Combined Prediction of PM10 Concentration at Smart Construction Sites Based on Quadratic Mode Decomposition and Deep Learning

Ming Li, Xin Li (), Kaikai Kang and Qiang Li
Additional contact information
Ming Li: College of Business, Hohai University, Nanjing 211100, China
Xin Li: College of Business, Hohai University, Nanjing 211100, China
Kaikai Kang: College of Business, Hohai University, Nanjing 211100, China
Qiang Li: Zhongbo Information Technology Research Institute Co., Ltd., Nanjing 210001, China

Sustainability, 2025, vol. 17, issue 2, 1-22

Abstract: The accurate prediction of PM10 concentrations at smart construction sites is crucial for improving urban air quality, protecting public health, and advancing sustainable development in the construction industry. PM10 concentrations at construction sites are influenced by the interaction of construction intensity and environmental meteorological factors, resulting in nonlinear and volatile data. To improve prediction accuracy, this paper presents a two-stage mode decomposition method that integrates Complementary Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Variational Mode Decomposition (VMD). This method is combined with a Bidirectional Long Short-Term Memory (BiLSTM) neural network, optimized using the Sparrow Search Algorithm (SSA), to establish a hybrid model for forecasting PM10 concentrations at construction sites. Initially, CEEMDAN decomposes the original sequence into several Intrinsic Mode Functions (IMFs). The sample entropy of each component is then calculated, and K-means clustering is used to group them. VMD is applied to further decompose the high-frequency components obtained after clustering. SSA is then employed to optimize the parameters of the BiLSTM network, which models all the components with the optimized predictive model. The predicted values of all components are aggregated to generate the final forecast. Real-time monitoring data from Construction Site A in Nanjing are used for case study validation. The empirical results demonstrate that the proposed hybrid prediction model outperforms comparison models on all evaluation metrics, offering a scientific foundation for sustainable and automated dust reduction decision-making at smart construction sites, thereby facilitating the shift toward greener, smarter, and more digitized construction practices.

Keywords: PM10 concentration prediction; smart construction site; quadratic mode decomposition; deep learning; combinatorial modeling (search for similar items in EconPapers)
JEL-codes: O13 Q Q0 Q2 Q3 Q5 Q56 (search for similar items in EconPapers)
Date: 2025
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2071-1050/17/2/616/pdf (application/pdf)
https://www.mdpi.com/2071-1050/17/2/616/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jsusta:v:17:y:2025:i:2:p:616-:d:1567108

Access Statistics for this article

Sustainability is currently edited by Ms. Alexandra Wu

More articles in Sustainability from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().