Multi-component rice, as a kind of prepared food mainly composed of the carbohydrates and proteins from rice, is popular among consumers due to the characteristics of balanced ratio that is rich in nutrition (Montero, Garrido, Gallardo, Tang, & Ross, 2021; Tang, Hong, Inanoglu, & Liu, 2018). To meet the needs of consumers in different regions of the world (Huang, Zhang, & Fang, 2023), many types and flavors of multi-component rice products have been produced by several food preparation industries around the Chinese market, highlighting their importance in the instant food industry. Although multi-component foods have the advantages of convenience, long shelf life, and stable flavor compared to traditional cooked dishes, most of them still require to be reheated before consumption (Huang et al., 2023; Lin & Li, 2023). This step of processing not only lead to food safety issues but also to sensory quality (Huang, Zhang, & Bhandari, 2019). Heretofore, the mainstream reheating methods are based on conventional heating (water bath and steam) (Dutta & Mahanta, 2014; Murphy, Duncan, Driscoll, Marcy, & Beard, 2003; Zhang et al., 2018), but microwave heating (915 and 2450 MHz) (Cai et al., 2018; Thanakkasaranee, Sadeghi, & Seo, 2023) appears to be the more popular method. The traditional and microwave heating methods could achieve relatively uniform temperature distribution for thin-layer products (Wang, Sun, Jiang, Xu, & Xia, 2024). However, the corresponding drawbacks of long time-consuming and non-uniform heating are often found in large-size materials due to the heat transfer mechanisms and low penetration depth, respectively (Tian et al., 2024). Due to the increasing variety of usage and ingredient ratios of multi-component rice found in the product information of different prepared food brands mentioned in Huang et al. (2023), it is necessary to introduce a novel heating technique with a wider range of applicability to solve this problem.

Radio frequency (RF) energy, as a dielectric heating technique (Miran & Palazoglu, 2023), causes dipolar rotation and ionic conduction within the material through high-frequency alternating electric fields (13.56, 27.12, and 40.48 MHz), resulting in the internal friction and heat generation (Guo, Mujumdar, & Zhang, 2019). Although this heating principle is similar to that of the microwave, RF treatment is more suitable for volumetric heating of large-size samples by virtue of its lower frequency and deeper penetration depth (Abea, Gou, Guardia, Banon, & Munoz, 2021; Zhang et al., 2024). Due to the difference in dielectric properties between the material and the external air medium during RF heating, the electric field lines in the transition region pass densely through the material edges and corners, leading to the local uneven temperature distributions called the edge-effect (Yu, Shrestha, & Baik, 2016). Correspondingly, several studies have proposed some methods to improve the RF heating uniformity by changing specific variables applicable to the relevant treated samples, such as electrode gap (Erdogdu, Altin, Marra, & Bedane, 2017), geometric shape (Li et al., 2018), spatial location (Li, Wang, Wang, & Ling, 2022), auxiliary materials (Villa-Rojas, Zhu, Marks, & Tang, 2017), and dielectric properties (Dong et al., 2021), etc.

Compared to studies focusing on the effect of a single specific variable on the improvement of RF heating performance, the exploration of the effects of multiple categories of influential factors in a uniform manner need to be more comprehensive and systematic. The results of these studies showed the relative influence of each parameter (Alfaifi et al., 2014) and also helped to identify the direction for subsequent improvement studies. In addition, due to the limited information provided by experiments on this complex process, such studies have mainly analyzed the RF heating process through finite element simulation (Huang, Marra, Subbiah, & Wang, 2018). Currently, many studies are still mostly centered on single-component materials. But the corresponding studies on multi-component food products, which involve more numbers of factors, such as the proportion between various parts in multi-component foods and the structural shape of internal materials, are relatively scarce. On the basis of exploring influence of RF system parameters and material properties, it is also necessary to involve factors, such as sample dosage and internal component shape, which could affect heating efficiency and uniformity.

A computer simulation model for multi-component rice heated by a small-scale 50 Ω RF heating system has been established and verified in the previous study (Tian et al., 2025). The IR30-ER40 sample scheme with relatively homogeneous heating has been identified by regularly varying the sizes of the corner radius of the internal and external components. To expand and promote the processing conditions and product types, further systematic exploration about the effect of the system parameter, sample structure, and material properties on the RF heating performance should be conducted on the basis of the initially established model (Tian et al., 2025). Therefore, combining the setting methods for parameters in previous studies (Huang, Zhu, & Wang, 2015; Li et al., 2024) and the present state for diversity of multi-component rice (Huang et al., 2023), the simulation model of multi-component rice for RF processing is necessary to be further validated and improved for conducting a systematic study of parameter optimization.

The main objectives of this study were to (1) further validate an established computer model based on a small-scale 50 Ω RF system through the selected multi-component rice scheme (IR30-ER40), (2) systematically explore the effects of changes in three types of parameters, including RF system parameters, sample structure, and material properties, on the characteristics of the RF treatments, and (3) compare and analyze the relative sensitivity of the overall heating rate and temperature uniformity index of the heated samples with respect to ±20% variation in the input parameters of model.