Study on peak hours, ventilation, and resident activities towards indoor air quality on PM2.5 in Surabaya

Several parameters contribute to the standard level of indoor air quality (IAQ), such as PM2.5 concentration, which is the primary parameter in this study. Several variables influence IAQ, such as infiltration and resident activities of outside air pollution through ventilation. Peak-hour traffic congestion can significantly raise outdoor air pollution. This study was aimed at finding out more about the influence of peak hours, the relationship between ventilation, resident activities, and indoor air quality (IAQ) in houses near the main road, as well as suggestions for improving IAQ. AirVisual Pro was the tool used to record PM2.5 concentration, temperature, and humidity; Kestrel 5500 was used to assess wind direction; and questionnaires were completed by residents. SPSS software is utilized for data analysis including the Pearson correlation coefficient, multicollinearity, autocorrelation, and multiple linear regression (MLR). The average indoor and outdoor PM2.5 concentrations in all units during peak and non-peak hours meet the quality standards on weekday and weekend. According to the correlation analysis, indoor and outdoor PM2.5 concentrations had no significant correlation regardless of whether the window is open or closed. Temperature (-1.473), humidity (-0.033), the number of furniture (3.660), fan usage (-8.005), cooking activities (3.755), cleaning activities (14.940), smoking behavior (4.545), and peak hours (6.896) associated with PM2.5 concentration, according to the MLR analysis. Increase fan use, no smoking inside the house, and less furniture are advised to improve IAQ.


Introduction
Air quality is a major risk factor for humans, whether it occurs during outdoor or indoor exposure.However, throughout the course of a lifetime, oxygen in enclosed area such as inside houses, is consumed by human around 50-90%, making indoor air a dominant exposure to risk for humans [1].According to the World Health Organization (WHO), household air pollution was responsible for 3.8 million deaths in 2016, contributing to 7.7% of the world's mortality [2].Indoor air quality (IAQ) can be affected by various parameters listed in Government Regulation No. 22, such as dust particulates which include TSP, PM10, and PM2.5 [3].Fine particles (PM2.5)can settle on a variety of indoor surfaces [4].
There are numerous factors at home that contribute to the IAQ, including the home's structure, cooking, cleaning, lighting, and tobacco smoking by residents.Furthermore, outside air pollution infiltration is a major source of indoor air pollution.Air movement in and out of dwellings caused by pressure and temperature changes through unintentional openings is referred to as infiltration.Outdoor 1263 (2023) 012046 IOP Publishing doi:10.1088/1755-1315/1263/1/012046 2 air pollution can enter the home via open windows and doors, as well as supply air ventilation systems.[2,5].However, increased traffic congestion and length during peak hours can significantly increase pollutant emissions and deteriorate air quality, particularly near large roadways [6].Therefore, this study aimed: to evaluate the obtained indoor PM2.5 concentration in houses near the main road during peak hours and non-peak hours compared the air quality standard; to identify the correlation of activities, ventilation, and peak hours with indoor air quality in houses near the main road regarding the PM2.5 concentration; and the recommendations for improving IAQ.

Research preparation and implementation
Research preparation is needed before conducting the research.These include the determination of peak hours, sampling location, questionnaire preparation, air visual pro preparation, and kestrel 5500 preparation.This research selects peak hours from 06:00 -09:00 AM and 04:00 -07:00 PM.The rest of the hours are classified as non-peak hours.The research uses a purposive sampling method which takes sampling according to the predetermined criteria.The criteria include: the type of the road of collector roads; minimum of five residents living in the house; the maximum distance between the house and the main road is 400 m; there are two floors in the house; areas of the building of each sampling house are like one another; the distance to the industrial area is at least 3 km from the sampling location.The sampling location was taken in 5 units of houses near the main road in Surabaya.The questionnaire is formed with structured questions addressed to residents.The sampling point of AirVisual Pro occurs indoors in the family room and outdoors on the terrace.The Kestrel 5500 was placed on the terrace by attaching it to a tripod.
AirVisual Pro measured indoor and outdoor PM2.5 concentration, humidity, and temperature.Kestrel 5500 measured wind speed and wind direction.Sampling is carried out for 2 x 24 hours, one working day, and one on the weekend, measured every 10 minutes using AirVisual Pro and Kestrel 5500.The data was then separated between peak hours and non-peak hours.The questionnaire was filled out by residents during the sampling measurement.

Data processing
Data processing is conducted to process the collected data (PM2.5 concentration measurement and questionnaire filled out) into usable information which supports the research objectives.This includes data examination, data input, and data tabulation using SPSS.The results of the data obtained from the measurement of PM2.5 concentration and the questionnaire obtained several independent variables (X) and the dependent variable (Y).The measured variables are shown in Table 1.

Evaluation of peak and non-peak hours PM2.5 with air quality standard
The obtained PM2.5 concentration has varied temperatures during measurements, which needs to be converted in a normal atmosphere (at 25°C).After the obtained PM2.5 concentration is converted at the normal temperature, the data is averaged per day for both on weekday and weekend.Then, each unit and all five units are compared to the PM2.5 quality standard, which is 55 μg/m 3 .The result comparison of peak and non-peak hours is shown in Table 2.The result shows that during peak hours, only unit 5 on the weekday exceeds the PM2.5 quality standards of 55 μg/m 3 in indoor and outdoor air.The PM2.5 concentration in indoor air is 55.49 μg/m 3 and for outdoor air is 57.14 μg/m 3 .Unit 1, unit 2, unit 3, and unit 4 meet the requirement of PM2.5 quality standards for both peak and non-peak hours.On weekdays and weekends, the average indoor and outdoor PM2.5 concentrations in all units met the quality criteria during peak hours and non-peak hours.Indoor PM2.5 concentration is slightly higher than outdoor PM2.5 concentration during peak hours.Meanwhile, during non-peak hours, the outdoor PM2.5 concentration is slightly higher than the indoor PM2.5 concentration.The increase in outdoor PM2.5 concentration is paralleled by an increase in indoor PM2.5 concentration.[7].Smoking, vehicle exhaust or ambient air infiltration into indoor areas, cooking, road dust, building materials, and cleaner sprays have all been linked to higher indoor PM2.5 concentration compared to outdoor concentration.Low indoor PM2.5 levels could be described as minimal human activities in indoor spaces [8].

Pearson correlation coefficient
The purpose of the Pearson correlation coefficient is needed for peak hours and non-peak hours to determine the correlation between PM2.5 concentration and the window opening (open and closed).The result calculation of all five units is shown in Table 3 and Table 4.  Based on Table 3 and Table 4 during peak hours and non-peak hours, the Pearson correlation indicated that all five units with open and closed windows have a positive correlation because the Pearson correlation ranges between 0 < 1.The increase of outdoor PM2.5 concentration increased indoor PM2.5 concentration.This indicated that there was no significant relationship between indoor and outdoor PM2.5 concentration regardless of whether the window is open or closed.Further studies need to be conducted to inspect other sources of air exchange, such as through air holes.Indoor PM2.5 sources include smoking, cooking, heating, the use of incense, candles, and insecticides, while the main sources of coarse particles are cleaning, housework, the presence of pets, and human movement.Outdoor air is also a significant source of PM2.5 in roadside residences with natural ventilation [9].

Multicollinearity test
The measured variable can be seen in Table 1 Window opening (X3) is removed from the regression model due to significance in Pearson correlation with PM2.5 concentration.The multicollinearity test determines whether a multiple regression model was having a high correlation between two or more predictor variables.If Variance Inflation Factors (VIF) > 10 and Tolerance > 0.1, there is no multicollinearity occurs.The result of the multicollinearity test during weekday and weekend can be seen in Table 5.Based on Table 5 there is no multicollinearity occurs on the multiple linear regression from X1, X2 and X4 -X9.This is due to no value of VIF greater than 10 and the Tolerance is greater than 0.1.

Autocorrelation test
After conducting the multicollinearity test and there is no multicollinearity occurs, the following test is autocorrelation.The autocorrelation test is used to detect if errors on a time-sorted sequence of data (time series) have a linear connection.One of the autocorrelation tests is the Durbin-Watson test which can determine whether there is an autocorrelation.If the Durbin Waston value (d) is between dU and 4 -dU (dU d 4 -dU), no autocorrelation exists.Based on the calculation, the result of the Durbin-Watson value (d) is 1.974.To prove there is no autocorrelation occurred, the d value must be greater than dU and smaller than 4 -dU, which can be seen as follows.
• X2 (Humidity) The regression shows the humidity is inversely proportional to PM2.5 concentration, where an increase of 1% can reduce the PM2.5 concentration by 0.033 μg/m 3 .When humidity was low, PM2.5 concentration increased; however, when humidity was high enough, PM2.5 concentration went down.This decrease is due to particles accumulating mass as a result of moisture, which eventually leads to the dry deposition of particles on the ground [11].
• X4 (Number of furniture) The MLR analysis shows that the number of furniture is directly proportional to PM2.5 concentration, where adding 1 unit of furniture in the house can increase the PM2.5 concentration by 3.660 μg/m 3 .The increase in particles corresponded with an increase in the number of furniture pieces, which could have been due to particle attachment to the furniture.The greater the number of pieces of furniture, the more dust collects on them [12].
• X5 (Fan's usage) The fans are generally located in the center of the room for better air circulation [13].The MLR shows the duration of fan usage is inversely proportional to PM2.5 concentration, where longer fan utilization can decrease the PM2.5 concentration by 8.005 μg/m 3 .

• X6 (Cooking activities)
The equation shows the duration of cooking activities is directly proportional to PM2.5 concentration, where the longer cooking duration can increase the PM2.5 concentration by 3.755 μg/m 3 .Cooking type can cause an increase in the concentration of PM2.5 indoors [14].
• X7 (Cleaning activities) The MLR shows that the duration of cleaning activities is directly proportional to PM2.5 concentration, where the longer the cleaning activities are conducted, can increase the PM2.5 concentration by 14.940 μg/m 3 .Indoor human activities, including cleaning, can raise the concentration of PM2.5 in the air to a higher level [15].

• X8 (Smoking behavior)
The regression shows smoking is directly proportional to PM2.5 concentration, where increased smoking activities in the house can raised the PM2.5 concentration by 4.545 μg/m 3 .Even in short-term measurements, cooking and smoking significantly increase PM2.5 concentration in other rooms including living rooms [16].
• X9 (Peak hours) The equation shows peak hours are directly proportional to PM2.5 concentration where during peak hours, PM2.5 concentration approximately increases by 6.896 μg/m 3 .On the contrary, during non-peak hours, the PM2.5 concentration approximately decreases by 6.896 μg/m 3 .There are more vehicles on the road during rush hour, which may result in greater PM2.5 exposure [17].
The R Square of the regression is 0.245, which is below 0.6.This means that the existing variables can only explain 24.5% of the correlation on the PM2.5 concentration.Meanwhile, 75.5% has not been explained.Therefore, it is necessary to add other variables for further study to increase R Square to a minimum of 60%, representing the correlation on the PM2.5 concentration.

Recommendation for improving IAQ
Several recommendations can improve indoor air quality in homes near the main road based on prior factors as indicators that influence PM2.5 concentration.The recommendations are: increase the usage of fans for better indoor air movement; no smoking inside the house; and minimize the amount of furniture or remove unused furniture inside of the house.

Conclusion
A conclusion can be established based on the data analysis and discussion.On weekdays and weekends, the average indoor and outdoor PM2.5 concentration of all units during peak and non-peak hours met the quality criteria.On weekdays and weekends, the average indoor and outdoor PM2.5 concentrations in all units met the quality criteria during peak hours and non-peak hours.The indoor PM2.5 concentration is slightly greater than the outdoor PM2.5 concentration during peak hours.Meanwhile, the outdoor PM2.5 concentration is slightly greater than the indoor PM2.5 concentration during non-peak hours.According to the correlation analysis, indoor and outdoor PM2.5 concentrations had no significant correlation regardless of whether the window is open or closed.Temperature (-1.473), humidity (-0.033), number of furniture (3.660), fan usage (-8.005), cooking activities (3.755), cleaning activities (14.940), smoking behavior (4.545), and peak hours (6.896) all associated with PM2.5 concentration, according to the MLR analysis.Increased fan use, do not smoke inside the house, and less furniture are advised to improve IAQ.

Table 3 .
Pearson correlation all units during peak hours.

Table 4 .
Pearson correlation all units during non-peak hours.

Table 6 .
Result calculation multiple linear regression analysis.