Research on the Rapid Detection of Formaldehyde Emission From Wood-Based Panels Based on the AMSHKELM

To enable quick and accurate detection of formaldehyde emissions from wood-based panels and to address the complexities of the full-scale chamber method and the inaccuracy of sensor-based methods, a modified formaldehyde emission model based on AdaBoost-MSBKA-SVMD-HKELM (AMSHKELM) is proposed. This model utilizes sensor data, which provides ease of use and rapid detection, as input and formaldehyde emission data measured by the full-scale chamber method as output. Initially, multiple strategies are employed to address the inherent limitations of the black-winged kite algorithm, which struggles to effectively find optimal solutions and is susceptible to getting trapped in local optima. The multi-strategy improved black-winged kite algorithm then optimizes key parameters of the successive variational mode decomposition (SVMD) and hybrid kernel extreme learning machine (HKELM). In addition, adaptive boosting is introduced to further improve the accuracy and robustness of the model. Then, the AMSHKELM formaldehyde emission modified deviation model is constructed to fit the decomposed subsequence. Moreover, an adaptive bandwidth kernel density estimation combined with the AMSHKELM is developed to construct an interval prediction model. Experimental results indicate that the AMSHKELM model achieves the coefficient of determination of up to 0.9767 and the root mean square error of 2.7141e-03, demonstrating higher fitting accuracy and stronger robustness compared to various other models. Additionally, the interval prediction model performs superiorly. This model can combine interval prediction information to effectively and comprehensively assess the pass rate of test samples, providing a fast and reliable solution for formaldehyde emission testing of wood-based panels.