A modified approach to metabolic rate determination for thermal comfort prediction during high metabolic rate activities

Highlights

• PMV index lacks accuracy during high metabolic activities.

• Modified approach offers improvement to comfort predictions during high metabolic activities.

• Current use of the PMV index to evaluate comfort in sport facilities may be unreliable.

• Existing guidance for PMV use may require further limitation.

Abstract

Environmental conditions in buildings are linked to the physical and mental wellbeing of occupants. Thus, it follows that the internal environment affects human performance and user experience during sport and activity. There are several indices that are used to evaluate occupant thermal comfort, the Predicted Mean Vote (PMV) index being the metric most commonly used. PMV is designed to evaluate comfort for sedentary occupants with low metabolic rates; however, PMV has also been used to evaluate comfort for individuals engaged in high metabolic rate activities, such as those common in sport facilities.

This paper investigates the implication of using PMV to evaluate thermal comfort in sport facilities using empirical data recorded over 24 months in a multi-purpose sports hall in the North of England. Data are used to develop and propose methodological modifications to improve the standard PMV model prediction to account for occupants having higher metabolic rates.

The paper evaluates the use of metabolic rate data from different sources including the Compendium of Physical Activities and quantifies the impact that the metabolic weighting approach has on predicted comfort. Finally, a novel method is proposed to modify PMV for use where occupants have high metabolic rates.

Despite the improvements made, the findings suggest that even a modified PMV may not be able to accurately evaluate the thermal comfort of people engaged in non-sedentary activity, recommending that use of the PMV index is restricted to activities with metabolic rates <2 MET.

Introduction

The sport and recreation sector has an important role in society [1], with sport and sport-related activity in the top 15 industry sectors in England, generating over £20bn annually and supporting over 400,000 jobs [2]. Sport and active lifestyles are also a key factor in wider societal issues, with physical inactivity a leading risk factor for global mortality [3,4]. The essential nature of sporting activity is recognised by UK government policy [5]. Studies have shown that sport participation rates are influenced by environmental conditions [6].

The internal environment of buildings has been shown to have a significant impact on the wellbeing and satisfaction of occupants [7] and this is particularly relevant in sport facilities, with common complaints resulting from poor conditioning including stuffiness and overheating in summer and cold draughts in winter. Inadequate lighting may cause visual discomfort [8] and insufficient ventilation may lead to poor air quality, moisture accumulation and odour [9]. At a critical level, the local environment can negatively influence the performance of an individual. Athletic performance becomes inhibited at extreme hot or cold temperatures [[10], [11], [12], [13], [14], [15]], meaning competitive athletic events will be affected by poor quality environmental conditions. In extreme cases this can lead to an enhanced risk of injury [16]. Similarly, environmental conditions and air quality can negatively affect cognitive performance [[17], [18], [19], [20], [21], [22], [23]], which is significant in mixed-use buildings where examinations or teaching may take place. A combination of undesirable environmental factors may lead to individual dissatisfaction and ultimately a reluctance to engage in sporting activity [6].

One metric by which environmental conditions may be assessed is thermal comfort, which is defined as the state of mind that expresses satisfaction with the thermal environment [24]. As thermal comfort is desirable, methods to measure and predict the thermal comfort of occupants have been developed. The Predicted Mean Vote (PMV) thermal comfort index developed by Fanger [25] is the most widely used of these, and has been accepted into an international standard [26].

The popularity of the PMV index has naturally attracted investigation into its validity, particularly in real-world contexts. Although international standards governing its use state that “although developed for the work environment, it is applicable to other kinds of environment as well” [26], research suggests that the PMV index is a poor predictor of actual thermal sensation across a wide range of contexts [[27], [28], [29], [30], [31], [32]]. Specifically, contextual factors found to influence predictive accuracy include: building type [29,33], local climate [34], cultural and regional difference [35], age [36], activity [37], gender [38] and the characteristics of the occupied space [[39], [40], [41]]. Despite this criticism, PMV remains the most used thermal comfort index [31]. Whilst investigation into model validity is important, identification of poor performance is unhelpful without a corresponding attempt to improve predictive performance [29].

The PMV index has been developed to evaluate comfort under the assumption that the occupants are undertaking low up to moderate work, although as noted previously, standards governing its use indicate that it is applicable to other kinds of environment [26]. Its guiding principle is that maximum comfort corresponds with neutral thermal sensation. This assumption has been challenged when considering contexts outside those used in the development of the PMV index, including during exercise [[42], [43], [44], [45], [46], [47], [48], [49], [50]]. Despite this, the PMV index has been applied to many situations where energetic leisure activities are likely to occur [46,47,[51], [52], [53], [54], [55], [56], [57]], with several authors attempting to evaluate thermal comfort in sport facilities using a mixture of modelling [[58], [59], [60], [61], [62]], laboratory [43,63,64] and field [47,54,65] methods. Existing studies are often limited in their scope or method [66], restricting their analysis to a subset of occupants, artificially limiting input values or applying the PMV index without providing detail on inputs and assumptions. This is relevant when considering that the PMV index places an upper limit on metabolic rate [26] that is routinely exceeded during exercise [67].

Inputs to the PMV index include environmental and personal parameters relevant to the human thermal balance, namely: air temperature; mean radiant temperature; humidity; air velocity; metabolic rate and clothing insulation. In field studies, environmental parameters may be measured directly [68], although this is often complex in practice [69] and the role of measurement accuracy is significant in PMV accuracy [70]. Measuring personal parameters is more complex, typically requiring laboratory-grade equipment and processes that involve thermal manikins [71] and calorimetry [72]. For this reason, in field studies the personal parameters are often assumed based on observation and comparison with tabulated reference values [32]. However, the PMV index is highly sensitive to personal parameters and the sensitivity of PMV to these assumptions is not habitually explored [73,74].

The implications of using the PMV index in sport facilities, despite recommendations in the standard for its limited use [24,26], requires further investigation. Additionally, existing PMV validation studies rely on databases of default values for predominantly sedentary (lower) metabolic rates [29,75]. Thus, there is a need to understand how PMV accuracy is affected when a greater variation of metabolic rates are used in the model, to accommodate a broader range of occupant activities and, specifically, those that take place in sport facilities. This paper will concentrate on the metabolic rate, due to its heightened relevance in sport facilities.

The metabolic rate for an activity is often presented as a MET value, which is a ratio between the energy intensity of an activity and a reference metabolic rate. For thermal comfort calculations, 1 MET = 58.2 W/m². International Standards for thermal comfort calculation and metabolic rate determination [26,72] contain the methods for measuring metabolic rate and the reference tables of metabolic rates for common activities. Determining metabolic rate by observing an activity and referring to the corresponding tabulated reference value is the prevalent approach in field studies, however this standard approach has been observed to lack accuracy [76]. One source of error is the requirement for researchers to use their own interpretation to select a comparable metabolic rate from the reference values where a specific activity is not represented [24]. More accurate approaches to establish metabolic rate, for example using wearable technology such as heart rate [73,77] and blood pressure [78] monitors, have been developed. However, the requirement for measurement equipment introduces the need for increased subject participation and has not been trialled in thermal comfort field tests comprising hundreds of participants. Consistent with the findings of Luo et al. [79] there is a need for improvement of methods using metabolic rate reference values.

The aim of this paper is to evaluate the implications of using the PMV index to evaluate comfort in sport facilities. As the previously acknowledged guidance for its use [26] suggests PMV predictions may be imperfect for individuals engaged in energetic activity, this paper explores the potential to improve predictive accuracy though investigation of the following two issues relating to the metabolic rate used during PMV calculation:

i) Using an alternative reference database of metabolic rates: The metabolic rates contained in the thermal comfort standards [26,72] are predominantly for sedentary or occupational activities rather than sport activities. This presents a possible source of error when applied to a sport facility; in the absence of sport specific values, the researcher is required to choose an activity deemed metabolically comparable for each sport that may take place in a sports hall. The Compendium of Physical Activities (CPA) [80] is an alternative and larger database of metabolic rates than that provided in the ISO standard [72], containing different activities that can be used where the MET data are not available in the ISO database. The MET values that make up the CPA are taken from robust laboratory classification studies, and are considered to be a valuable resource in studies of physiology and sport science [81]. Using the CPA for thermal comfort calculations may be problematic, given that MET values are presented using different units (W/kg as opposed to W/m²), and so directly using MET values from the CPA may cause methodological uncertainty. This paper will investigate the impact of this difference on the PMV for the case study building.

ii) Varying MET weighting approach: When occupants undertake multiple activities with varying metabolic rates consecutively while in a building, researchers often simply use time-weighted average of the metabolic rates to input into PMV calculations [26]. However, this has been shown to cause overestimations in PMV, and not adequately account for the transient effects of metabolic activity and thermal sensation [64,82,83]. Furthermore, the insulating effect of clothing (clo) is reduced when the body moves, through increased relative air velocity compressing fabrics and by the ‘pumping’ effect as motion increases air exchange between the body surface and the external environment via cuffs, collars and sleeve openings [71,84]. As metabolic rate influences insulating performance, applying an average metabolic rate may incorrectly estimate the actual reduction in clothing insulation. Thus, applying a time-weighted average metabolic rate to work cycles that include different activities may not be appropriate, and risks obscuring dynamic effects. To overcome this issue, rather than calculate a single PMV using the average MET rates this paper will evaluate the impact of calculating a time-weighted average PMV for the activity period, based on discreet PMV values generated for each activity phase [47].

Using a case study of a Multi-Purpose Sports Hall (MPSH), the aggregated discrepancy between the PMV and the occupant's actual thermal sensation vote (TSV) as reported via occupant survey was calculated. The effectiveness of novel modifications to the PMV method to reduce predictive bias and improve accuracy were then evaluated, incorporating an alternative database of metabolic rates, the CPA.