceiling fan airflow pattern, which is helpful to the new concept of integrating ceiling fan with air conditioning system.

Cover page of The Squeaky Wheel: Machine learning for anomaly detection in subjective thermal comfort votes

The Squeaky Wheel: Machine learning for anomaly detection in subjective thermal comfort votes

(2019)

Anomalous patterns in subjective votes can bias thermal comfort models built using data-driven approaches. A stochastic-based two-step framework to detect outliers in subjective thermal comfort data is proposed to address this problem. The anomaly detection technique involves defining similar conditions using a k-Nearest Neighbor (KNN) method and then quantifying the dissimilarity of the occupants’ votes from their peers under similar thermal conditions through a Multivariate Gaussian approach. This framework is used to detect outliers in the ASHRAE Global Thermal Comfort Database I & II. The resulting anomaly-free dataset produced more robust comfort models avoiding dubious predictions. The proposed method has been proven to effectively distinguish outliers from inter-individual variabilities in thermal demand. The proposed anomaly detection framework could easily be applied to other applications with different variables or subjective metrics. Such a tool holds great promise for use in the development of occupancy responsive controls for automated building HVAC systems.

Cover page of Personal CO2 bubble: Context-dependent variations and wearable sensors usability

Personal CO2 bubble: Context-dependent variations and wearable sensors usability

(2019)

High CO2 concentration in inhaled air has been shown to negatively impact work performance and increase acute health symptoms. As respiratory CO2 is constantly exhaled, it may not dissipate in surrounding air in absence of adequate air movement and is instead re-inhaled into the airways (breathing in a CO2-rich bubble). In this study, we explored the impacts of context-dependent factors such as office activities, desk settings, and personal differences on the inhalation zone CO2 concentration and on concentrations at a below-neck wearable sensor. While all factors were found to significantly impact measurements at both measuring points, desk settings (empty desk, desk with a fan, desk with laptop, desk with monitor) was found to be the most dominant factor. Presence of a small portable desk fan was found to significantly reduce the CO2 concentration. On average, we observed a 177 ppm reduction in CO2 concentration when using a fan, which is 25 ppm higher than the background CO2 measurement (650 ppm). Among 41 test subjects, we found distinct relationships between the inhalation zone CO2 concentration and the wearable sensor measurements and, by applying a hierarchical clustering algorithm, we found 4 clusters of relationships. While below-neck wearable sensors could be used as an exact measure of inhalation of CO2 concentration for 29% of the subjects, we identified a boundary point (917 ppm) separating high and low inhalation zone CO2 concentration measurements.