Study on Optimization of Rice-Drying Process Parameters and Directional Regulation of Nutrient Quality

Abstract

The physicochemical components of rice such as starch, protein, fat and water have significant influence on its nutritional value, and the drying process can easily cause changes in these components. In this paper, the effect of technical parameters on the nutritional quality of rice during hot-air drying was studied, and a control method of the rice-drying process based on effective accumulated temperature was proposed to ensure the drying quality and improve the drying efficiency. Through thin-layer drying experiments, hot-air temperature (T), humidity (RH), initial moisture content (MC), wind speed (V) and tempering ratio (TR) were selected as control factors, and the central composite design was adopted to optimize the experimental scheme. The relationship between each factor and nutrient quality was revealed through response surface analysis, and the regression model and process optimization parameters were established. The results show that the optimization parameters are as follows: hot-air temperature, 48.87 °C; humidity, 30.12%; initial moisture content, 21.31%; wind speed, 0.62 m/s; tempering ratio, 2.87; the optimized total drying time is 4.23 h; the effective accumulated temperature is 214.44 °C·h. The contents of protein, fat, amylose and amylopectin were 8.47 g/100 g, 1.97 g/100 g, 15.33 g/100 g and 60.50 g/100 g, respectively. The relative error of the verification test was 4.17%. The optimized process can effectively maintain the nutritional quality of rice and improve drying efficiency. This paper provides a new way to deeply explore the mechanism of rice quality change, and the established process reference chart provides a scientific basis for actual drying operations and the development of an intelligent control system.

1. Introduction

Rice is an important food crop in the world and one of the main dietary sources for human beings. Drying is the precursor of rice harvesting, storage and processing, and the drying process of rice plays a crucial role in grain quality. Scientific and effective drying technology not only helps to maintain grain quality standards but also helps to achieve energy saving and loss reduction. With the increasing attention of the public on food safety and nutritional value, how to improve drying efficiency while ensuring the rice quality and taking into account the overall quality of rice has become an urgent issue to be solved. The nutritional value of rice, especially the key physical and chemical properties such as starch, protein, fat and water, is directly related to human health and quality of life [1,2,3,4]. However, the drying process, especially at high temperatures, can lead to denaturation of the proteins in rice. Siddiqui et al. [5] found that protein denaturation would reduce its solubility and digestibility, thus affecting the nutritional value of rice. The longer the drying time, the more obvious the degree of protein degradation, which is related to the accumulation of carbohydrates, possibly due to the conversion and accumulation of starch [6]. Amanda Muller et al. [7] showed that, with the increase of drying temperature, the crystallinity of starch in rice decreased, resulting in its structure becoming loose, which further affected the gelatinization characteristics and digestion rate of starch. In addition, Guiyuan Xiang et al. [8] also showed that the drying process would also change the ratio of amylose to amylopectin, which might affect the performance of rice in the cooking process. Ragaee et al. [9] found that far-infrared radiation drying can inhibit the biological activity of α-amylase, thus affecting the content of starch. Therefore, during rice drying, temperature and time have significant effects on the content and properties of nutrients, and thus affect the food quality and nutritional value of rice. A thorough understanding of these changes is important for optimizing the drying and storage process of rice and provides a theoretical basis for improving the overall quality of rice.

Accumulated temperature is an important indicator of crop growth, which was first proposed by Tongbin et al. in 1735 [10], and it is believed that plants are more consistent in their temperature demand at all growth stages and that the accumulated temperature can effectively reflect the maturity of crops. Dong et al. [11] showed that cumulative temperature changes significantly affect crop growth, planting methods and agricultural production, and maize growth can be regulated by cumulative temperature calculation [12]. Qin et al. [13] pointed out that cumulative temperature is crucial for improving maize yield, while Wu et al. [14] in this class group explored the application of cumulative temperature in maize drying regulation. Previous studies have demonstrated the importance of cumulative temperature for seed germination [15,16], vegetative growth [17], yield enhancement [18] and pest and disease reduction [19]. Similarly, we believe that cumulative temperature has a similar role in the dry seed growth of rice because the cumulative temperature required for rice grains to drop to a safe moisture content (around 14.5%) is relatively stable under certain drying conditions. Accordingly, the study in this paper is the first to use the amount of cumulative temperature in a model to reflect the drying process, which will improve the accuracy of the model. Therefore, it is a challenging task to precisely regulate the drying quality of rice grains. For this reason, research teams at home and abroad have carried out a large number of related studies, and our group has also completed a series of experiments on the directional regulation of the quality of high-quality rice grains [18,20,21,22,23,24] The results showed that factors such as drying medium temperature, relative humidity, wind speed and tempering ratio had a significant effect on the physicochemical indexes of rice grains, and there was a complex stress coupling and interconversion effect between a variety of environmental factors during the drying process [25]. In the drying process, there are complex stress coupling and interconversion effects between multiple environmental factors [18].

Based on this, the present study introduces accumulated temperature parameters to regulate nutrient quality during the drying process. It utilizes regression models to create reference process diagrams that visualize the relationship between drying parameters and nutrient indices, thereby systematically enhancing the accuracy of the drying procedure. The effects of various technological parameters on the nutritional quality of rice during its drying were examined, leading to a proposed control strategy based on effective accumulated temperature. The key control variables selected include hot-air temperature, humidity, initial moisture content, wind speed and tempering ratio. A central composite design scheme was employed for experimentation, allowing for an investigation into the interactions between process parameters and nutrient quality through response surface analysis. Following data analysis, a regression model was established to optimize comprehensive process parameters; additionally, a reference chart featuring both process retrieval and prediction functions was developed. The results indicated that the optimized drying process not only effectively preserves the nutritional quality of rice but also significantly enhances drying efficiency. This study enriches our understanding of the mechanisms underlying changes in rice quality while providing a scientific basis for practical rice-drying operations. Furthermore, the constructed process reference map facilitates both query and prediction capabilities regarding processes; it offers theoretical support and practical guidance for designing intelligent control systems as well as visualizing regulatory adjustments in future applications—thus laying a solid foundation for optimizing and intelligently implementing rice-drying processes.

2. Materials and Methods

2.1. Experimental Materials

This study selected samples from the high-quality variety Nanjiang 5055, procured in December 2023 from Zhangjiagang City. During the sample preparation process, an oven method at 105 °C was employed to determine the initial moisture content of the rice, which was found to be 30% (w.b.), and a cracking rate of 1% was confirmed. The moisture analyzer was calibrated using the oven method, and, prior to commencing experiments, the moisture content of the paddy rice was adjusted to the desired level. Throughout the experiment, real-time monitoring of moisture levels was conducted using an HS153 moisture analyzer to ensure compliance with experimental gradients. Before testing commenced, damaged and shriveled grains were removed; only plump and intact samples were selected and stored in a sealed container within a cold storage facility maintained at 4 ± 1 °C.

2.2. Equipment and Instruments

The instruments used were a JK-LB1700 multi-parameter controllable in situ precise drying experimental bench (Changchun Jida Scientific Instrument Equipment Co., Ltd.; Changchun, China), PQ-520 single-grain moisture meter for rice and wheat (Mettler Toledo Group, Zurich, Switzerland), BLH-3250 huller (Zhejiang Bethlehem Instrument Equipment Co., Ltd.; Hangzhou, China), DGG-9250GD electric constant-temperature blast-drying oven (Shanghai Senxin Experimental Instrument Co., Ltd., Shanghai, China), PL3002-IC electronic analysis balance (Mettler Toledo Instruments Ltd.; Zurich, Switzerland), FSJ-IIA Hammer Cyclone Mill (China Grain Reserve Management Group Co., Ltd. Beijing, China), Climacell Series Constant Temperature and relative humidity Chamber (Leica Instruments GMBH; Brunswick, Germany), OS-200 Orbital Oscillator (Hangzhou Aosheng Instrument Co., Ltd.; Hangzhou, China), TG16K Centrifuge (Changsha Dongwang Experimental Instrument Co., Ltd.; Changsha, China), and HS153 moisture analyzer (Mettler Toledo Group, Zurich, Switzerland).

The test was conducted in the Changping pilot Test Base of the Scientific Research Institute of the National Food and Strategic Reserves Administration. The drying test device was developed independently [26]. The structure of the test device is shown in Figure 1. The test device can ensure that the temperature and flow rate of the drying medium inside the experimental chamber is consistent and uniform when the test requires that the material is located in an environment of high relative humidity conditions, and, if the evaporation of water from the grain itself cannot meet the requirements, the test device has the function of automatic humidification. The test device can control the temperature of the drying medium, relative humidity and flow rate and other parameters, monitoring temperature, relative humidity, moisture, wind speed, etc. (the details are shown in Figure 1d); can restore the grain in the dryer in environmental conditions simulating different factors and levels of grain drying; can be adjusted at any time in its micro-environmental drying parameters; and can be completed to approximate the level of in situ measurements of the test, such as constant temperature or temperature-varying forms of drying tests.

2.3. Experimental Methods

Before the start of the experiment, the researchers first precisely preset the core operating parameters such as drying temperature, wind speed and relative humidity in the central control system of the experimental device to ensure the consistency and accuracy of the experimental conditions. The parameter control operation page is shown in Figure 1d. After the preset parameters were entered, a 20-min preheating procedure was performed to fully achieve and maintain the set operating stable state. At the same time, the experimental rice samples were removed from the low temperature storage environment (cold storage) and left in a constant indoor environment to acclimate to room temperature conditions. In the experimental stage, 1000 g of rice samples were accurately weighed (using a digital electronic balance with an accuracy of ±0.01 g) and then placed in the material chamber of the drying equipment to start the drying process. Throughout the drying process, the rice samples were weighed every 15 min to monitor real-time water loss. At the end of the drying stage, the samples were transferred to an oven with the same set temperature for tempering treatment according to the preset tempering ratio. The time length of the tempering ratio was strictly calculated according to the preset tempering ratio. In order to simulate a real tempering environment, the special tempering box used in this study has good sealing performance to ensure that the influence of external environmental factors on the sample is minimized during the tempering process.

All rice-drying experiments were continued until the samples reached the predetermined moisture content threshold of 14% (w.b.). The dried rice samples were kept stable at room temperature for 48 h, and then a series of quality indicators were determined, including the percentage of burst waist growth, the percentage of whole milled rice, the content of resistant starch and the percentage of germination. After that, the samples were stored in a cold storage environment at 4 °C, so as to detect the deep nutritional quality of the rice such as fatty acid value, protein content and fat content. In order to ensure the reliability and statistical validity of the experimental data, all the quality tests under each group of experimental conditions were repeated in three parallel experiments.

2.4. Measurement of Experimental Indicators

(1) Wet base moisture content (MC)

The water content of rice can be calculated according to Formula (1):

$$MC={\displaystyle \frac{m_t-m_d}{m_t}}$$

(1)

In the formula, MC denotes wet basis moisture content (%); $m_d$ is the mass of the dry sample (g); and $m_t$ is the mass of the wet sample (g).

(2) Effective accumulated drying temperature (AT)

Total drying time represents the total residence time of rice in the dryer, and its formula is shown in Equation (2):

$$Dtt=Dtn+Tt$$

(2)

where Dtt total drying time; Dtn represents net drying time; and Tt represents tempering time.

The effective cumulative temperature is the total effective drying temperature of the paddy throughout the drying process, which is borrowed from the concept of the cumulative temperature of plant growth in agronomy. According to the results of previous research [18,27], the desorption equilibrium temperature was uniformly set to 0 °C. The calculation method is as follows:

$$A{T}_n=\sum _{i=1}^n\left(T-T_{ei}\right)\times t_i$$

(3)

ATn is the accumulated drying temperature of the thin layer of rice (°C·h); T is grain temperature (°C); $t_i$ is the weighing period of rice; T_ei is the desorption equilibrium temperature of rice in the ith weighing cycle, °C.

(3) Tempering Ratio (TR)

Tempering Ratio is the ratio between “drying time” and “tempering time” during the drying process. This ratio can help assess whether adequate moisture regain (tempering) is required during drying to avoid excessive drying. The calculation formula is provided as follows:

$$TR={\displaystyle \frac{D\mathit{tn}-T_t}{T_t}}$$

(4)

Tt: the time spent in the drying process in order to improve the quality of the material and avoid excessive drying.

(4) Protein content (PC)

This was according to GB/T 24897-2010 grain and oil inspection “Determination of crude protein content of rice” [28].

(5) Fat content (FC)

Refer to GB 5009.6-2016 “National Standard for Food safety—Determination of fat in food” [29].

(6) Amylose content (AC)

Refer to GB/T 15683-2008 “Determination of amylose content in rice” [30].

(7) Amylopectin content (APC)

Amylopectin content is determined by DB32/T 2265-2012 [31].

2.5. Data Analysis

SPSS18.0 was used for statistical analysis; Origin 8.0 was used for drawing charts; and Design Expert V8.0.6.1 software was used for analysis and processing.

2.6. Experimental Design Scheme

Response Surface Methodology (RSM) was used to construct a five-factor and five-level orthogonal experimental design covering five independent variables with five different levels for each variable. Compared with previous research schemes [18,22], this scheme included more dimensions of response indicators in the evaluation of the dry rice quality, so as to comprehensively consider the influencing factors of the dry rice quality. In this study, the Central Composite Design (CCD) was selected as the basic framework of RSM experimental design. A total of 59 experimental units were designed, including 32 factor design points, 10 axial points and 17 central points, in order to deeply analyze the interaction effects between the studied variables.

The effect of temperature on the output results in this study can be found in detail in previous studies by scholars. The temperature selection interval is determined by GB/T21015-2007 “Technical Specification for Paddy Drying” [32], which states that the permissible heating temperature of the paddy in drying is ≤40 °C. And, when the paddy drying temperature exceeds 45 °C, the internal starch arrangement of rice is chaotic, free fatty acid content increases, and the degree of aging of rice increases, resulting in a decline in taste, so a drying temperature not exceeding 45 °C is appropriate [33]. Therefore, the drying temperature should not exceed 45 °C. The relative humidity selection interval was obtained from the relative humidity inside the machine collected in the previous period on multiple circulating paddy dryers. The collected data showed that the ambient relative humidity interval was about 40–65% during the drying operation of paddy grains undergoing drying in the dryer. The initial moisture interval selected in this study covers the moisture of the drying object in the actual dryer. The purpose of studying the effect of initial moisture on the output term is to demonstrate that the drying profiles of grains with different moisture gradients under the same drying conditions are not the same. The wind-speed selection interval was obtained from the in-machine wind speeds collected in the previous period on multiple circulating paddy dryers [27]. Therefore, five key process parameters and their corresponding variation ranges were selected: drying temperature (T, 35 °C to 55 °C), air relative humidity (RH, 30% to 50%), initial moisture content (MC, 20% to 28%), wind speed (V, 0.36 m/s to 0.84 m/s) and tempering ratio (TR, 1 to 4). The specific level settings for each variable are detailed in

Table 1. Factor levels of response surface test.

Factors	Symbols	Coded Versus Uncoded Values
		−α = 2.378	−1	0	1	α = 2.378
Drying temperature (°C)	T	35	40.8	45	49.2	55
Relative humidity (%)	RH	30	35.8	40	44.2	50
Initial moisture content (%)	MC	20	22.3	24	25.7	28
Wind speed (m/s)	V	0.36	0.5	0.6	0.7	0.84
Tempering ratio	TR	1	1.9	2.5	3.1	4

Note: hot-air temperature, T; hot-air relative humidity, RH; initial moisture content, MC; hot-air wind speed, V; tempering ratio, TR.

. The response parameters of interest in this study covered the key characteristics of rice during drying, such as net drying time and cumulative effective accumulated temperature, and multiple nutritional quality indicators of dried rice, including protein content, fat content, amylose content and amylopectin content. The protocol and results of the response surface experiment are shown in Table 1.

3. Results

After rigorous statistical analysis of the experimental data, this study successfully identified and quantified the main effective parameters that dominated the rice-drying process and their mutual interaction effect responses. In order to further confirm whether there were significant differences in these response indicators, and to screen out the most explanatory and consistent regression models, analysis of variance (ANOVA) was performed on the obtained experimental results.

3.1. Experimental Results and Analysis of Variance

The experiment investigated the effects and interactions of various factors on the response.

Table 2. Protocol and results of response surface experiment.

Std	T (°C)	RH (%)	MC (%)	V (m/s)	TR	Dtt (min)	AT (°C·h)	PC (%)	FC (%)	AC (g/100 g)	APC (g/100 g)
1	40.8	35.8	22.3	0.5	1.9	7.38	301.17	8.35	2.46	15.22	58.02
2	49.2	35.8	22.3	0.5	1.9	5.28	259.86	8.31	2.21	11.82	61.11
3	40.8	44.2	22.3	0.5	1.9	9.54	389.10	8.45	2.30	11.64	62.44
4	49.2	44.2	22.3	0.5	1.9	8.22	404.34	8.38	2.49	7.52	66.62
5	40.8	35.8	25.7	0.5	1.9	9.52	388.21	8.41	2.95	19.33	53.37
6	49.2	35.8	25.7	0.5	1.9	6.61	325.13	8.46	2.63	16.01	55.33
7	40.8	44.2	25.7	0.5	1.9	12.39	505.44	8.43	2.52	13.66	60.88
8	49.2	44.2	25.7	0.5	1.9	9.65	474.86	8.50	2.36	9.44	64.78
9	40.8	35.8	22.3	0.7	1.9	5.96	243.17	8.75	2.71	9.07	63.78
10	49.2	35.8	22.3	0.7	1.9	4.49	220.99	8.65	2.31	11.79	63.31
11	40.8	44.2	22.3	0.7	1.9	7.42	302.87	8.81	2.55	11.79	61.71
12	49.2	44.2	22.3	0.7	1.9	7.29	358.67	8.71	2.53	6.77	66.31
13	40.8	35.8	25.7	0.7	1.9	9.51	388.14	8.56	2.75	17.73	55.12
14	49.2	35.8	25.7	0.7	1.9	5.92	291.18	8.49	2.34	18.26	55.4
15	40.8	44.2	25.7	0.7	1.9	12.38	504.97	8.66	2.32	18.04	56.2
16	49.2	44.2	25.7	0.7	1.9	8.19	403.11	8.49	2.27	15.93	59.17
17	40.8	35.8	22.3	0.5	3.1	9.33	380.73	8.57	2.04	8.23	65.31
18	49.2	35.8	22.3	0.5	3.1	6.19	304.38	8.60	2.22	10.23	63.31
19	40.8	44.2	22.3	0.5	3.1	12.36	504.36	8.54	2.27	12.76	62.69
20	49.2	44.2	22.3	0.5	3.1	7.37	362.36	8.57	2.58	7.22	64.89
21	40.8	35.8	25.7	0.5	3.1	12.44	507.55	8.67	2.32	9.30	62.4
22	49.2	35.8	25.7	0.5	3.1	8.37	411.64	8.74	2.27	12.99	61.19
23	40.8	44.2	25.7	0.5	3.1	15.40	628.32	8.61	2.02	7.18	65.41
24	49.2	44.2	25.7	0.5	3.1	10.47	514.96	8.70	2.34	8.28	64.72
25	40.8	35.8	22.3	0.7	3.1	8.26	337.01	8.70	2.06	6.37	67
26	49.2	35.8	22.3	0.7	3.1	5.16	254.04	8.63	2.02	9.53	64.81
27	40.8	44.2	22.3	0.7	3.1	9.31	379.78	8.68	2.23	13.88	58.56
28	49.2	44.2	22.3	0.7	3.1	7.23	355.55	8.56	2.78	10.36	60.63
29	40.8	35.8	25.7	0.7	3.1	11.31	461.31	8.69	2.10	11.68	61.8
30	49.2	35.8	25.7	0.7	3.1	7.21	354.57	8.71	2.03	18.23	57.55
31	40.8	44.2	25.7	0.7	3.1	12.30	501.84	8.66	1.98	14.74	55.06
32	49.2	44.2	25.7	0.7	3.1	8.43	414.59	8.64	2.16	16.37	55.6
33	35	40	24	0.6	2.5	13.20	462.06	8.62	2.60	15.20	56.81
34	55	40	24	0.6	2.5	5.44	299.11	8.61	2.61	13.52	57.19
35	45	30	24	0.6	2.5	5.49	246.83	8.64	2.53	18.49	55.78
36	45	50	24	0.6	2.5	13.16	592.13	8.61	2.58	14.15	58.72
37	45	40	20	0.6	2.5	5.32	239.25	8.46	2.02	9.93	64.33
38	45	40	28	0.6	2.5	9.87	444.30	8.48	1.97	17.87	56.69
39	45	40	24	0.36	2.5	9.70	436.58	8.68	2.47	3.78	65.63
40	45	40	24	0.84	2.5	7.19	323.63	8.86	2.40	9.22	63.14
41	45	40	24	0.6	1	6.61	297.60	8.48	2.39	10.09	63.48
42	45	40	24	0.6	4	9.00	405.00	8.63	1.72	10.47	64.78
43	45	40	24	0.6	2.5	8.13	365.63	8.80	2.17	12.58	62.92
44	45	40	24	0.6	2.5	8.03	361.43	8.77	2.00	8.88	64.58
45	45	40	24	0.6	2.5	8.05	362.10	8.73	2.13	8.78	64.72
46	45	40	24	0.6	2.5	8.08	363.60	8.66	2.38	10.54	63.47
47	45	40	24	0.6	2.5	8.88	399.53	8.72	2.19	9.00	64.36
48	45	40	24	0.6	2.5	7.92	356.18	8.73	2.06	10.82	63.33
49	45	40	24	0.6	2.5	8.15	366.90	8.80	2.17	12.59	62.84
50	45	40	24	0.6	2.5	7.99	359.40	8.75	2.08	8.90	64.42
51	45	40	24	0.6	2.5	7.24	325.80	8.72	2.14	8.79	64.58
52	45	40	24	0.6	2.5	8.08	363.60	8.67	2.27	10.51	63.58
53	45	40	24	0.6	2.5	8.88	399.45	8.73	2.17	9.15	64.2
54	45	40	24	0.6	2.5	8.17	367.73	8.73	2.11	10.78	63.86
55	45	40	24	0.6	2.5	7.24	325.80	8.74	2.08	8.90	64.45
56	45	40	24	0.6	2.5	8.04	361.88	8.72	2.14	8.98	64.47
57	45	40	24	0.6	2.5	8.83	397.28	8.68	2.27	10.51	63.37
58	45	40	24	0.6	2.5	8.08	363.53	8.72	2.17	9.10	63.44
59	45	40	24	0.6	2.5	7.97	358.43	8.73	2.11	10.71	63.41

Note: hot-air temperature, T; hot-air relative humidity, RH; initial moisture content, MC; hot-air wind speed, V; tempering ratio, TR; total drying time, Dtt; effective accumulated temperature, AT. The protein content of undried rice was 7.58 g/100 g; fat content was 1.9 g/100 g; amylose AC was 6.62 g/100 g; amylopectin APC was 118.7 g/100 g.

presents the factor levels for each specific experimental group along with their corresponding response results. In

Table 3. Quadratic multinomial regression model.

Responses	Polynomial Regression Model Equation	R2	CV
Dtt	$$\begin{gathered} {\mathrm{Y}}_1=-22.9048-0.0481367\times \mathrm{A}-0.876145\times \mathrm{B}+3.05861\times \mathrm{C}-6.0108\times \mathrm{D}+7.64116\times \mathrm{E} \\ \hspace{1em}-0.0527909\times \mathrm{A}\times \mathrm{C}-0.147053\times \mathrm{A}\times \mathrm{E}+0.0145935\times {\mathrm{A}}^2 \\ \hspace{1em}+0.0146102\times {\mathrm{B}}^2 \end{gathered}$$	0.9398	7.06
AT	$$\begin{gathered} {\mathrm{Y}}_2=-1532.43+21.2508\times \mathrm{A}-39.7059\times \mathrm{B}+138.353\times \mathrm{C}-819.682\times \mathrm{D}+299.058\times \mathrm{E} \\ \hspace{1em}-1.65332\times \mathrm{A}\times \mathrm{C}-5.50444\times \mathrm{A}\times \mathrm{E}+0.272099\times {\mathrm{A}}^2+0.661016\times {\mathrm{B}}^2 \\ \hspace{1em}-0.706456\times {\mathrm{C}}^2+460.622\times {\mathrm{D}}^2-1.4306\times {\mathrm{E}}^2 \end{gathered}$$	0.9190	7.07
PC	$$\begin{gathered} {\mathrm{Y}}_3=-13.531+0.166811\times \mathrm{A}+0.132395\times \mathrm{B}+0.942584\times \mathrm{C}+11.2665\times \mathrm{D} \\ \hspace{1em}+0.720474\times \mathrm{E}-0.0662478\times \mathrm{A}\times \mathrm{D}-0.0100738\times \mathrm{B}\times \mathrm{E} \\ \hspace{1em}-0.2468\times \mathrm{C}\times \mathrm{D}+0.029186\times \mathrm{C}\times \mathrm{E}-0.807224\times \mathrm{D}\times \mathrm{E} \\ \hspace{1em}-0.00143806\times {\mathrm{A}}^2-0.00133621\times {\mathrm{B}}^2-0.0180165\times {\mathrm{C}}^2+0.202091\times {\mathrm{D}}^2 \\ \hspace{1em}-0.0901186\times {\mathrm{E}}^2 \end{gathered}$$	0.8956	0.5380
FC	$$\begin{gathered} {\mathrm{Y}}_4=3.17251+0.19669\times \mathrm{A}+0.68291\times \mathrm{B}-0.303304\times \mathrm{C}+0.0720186\times \mathrm{D} \\ \hspace{1em}+0.069834\times \mathrm{E}+0.0943789\times \mathrm{A}\times \mathrm{B}-0.0434371\times \mathrm{A}\times \mathrm{C}+0.145768\times \mathrm{A}\times \mathrm{E} \\ \hspace{1em}-0.158044\times \mathrm{B}\times \mathrm{C}+0.152018\times \mathrm{B}\times \mathrm{E}-0.0878355\times \mathrm{C}\times \mathrm{D} \\ \hspace{1em}+0.0767758\times {\mathrm{A}}^2+0.10754\times {\mathrm{B}}^2-0.0445448\times {\mathrm{C}}^2+0.0529495\times {\mathrm{D}}^2 \\ \hspace{1em}-0.0366893\times {\mathrm{E}}^2 \end{gathered}$$	0.8974	3.96
AC	$$\begin{gathered} {\mathrm{Y}}_5=452.644-5.27827\times \mathrm{A}-2.88352\times \mathrm{B}-13.8538\times \mathrm{C}-273.98\times \mathrm{D}-21.8944\times \mathrm{E} \\ \hspace{1em}-0.0597475\times \mathrm{A}\times \mathrm{B}+0.076842\times \mathrm{A}\times \mathrm{C}+1.3194\times \mathrm{A}\times \mathrm{D} \\ \hspace{1em}+0.34728\times \mathrm{A}\times \mathrm{E}-0.0857762\times \mathrm{B}\times \mathrm{C}+2.28081\times \mathrm{B}\times \mathrm{D} \\ \hspace{1em}+0.355633\times \mathrm{B}\times \mathrm{E}+7.33367\times \mathrm{C}\times \mathrm{D}-0.692551\times \mathrm{C}\times \mathrm{E} \\ \hspace{1em}+10.5348\times \mathrm{D}\times \mathrm{E}+0.0454208\times {\mathrm{A}}^2+0.0650667\times {\mathrm{B}}^2+0.255692\times {\mathrm{C}}^2 \\ \hspace{1em}-57.6209\times {\mathrm{D}}^2+0.200655\times {\mathrm{E}}^2 \end{gathered}$$	0.9028	11.80
APC	Y6 = −363.404 + 5.3093 × A + 4.7274 × B + 7.12924 × C + 319.431 × D + 33.879 × E + 0.0435091 ×   A × B − 0.322917 × A × E + 0.0983018 × B × C − 3.56548 × B × D − 0.60119 × B × E − 6.23162 ×  C × D + 0.599877 × C × E − 14.25 × D × E − 0.0683901 × A2 − 0.0658901 × B2 − 0.207833 × C2	0.9625	1.32

Note: A stands for dry temperature; B stands for hot-air relative humidity; C stands for initial moisture content; D stands for wind speed; and E stands for tempering ratio. The terms in the equation are non-coded values.

, polynomial quadratic regression models for each response were obtained after excluding non-significant factors, and corresponding r² and CV values were calculated as standards to assess model accuracy. Here, r² represents the correlation coefficient, while CV denotes the coefficient of variation expressed as a percentage of standard deviation to mean ratio. A smaller value indicates higher reliability of data; typically, a CV value is expected to be less than 12%. The results of the analysis of variance are shown in

Table 4. Results of variance analysis of regression model.

Indicators	Sources of Variation	Sum of Squares	Degrees of Freedom	Mean Square	F-Measure	p-Value	Significant or Not
Dtt	Model	277.86	9	30.87	84.95	<0.0001	Yes
	Residuals	17.81	49	0.3634
	Misfitting term	14.50	33	0.4395	2.13	0.0551	NO
	Error	3.30	16	0.2065
AT	Model	3.753 × 105	12	31,273.03	43.49	<0.0001	Yes
	Residuals	33,074.27	46	719.01
	Misfitting term	26,382.28	30	879.41	2.10	0.0596	NO
	Error	6691.98	16	418.25
PC	Model	0.7948	15	0.0530	24.58	<0.0001	Yes
	Residuals	0.0927	43	0.0022
	Misfitting term	0.0709	27	0.0026	1.93	0.2879	NO
	Error	0.0218	16	0.0014
FC	Model	3.02	16	0.1885	22.95	<0.0001	Yes
	Residuals	0.3450	42	0.0082
	Misfitting term	0.2160	26	0.0083	1.03	0.8773	NO
	Error	0.1290	16	0.0081
AC	Model	659.89	20	32.99	17.64	<0.0001	Yes
	Residuals	71.09	38	1.87
	Misfitting term	44.89	22	2.04	1.25	0.3301	NO
	Error	26.20	16	1.64
APC	Model	15,217.98	15	1014.53	55.19	<0.0001	Yes
	Residuals	790.48	43	18.38
	Misfitting term	557.73	27	20.66	1.42	0.2341	NO
	Error	232.74	16	14.55

Note: total drying time, Dtt; effective accumulated temperature, AT. The protein content of undried rice was 7.58 g/100 g; fat content was 1.9 g/100 g; amylose AC was 6.62 g/100 g; amylopectin APC was 118.7 g/100 g.

, indicating that the model should exhibit significance (p < 0.05), while pseudo-loss should not demonstrate significance (p > 0.05) [22,34].

Quadratic and cubic regression models were developed for various factors including drying temperature, relative humidity, tempering ratio, initial moisture content, total drying time, effective cumulative temperature, protein content, fat content, amylose levels and amylopsin activity. A comparison of the coefficient of determination (R²) between the quadratic and cubic regression models revealed that, while the R² value improved with the inclusion of cubic terms, this enhancement was accompanied by a significant increase in model complexity and computational demands. Consequently, a quadratic mathematical model—excluding non-significant variables—was employed to assess the experimental results. The interaction between input variables and responses was further examined through analysis of variance (ANOVA). During this analytical process, non-significant model terms were eliminated, ultimately resulting in the quadratic model presented in Table 3. The ANOVA results indicated that the regression equations for all six indicators were statistically significant (p < 0.05), whereas lack-of-fit was not significant (p > 0.05), suggesting a strong correlation between experimental values and the regression model.

3.2. Response Surface Plot Analysis

(1) Response analysis of protein content (PC)

The protein content is one of the important indicators of the nutritional quality of rice, directly influencing its energy supply for human consumption. According to the analysis results presented in

Table 5. F value of each factor on each response index.

Factor F-Measure	Dtt	AT	PC	FC	AC	APC
A-T	41.49	3.42	4.23	12.60	37.04	8.88
B-RH	35.54	36.35	15.70	57.51	48.24	44.87
C-MC	154.04	135.84	70.75	12.92	7.06	0.2622
D-V	43.26	42.79	85.59	6.97	4.24	1.53
E-TR	57.17	49.75	6.27	0.5822	8.27	12.00

Note: hot-air temperature, T; hot-air relative humidity, RH; initial moisture content, MC; hot-air wind speed, V; tempering ratio, TR.

and Figure 2, the F-value reflects the relative impact of various factors on changes in rice protein content. Data analysis indicates that the primary factors affecting rice protein content, ranked by importance, are as follows: wind speed (V) > moisture content (MC) > relative humidity (RH) > tempering ratio (TR) > temperature (T). To further explore the patterns of variation in protein content, we conducted a detailed analysis of the regression equation provided in Table 3. The calculated coefficient of determination (R²) for the regression model was found to be 0.8956, indicating that this model fits well with the relationships among variables and possesses strong explanatory power. Additionally, the coefficient of variation (CV) was determined to be 0.5380%, demonstrating that this model exhibits high robustness and consistency when predicting protein content. Therefore, analyses based on this model not only contribute to revealing key factors influencing protein content but also provide scientific evidence and practical guidance for optimizing post-harvest drying processes for rice preservation [34,35].

The significant interaction effect between initial moisture content and wind speed can be clearly revealed by looking at Figure 3. The effect of wind speed on the protein content of rice grains was more pronounced when the wind speed was lower (compared to a high initial moisture content) and when drying at a low tempering ratio (compared to a high tempering ratio), and this effect intensified as the wind speed increased. The rice grain protein content peaked when dried at a medium initial water content ratio (24%), lower tempering ratio (1) and high wind speed (0.84 m/s). The reason behind this phenomenon is that the medium initial moisture content may have maintained the proper moisture content of the rice grains so that the proteins would not change due to over drying or over wetting. Lower tempering may help to slow down the thermal stress during drying and protect the protein structure. High wind speed may accelerate water evaporation, but, at the proper relative humidity, it does not cause too much thermal damage to the paddy, thus helping to maintain the protein content [35]. Yang Huiping et al. [36] confirmed that the soluble protein content of paddy grain after high-temperature hot-air treatment decreased significantly compared with natural air-drying, which may be attributed to the denaturation of proteins within the paddy grain caused by a high temperature for a long time, resulting in a decrease in the water-solubility of the proteins. Da Silva Timm et al. [37] investigated the changes in protein and starch characteristics of corn cobs at different drying temperatures. They found that under high-temperature conditions, the proteins in corn underwent denaturation, which subsequently affected their solubility. Similarly, Hassan et al. [38] examined the effects of high temperature on the physicochemical properties, color and viscosity of corn grains through microwave heating treatment. Their findings also indicated that prolonged exposure to high temperatures could lead to structural changes in proteins and a reduction in their solubility. Based on these research results, we hypothesize that rice—being a similar cereal crop—may experience analogous denaturation phenomena after prolonged exposure to elevated temperatures, resulting in decreased protein solubility. Therefore, high temperatures should be avoided as much as possible during the drying process.

(2) Response analysis of fat content (FC)

In the process of drying rice, it is essential to manage the drying conditions effectively to enhance the quality, nutritional content and shelf life of the rice. Environments characterized by high temperatures and humidity can accelerate fat hydrolysis in rice, resulting in the formation of glycerol and free fatty acids. This may also lead to lipid oxidation, producing compounds such as peroxides, aldehydes and ketones. These alterations not only impact the acid value but can also influence the aging processes and flavor characteristics of rice. Specifically, aldehydes generated from lipid oxidation might impair the flavor profile of rice, while free fatty acids arising from hydrolysis could compromise its storage stability and overall eating quality.

Based on the F values presented in Table 5 and Figure 2, factors influencing changes in fat content are prioritized as follows: relative humidity (RH) > moisture content (MC) > temperature (T) > wind speed (V) > rotational speed (TR). The regression equation detailing variations in fat volume is provided in Table 3 with a coefficient of determination R² at 0.8. During the drying phase, there was a reduction in fat levels within the rice grains. This decrease can be attributed partly to peroxide production along with glycerol and free fatty acids due to fat hydrolysis. Additionally, elevated temperatures facilitate lipid oxidation that leads to an increase in aldehydes and ketones; this effect is particularly pronounced at 37 °C where fat loss is most notable. Fat enzyme acts as a key hydrolytic enzyme responsible for breaking down fats. As illustrated in Figure 4, there is a gradual decline in fat enzyme activity throughout drying; this trend correlates with rising fatty acid concentrations alongside decreasing pH levels. Higher temperatures further stimulate fatty acid production, which results in lower pH values that inhibit fat enzyme activity [39,40].

(3) Amylose content response analysis (AC)

The content of amylose in rice is the key factor affecting the edible and nutritional quality of rice. According to the results in Table 5 and Figure 2, F value reflects the degree of influence of various factors on the response, and it can be seen that the order of factors affecting AC is RH > T > TR > V > MC. The regression equation of protein content is shown in Table 3. The coefficient of determination of the model R² is 0.9028 and the value of CV is 11.80%, indicating that the model has good fitting ability and reliability.

As can be seen from Figure 5, the interaction of temperature and relative humidity during the drying of paddy grains had a significant effect on the content of straight-chain starch within the experimental protocol. Compared with the undried rice grains, the whole straight-chain starch content (soluble and insoluble straight-chain starch content) of the rice grains treated with different drying parameters were increased as a whole; since branched-chain starch would be partially converted into straight-chain starch or other sugars by the drying environment, the content of straight-chain starch would be increased. Figure 5 shows that in the case of low relative humidity drying, AC is weakly decreasing and then increasing all the time as the drying temperature increases; in the case of high relative humidity drying, AC is negatively correlated with the drying temperature. The lowest values of AC appeared when the temperature was around 45–50 °C and the relative humidity was 42–46% [41], while the peak values of AC appeared at a low temperature and high relative humidity, and a high temperature and low relative humidity. Drying rice grains at a high temperature and low relative humidity promotes the activity of starch synthase and accelerates the process of starch synthesis, thus increasing the content of straight-chain starch. Whereas, drying at a high temperature and high relative humidity slows down the activity of starch synthase and inhibits the starch synthesis process, leading to a decrease in the content of straight-chain starch. Li et al. [42] studied the effects of high temperature on starch morphological changes and expression related to starch biosynthesis and degradation. This study provides an important reference for us to understand the effect of temperature changes on the structural characteristics and internal components of rice, especially the effect of temperature on the starch change of rice. Therefore, high temperature and low relative humidity promoted the formation of straight-chain starch, while high temperature and dry conditions with high relative humidity;, on the contrary, inhibited the formation of straight-chain starch. The main increase is in soluble straight-chain amylopectin, the relatively high content (13–20%) of which gives rice a good texture when cooked. A rice with a high content (13–20%) has a good texture [38,41,42]. According to China’s national standard GB/T 19266-2008 [43] for high-quality round-grained rice, a straight-chain starch content of 13% to 20% gives a better taste; a straight-chain starch content of more than 20% results in rice that is viscous, small, with a hard texture, no luster and poor taste; if the straight-chain starch content is too low, then the rice is soft, sticky and greasy, and elasticity is poor.

(4) Response analysis of amylopectin (APC)

Amylopectin content can affect the nutritional quality of rice. According to the result of Table 5, Figure 2, the F value reflects the influence degree of each factor on the response to the influencing factors of AC in the following order: RH > TR > T > V > MC. The regression equation of the protein content is as shown in Table 3; the model of the decision coefficient R² is 0.9625; the CV value is 1.32%; the model has good fitting capability and reliability.

It can be seen from Figure 6 that, within the scope of the test scheme, the interaction of temperature, humidity and tempering ratio of rice in the drying process has a significant impact on amylose content. Figure 6a,b shows that the drying temperature is 45 °C and humidity is 42%, and the amylopectin content reaches its peak. At a high relative humidity, TR was negatively correlated with amylopectin content. In the case of a high tempering ratio and low relative humidity, the correlation is positive. The content of amylopectin was the highest under the condition of a low humidity and high tempering ratio. The reasons are the slow water migration in rice, more crystallized starch particles and inhibition of starch synthase activity. The activity of starch synthetase may be inhibited to a certain extent, thus slowing down the speed of starch synthesis. This leads to a relative increase in amylopectin content at the same time as starch synthesis because amylopectin synthesis usually takes precedence over amylose synthesis. This results in the highest amylopectin content [1,44]. Li Hongyan et al. [45] on the study of starch microstructure, the relationship between amylase activity and different starch content found that temperature and amylase also have a significant impact on the decomposition and conversion of starch.

3.3. Parameter Optimization and Model Validation

A single experiment was conducted to evaluate the effectiveness of the optimization conditions provided by the model using the specific parameters suggested by the model. The results showed acceptable agreement between the experimental and predicted values, and, thus, the model effectively predicted the response [46]. In order to obtain the optimum combination of retarded drying parameters, the target values of total drying time were taken, cumulative temperature values were limited at a minimum, and the target ranges of values of protein content, fat content, straight-chain starch content and branched-chain starch content were taken. According to Liu Qiuyuan et al. [47], it was shown that the protein range of high-quality rice is 7.86–9.32%, and the fat content is generally between 1% and 2%. According to the Chinese national standard quality for japonica rice GB/T 19266, the straight-chain starch content range is 15–20 g/100 g and branched-chain starch accounted for the content range of 60–70 g/100 g; a straight-chain starch content of more than 20% results in rice that is viscous, small, with a hard texture, no luster and poor taste; so a straight-chain starch content of 13% to 20% gives a better taste. If the content of straight-chain starch is too low, the rice will be soft, sticky and greasy with poor elasticity. These ranges are applicable to most high-quality rice varieties and maintain their nutritional and flavor values [41,47].

The parameters were optimized using the central composite design of experiment in Design-Expert-13 software, and the optimal combination of process parameters within the experimental range was obtained as a hot-air temperature of 54.58 °C, relative humidity of 38.95%, initial moisture content of 20.54%, air velocity of 0.63 m/s, and tempering ratio of 3.74. Under this optimized parameter, the total time consumed by drying was 4.40 h, effective cumulative temperature was 216.40 °C-h, protein content was 8.19 g/100 g, fat content was 2.67 g/100 g, straight-chain starch content was 16.08 g/100 g and branched-chain starch content was 81.59 g/100 g. Validation experiments were carried out using the optimized parameters, and, in order to eliminate the random errors, three parallel experiments were carried out, and the results of the experiments are shown in

Table 6. Experimental verification results of optimized parameters and response indexes.

		T	RH	MC	V	TR	Dtt	AT	PC	FC	AC	APC	Total Error %
Optimal parameter validation	Optimal parameter prediction value	48.87	30.12	21.31	0.62	2.86	4.23	214.44	8.466	1.96	15.33	60.49
	Experimental Group 1						4.18	223.15	7.92	1.98	16.49	61.07
	Experimental Group 2						4.34	208.55	8.14	2.02	15.05	57.87
	Experimental Group 3						4.37	224.75	7.83	1.9	15.88	64.18
	Average error						4.73%	3.87%	5.94%	2.38%	4.33%	3.80%	4.17%
Response metric validation	Model predictions	46.4	52.99	19.2	0.7	1.5	9.59	446.4	8.42	3.71	15.14	100.12
	Experimental Group 1						9.57	453.49	8.57	3.59	15.45	100.7
	Error						0.23%	1.59%	1.78%	3.34%	2.03%	0.58%
	Model prediction	41.33	49.65	20.11	0.65	1.45	8.51	365.02	8.49	2.95	16.97	98.84
	Experimental Group 2						8.22	348.14	8.14	2.81	17.99	93.92
	Error						3.47%	4.62%	4.17%	4.60%	5.99%	4.98%
	Model predicted value	40.57	48	19.49	0.7	1.99	7.75	321.92	8.53	2.91	17.22	97.04
	Experimental Group 3						7.6	308.92	7.99	2.75	16.36	87.85
	Error						2.04%	4.04%	6.36%	5.59%	5.03%	9.47%
	Average error						4.79%	3.25%	6.89%	7.35%	6.57%	6.68%	5.92%

. Analysis shows that the average error between the experimental values and optimized parameter values is 4.17%, indicating that the experimental results are basically consistent with the optimized results, which are all in line with the standard parameter range of high-quality Japonica rice, and, therefore, the choice of factors affecting the drying quality of rice grain is reasonable. Therefore, the selection of factors affecting the drying quality of rice is reasonable [47,48,49].

In order to verify the fitting effect of the established regression model, three sets of values were randomly generated within its parameter range for validation testing. The predicted values were obtained according to the regression equation and the relative error between the experimental and predicted values was calculated. The results from Table 6 show that the relative average error between the experimental and model values is 5.92%; the predicted values are in acceptable agreement with the experimental values [46,50,51], which indicates that the model fitting is credible.

3.4. Process Reference Drawing

Based on the previous research in our group, by conducting actual measurements on a continuous rice dryer, we have developed a time-and-quality process reference chart for polished rice production, which has improved the overall polished rice rate and other processing indicators, making the use of the process chart more effective in improving rice quality indicators [20,21,22,23]. In this paper, a set of new process reference charts aimed at optimizing drying time and guaranteeing the nutritional quality of rice grains was constructed by using effective cumulative temperature in an innovative way. The maps visualize the expected performance of rice grains for various combinations of drying process parameters and support the user in backchecking the appropriate drying process conditions according to specific quality objectives. Focusing on the experimental data of this study, the sensitivity and controllability of the process parameters were considered, so the hot-air temperature was selected as the core query variable, and the initial moisture content and ambient relative humidity were included as auxiliary process parameters, which were reflected in the process reference diagram.

Figure 7a,b illustrates the process reference charts for multiple indicators at tempering ratios of 2 and 3, respectively, for different grain dryer tempering ratios. Figure 7 illustrates the process reference charts combining AT and PC. It can be observed that, at high initial moisture (~28%), the effective cumulative temperature decreases significantly with increasing temperature, while, at low initial moisture (~20%), the effective cumulative temperature is less sensitive to changes in temperature. Protein content did not change significantly at high and low moisture conditions, whereas, at medium moisture (~24%), protein content decreased with increasing temperature and peaked. It was observed that the effective cumulative temperature of the paddy decreased when the initial moisture decreased and the drying temperature increased, whereas it first increased and then decreased, indicating that the protein content was higher (higher than about 8.5%) at temperatures of 45–47 °C. After drying, the rice grain protein content over 8.5% has a favorable value for its palatability; for the effective cumulative temperature, the results of the experimental model prediction selected a more moderate value (300 °C-h); this value is more in line with most of the requirements of the quality indicators; both meet the ACP ≤ 8.5%, the Dtn ≤ 300 °C-h region that is the preferred interval. For example, In Figure 7a, 40% RH is represented by the AT-40%RH, PC-40%RH curve, which is surrounded by the region (the shaded diagonal part), i.e., the preferred region required.

For example, the reference process diagram illustrates usage; to give a protein content of more than 8.5%, if the raw grain initial moisture content is set at 22.8% and the inside of the dryer is 40%, the degree of relative humidity measurement results, according to process chart 7a, as an available drying temperature of 42 °C under drying (point A) and an effective accumulated temperature of 350 °C·h.

In addition to coupling effective accumulated temperature and protein content, other nutritional quality indicators can be coupled in order to regulate several quality indicators at the same time, e.g., Appendix A Figure A1 and Figure A2 is a process reference diagram of effective cumulative temperature coupled with other quality indicators.

4. Conclusions

This study used the Design Expert software to establish a regression mathematical model to fit the relationships between five drying process parameters (hot-air temperature, humidity, initial moisture content, wind speed and resting time ratio) and drying heat accumulation and key nutrient indicators (protein, fat and starch content). The variance analysis results of the model showed that the R² values of all models were above 0.89 and had significant statistical significance (p ≤ 0.05). The study used response surface analysis to investigate the influence of multiple factors on the nutritional quality of rice during drying. The findings showed that the interaction effect between initial moisture content and wind speed was significant, and an increase in wind speed may exacerbate protein oxidation and cause a decrease in protein content. Under high-temperature and high-humidity conditions, fat content decreased, which may be due to increased fat hydrolysis and oxidation. Low temperature and high humidity are conducive to the formation of linear starch, while high humidity conditions inhibit its formation. The interaction effect of humidity and temperature significantly affects the content of branched starch, and the highest content of branched starch is observed under low-temperature and low-humidity conditions, which may be related to the reduction in activity of starch synthesis enzymes. Through a central composite experimental design, the optimal drying process parameters were optimized: hot-air temperature 48.871 °C, humidity 30.120%, initial moisture content 21.312%, wind speed 0.621 m/s and resting time ratio 2.867. Under these optimized conditions, the total drying time was 4.232 h; the effective heat accumulation was 214.442 °C·h; and the main nutrient components (protein content 8.466 g/100 g, fat content 1.968 g/100 g, linear starch 15.334 g/100 g and branched starch 60.498 g/100 g) all met the expected values. The error rate of the validation experiment was 4.17%, which proved the effectiveness and reliability of the optimization scheme. Furthermore, this study innovatively proposed the effective heat accumulation as a reference index for optimizing drying time. At the same time, it provides theoretical support for understanding the mechanism of rice quality change during rice drying.