High-Accuracy Dynamic Weighing Framework for Checkweighers Using IRVMD and MKSVR

Conveyor belt-type checkweighers play a crucial role in ensuring product quality, improving production efficiency, and reducing operating costs. However, due to the combined effects of inherent system nonlinearities and multisource interference, achieving high-speed and high-accuracy dynamic weighing continues to pose significant challenges. To address this issue, this article proposes a dynamic weighing framework that integrates improved recursive variational mode decomposition (IRVMD) with multikernel support vector regression (MKSVR). First, IRVMD employs a single-mode recursive extraction mechanism and incorporates an adaptive stopping criterion constructed based on the characteristics of the load cell signals, enabling adaptive suppression of multisource interference. Then, building upon the denoised signals, MKSVR is employed to establish a nonlinear mapping between the weighing-segment signals and the true load mass in a more expressive composite-kernel space, thereby enhancing the model’s capability to characterize nonlinear dynamic weighing relationships. Based on the developed checkweigher experimental platform, dynamic weighing experiments under multiple load levels and conveyor belt speeds are conducted. The experimental results demonstrate that, under the considered operating conditions, IRVMD-MKSVR fulfills the accuracy requirements specified in OIML R 51 for class XIII(0.02) checkweighers, thereby validating its effectiveness and feasibility for high-speed and high-accuracy dynamic weighing.