Changes in the Properties of Hazelnut Shells Due to Conduction Drying

Abstract

In this study, the physical properties of two hazelnut species were investigated before and after drying at different temperatures and durations. The results showed that the physical properties of the hazelnut samples, including size, volume, density, weight, kernel mass, and shell mass, were significantly affected by temperature, duration, and their interactions. In addition, the moisture content of the samples decreased with increasing temperature and drying duration. The lowest value for the Istarski duguljasti variety was 5.36% (160 °C and 45 min), while the lowest value for Rimski okrugli was measured at 160 °C and 60 min (5.02%). Ash content was affected by both temperature and time, with the Istarski duguljasti variety having a minimum value of 0.84% at 120 °C and 60 min and Rimski okrugli a maximum value of 1.24% at 100 °C and 30 min. The variables of the ultimate analysis, such as nitrogen, carbon, sulfur, and hydrogen, increased with increasing temperature and time. The oxygen content and the higher heating value decreased with increasing temperature. Energy optimization in the drying process is crucial to reduce costs and save time. Effective energy optimization measures can lead to significant cost savings and improved operational efficiency in the drying process.

1. Introduction

Recently, there has been an increased focus on the use of sustainable energy sources to minimize negative impacts on the environment [1]. Biomass energy production is limited by its renewability and utilization efficiency [2]. The utilization of biomass and biofuels has been demonstrated to mitigate the European Union’s dependence on external energy sources while concurrently contributing to the reduction of greenhouse gas emissions. Biomass constitutes a vital renewable energy source in the European Union, obtained from organic matter [3]. The utilization of biomass for energy offers numerous benefits, including a decrease in greenhouse gas emissions, such as carbon dioxide and toxic exhaust gas components. The feasibility of biomass for heat and power generation is dependent on various physicochemical and thermal properties, particularly the calorific value [4]. The hazelnut industry produces significant amounts of by-products, including waste hazelnut shells, which account for about 50% of the nut weight and can be processed into valuable products [5]. Hazelnut biomass, which is a byproduct of production, can be successfully used as a renewable energy source [6]. Hazelnuts are one of the most popular nuts in the world [7], whose proper drying method is necessary to maintain the overall quality of the raw material [8]. Hazelnuts are an important food source due to their beneficial organoleptic and nutritional properties [9,10,11]. The nutritional properties of hazelnuts are strongly influenced by the processing method [12]. The hazelnut shell is a waste material generated during the production of hazelnuts. As production residues, hazelnut shells have great potential for various bioproducts, but also the possibility of conversion into energy as biofuel [13]. The characteristics of hazelnuts, including chemical composition, are influenced by hazelnut variety, environmental factors, and processing method [14,15]. To maintain the quality of hazelnuts, it is necessary to choose the right processing method [16]. In order to use hazelnut shells as an energy source, the shells must be ground and dried to a certain amount of moisture [17].

Moralar and Çelen (2022) [17] studied the drying of hazelnut shells using a laboratory microwave dryer, which gave good results in terms of energy and time consumption. In order to study the influence of drying on maintaining the quality of hazelnuts, Turan (2019) [8] conducted research in which he studied the change in chemical properties of hazelnuts after natural and artificial drying. The author concluded that the application of artificial drying ensures more desirable properties of the raw material. In a study by Ercisli et al. (2011) [18], significant variations were observed among 12 hazelnut genotypes concerning their physical and mechanical properties, specifically within the Corylus avellana L. species. Maleki et al. (2013) [19] conducted a study to investigate the physical properties of hazelnuts. Their results showed that an increase in moisture content led to an increase in the true density of hazelnuts, while porosity decreased proportionally with decreasing moisture content. Turan et al. (2015) [20] used Response Surface Methodology (RSM) in their research to evaluate the effect of roasting on the functional properties of defatted hazelnut flour using two independent variables (time and temperature). The research demonstrated the ability of the RSM method to set conditions for optimizing the roasting process to preserve the functional value of hazelnuts. Sarkar et al. (2021) [21] conducted a study for mango leather production through dehydration and optimized its processing parameters using Response Surface Methodology (RSM). The RSM model was constructed using the central composite design and mathematical operations were applied to describe the model. The study results indicate that Response Surface Methodology (RSM) could serve as a beneficial technique for enhancing the production process of mango leather.

The objective of this investigation is to evaluate the physical and chemical characteristics of hazelnut biomass pre- and post-drying at four distinct temperatures (100, 120, 140, and 160 °C) and for four different durations (15, 30, 45, and 60 min). Physical properties include dimensions such as length, width, thickness, density, volume, fruit mass, shell mass, and kernel mass. Chemical properties include moisture content, dry mass, ash, coke, solid carbon, volatile matter, nitrogen, carbon, sulfur, oxygen, hydrogen, and higher heating value. The aim of this study is to comprehend the changes in physical and chemical properties that occur during the drying process and to evaluate the effects of individual parameters such as type of sample, drying temperature, and duration of the process. The researchers aim to measure the effects of these parameters on the changes in physical and chemical properties of hazelnut shells during the drying process in order to determine the optimal drying conditions and improve the uniformity of the final product. Studying the changes in the chemical and physical properties of hazelnut biomass caused by the drying process is a critical element in optimizing and developing strategies for using the feedstock as a sustainable energy source.

2. Materials and Methods

For this study, hazelnut cultivars Istarski duguljasti and Rimski okrugli (Virovitica Podravka County, geographic coordinates 45.62885123923944, 17.883251895676104) of the species Corylus sp. were used. The fruits of both species were harvested manually to avoid prevent shell damage to the peel and selected randomly from each bush. The sample was then cleaned and prepared for analysis using the quartering method.

Before and during a set of conductive drying processes at different temperatures (100 °C, 120 °C, 140 °C and 160 °C) and durations (15, 30, 45 and 60 min), hazelnut fruit and shells were evaluated for their physical and chemical properties. The hazelnut shells underwent both ultimate and proximate analysis to determine their chemical composition.

The analysis of hazelnut shells involves both ultimate and proximate measurements. The ultimate analysis quantifies the percentage of C, H, N, S, and O, whereas the proximate analysis variables include moisture, DM, ash, coke, and FC. The higher heating value (HHV) was determined by calorimetric analysis. The volatile matter (VM) content of the sample was computed following the procedure outlined in the HRN EN 15148:2009 standard [22]. The carbon (C) and hydrogen (H) content were measured utilizing a CHNS analyzer (Elementar Analyze Systeme GmbH, Langenselbold, Germany) [23], in accordance with the HRN EN ISO 16948:2015 standard protocol [24], while the oxygen (O) content was estimated as the residual of the elements C, H, N, S. The higher heating value (HHV) was determined via an IKA C200 oxygen bomb calorimeter (IKA Analysentechnik GmbH, Heitersheim, Germany) [25], following the HRN EN 14918:2010 standard protocol [26]. Utilizing a Mettler Toledo scale, the mass ratio of hazelnut fruit, kernel, and shell was determined. Initially, the weight of the entire fruit was measured, followed by separation of the shell from the kernels. The weights of the kernels and shells were then recorded. Subsequently, the physical and chemical properties of the hazelnut cultivars were assessed, and statistical analysis was performed using the TIBCO STATISTICA 13.3.0 software (StatSoft TIBCO Software Inc., Palo Alto, CA, USA) [27] to identify the variations in the aforementioned properties across the observed cultivars. The statistical analysis is presented as mean and standard deviation. The differences and statistical significance between the variables were determined with the aid of Analysis of Variance (ANOVA) and Tukey’s test, in relation to the different factors of drying temperature and process duration.

In addition, 3D plots (second-degree polynomials) were created using Response Surface Methodology (RSM) to model the proximate and ultimate analysis parameters as a function of process duration and drying temperature. The analysis was performed using the Python programming language (Python 3.10.0) [28] and associated packages (pandas, numpy, matplolib, and scikit-learn) [29,30,31,32]. The data were read from the database and sorted as x, y, and z axis variables. Next, a second-degree feature transformer was created using “PolynomialFeatures” from scikit-learn. This transformer was used to convert the temperature and time data into a matrix of polynomial features, where each row represents a unique combination of temperature and time values. A linear regression model was then fitted to the polynomial features using the “LinearRegression” function of scikit-learn. This model was used to predict the response surface, which is the relationship between temperature, time, and variables from ultimate and proximate analysis of the observed biomass. The response surface was then plotted in a 3D plot (using Axes3D from mpl_toolkits.mplot3d). A graphical representation of the relationship between temperature, time, and observed values obtained from ultimate and proximate analysis was created, which enables making predictions for new combinations of temperature and time. The univariate analysis was performed to determine the effect of individual parameters, as well as their interactions, on changes in physical and chemical properties of the biomass.

3. Results

Table 1. Physical properties of observed hazelnut varieties from natural samples.

Sample	Len. (mm)	Wid. (mm)	Thick. (mm)	Den. (g/cm3)	Vol. (cm3)	FM (g)	KM (g)	SM (g)
Istarski duguljasti	25.93 ± 1.24 b	19.93 ± 0.88 a	16.36 ± 1.28 a	1.05 ± 0.09 a	0.96 ± 0.08 b	3.85 ± 0.44 a	1.78 ± 0.21 a	2.07 ± 0.31 a
Rimski okrugli	21.41 ± 1.8 a	22.79 ± 1.2 b	15.92 ± 1.05 a	1.3 ± 0.21 b	0.8 ± 0.16 a	3.85 ± 0.45 a	1.75 ± 0.27 a	2.1 ± 0.33 a
Statistical significance	*	*	n.s.	*	*	n.s.	n.s.	n.s.

Len—Length; Wid—Width; Thick—Thickness; Den—Density; Vol—Volume; FM—Fruit mass; KM—Kernel mass; SM—Shell mass. Tukey’s HSD test identified statistically significant differences (p < 0.05) between means in the same column with distinct lowercase superscripts. Statistical significance is denoted as follows: * p ≤ 0.01, while the abbreviation n.s. signifies non-significance.

shows the physical properties of hazelnuts of the Istarski duguljasti and Rimski okrugli varieties before the drying treatment.

Statistical analysis of the physical properties of hazelnut varieties from natural samples revealed differences in mean values. The Istarski duguljasti variety had higher average values for length, thickness, volume, and kernel mass. In the observed Rimski okrugli sample, the values for width, density, and shell mass were higher on average. In the analysis of variance, significant differences were found for the variables of length, width, density and volume with a coefficient p ≤ 0.01. Ferrão et al. (2021) [9] found in their study that the weight of the hazelnut fruit ranged from 2.23 to 3.82 g, while the weight of the kernel ranged from 1.12 to 1.70 g for the different varieties studied. In the studies conducted, the width of the hazelnut varieties studied was 1.08 to 1.60 cm and the thickness was 1.00 to 1.42 cm. The authors also report the density (including the hazelnut shell) as 0.73–1.56 g/mL. In a study to determine the physical properties of 12 hazelnut genotypes, the length was 18.91–25.47 mm, the width was 15.09–21.20, and the thickness was 12.76–21.20 mm [18].

Table 2. Proximate, ultimate, and calorimetric analysis of natural samples.

Sample	Moisture (%)	DM (%)	Ash (%)	Coke (%)	FC (%)	VM (%)	N (%)	C (%)	S (%)	O (%)	H (%)	HHV (MJ kg−1)
Istarski duguljasti	17.14 ± 0.09 a	82.86 ± 0.09 a	0.99 ± 0.14 a	22.58 ± 0.86 a	17.73 ± 0.82 a	64.14 ± 0.69 a	0.31 ± 0.08 a	57.6 ± 2.68 a	0.1 ± 0.01 a	34.7 ± 3.25 a	7.29 ± 0.51 b	20.72 ± 0.21 a
Rimski okrugli	17.41 ± 1.19 a	82.59 ± 1.19 a	1.12 ± 0.13 a	23.22 ± 0.4 a	18.06 ± 0.38 a	63.41 ± 0.88 a	0.25 ± 0.09 a	57.95 ± 3.15 a	0.21 ± 0.14 a	34.94 ± 3.07 a	6.65 ± 0.16 a	21.12 ± 0.76 a
Statistical significance	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	**	n.s.

DM—Dry matter; FC—Fixed carbon; VM—Volatile matter; N—Content of nitrogen; C—Content of carbon; S—Content of sulfur; O—Content of oxygen; H—Content of hydrogen; HHV—Higher heating value. Tukey’s HSD test identified statistically significant differences (p < 0.05) between means in the same column with distinct lowercase superscripts. Statistical significance is denoted as follows: ** p ≤ 0.05, while the abbreviation n.s. signifies non-significance.

shows ultimate and proximate analysis of hazelnut samples (Istarski duguljasti and Rimski okrugli varieties) before the drying treatment.

The average percentage of moisture, ash, coke, fixed carbon, carbon, sulfur, oxygen, and HHV is higher for the observed Rimski okrugli variety. In the Istarski duguljasti variety, the proportion of DM, VM, N, and H is higher on average.

Analysis of variance showed a significant difference in the percentage of H in the observed variables, with a coefficient of p ≤ 0.05. The other observed variables are not statistically significant. Mueller et al. (2020) give the range of ash content in the study as 1.9–3.1%, while Sahin et al. (2022) give values in the range of 2.34–3.79%. Bohnhoff et al. (2019) [33] give a moisture content in the range of 6.2–18.9%. Smith et al. (2019) [34] give proximate analysis values for hazelnut shells: moisture (12.45%), VM (62.70%), FC (24.08%), and ash (0.77%); while Marcantonio et al. (2019) [35] give the following proximate analysis values: moisture content (7.90%), FC (26.39%), VM (72.45%), and ash (1.16%). Additionally, Smith et al. (2019) [33] give ultimate analysis values for hazelnut shells: C (46.76%), H (5.76%), O (45.83%), N (0.22%), and S (0.67%). Marcantonio et al. (2019) [35] indicate the proportion of ultimate analysis variables: C (50.38%), H (6.03%), N (0.22%), O (42.32%), and S (0.67%). In addition, the authors indicate an HHV value of 20.20 MJ/kg.

Table 3. Physical properties of observed hazelnut varieties after drying treatment.

Sample	Temp. (°C)	Time (min)	Len. (mm)	Wid. (mm)	Thick. (mm)	Den. (g/cm3)	Vol. (cm3)	FM (g)	KM (g)	SM (g)
Istarski duguljasti	100	15	25.31 ± 1.16 bcde	19.79 ± 0.46 abcde	15.56 ± 1.21 a	2.09 ± 0.12 abcdefg	0.48 ± 0.03 a	4.23 ± 0.06 a	1.91 ± 0.1 a	2.32 ± 0.11 ab
		30	26.97 ± 0.68 e	19.1 ± 1.6 abcde	14.34 ± 1 a	2.7 ± 0.21 efghijk	0.37 ± 0.03 a	3.95 ± 0.5 a	1.87 ± 0.31 a	2.07 ± 0.21 ab
		45	24.62 ± 0.33 abcde	18.99 ± 0.21 abcde	14.17 ± 1.48 a	3.12 ± 0.24 jk	0.32 ± 0.03 a	4.18 ± 0.74 a	1.99 ± 0.24 a	2.19 ± 0.51 ab
		60	24.81 ± 1.18 abcde	19.22 ± 0.83 abcde	15.25 ± 1.31 a	1.65 ± 0.54 a	1.65 ± 0.54 b	3.44 ± 0.29 a	1.59 ± 0.15 a	1.85 ± 0.15 ab
	120	15	25.26 ± 1.3 bcde	18.83 ± 1.11 abcde	14.45 ± 0.9 a	2.02 ± 0.11 abcdef	0.5 ± 0.03 a	4.43 ± 0.22 a	1.83 ± 0.21 a	2.6 ± 0.01 b
		30	24.35 ± 1.45 abcde	18.08 ± 1.18 ab	13.24 ± 0.33 a	1.78 ± 0.14 ab	0.57 ± 0.04 a	3.92 ± 0.22 a	1.85 ± 0.1 a	2.07 ± 0.21 ab
		45	26.23 ± 2.86 cde	19.04 ± 1.33 abcde	15.48 ± 1.61 a	1.82 ± 0.32 abc	0.56 ± 0.11 a	4.37 ± 0.34 a	1.91 ± 0.17 a	2.46 ± 0.19 ab
		60	26.01 ± 1.09 bcde	19.14 ± 0.78 abcde	15.95 ± 0.81 a	3.31 ± 0.23 k	0.3 ± 0.02 a	3 ± 0.21 a	1.43 ± 0.24 a	1.57 ± 0.03 a
	140	15	25.62 ± 1.62 bcde	18.82 ± 0.52 abcde	15.66 ± 0.88 a	2.31 ± 0.23 abcdefghi	0.44 ± 0.04 a	4.02 ± 0.21 a	1.9 ± 0.06 a	2.12 ± 0.16 ab
		30	24.81 ± 0.43 abcde	17.87 ± 0.52 a	14.46 ± 1.39 a	2 ± 0.34 abcde	0.51 ± 0.1 a	3.79 ± 0.22 a	1.79 ± 0.11 a	2 ± 0.13 ab
		45	26.7 ± 1.05 de	18.57 ± 0.44 abcd	14.89 ± 0.07 a	2.54 ± 0.22 cdefghij	0.4 ± 0.04 a	3.57 ± 0.56 a	1.51 ± 0.37 a	2.06 ± 0.27 ab
		60	25.64 ± 1.72 bcde	18.76 ± 0.49 abcde	15.38 ± 0.74 a	2.4 ± 0.17 bcdefghij	0.42 ± 0.03 a	4.06 ± 0.74 a	1.88 ± 0.32 a	2.19 ± 0.42 ab
	160	15	25.11 ± 0.7 bcde	18.18 ± 0.68 abc	13.9 ± 0.61 a	2.17 ± 0.09 abcdefg	0.46 ± 0.02 a	3.74 ± 0.36 a	1.69 ± 0.14 a	2.06 ± 0.29 ab
		30	25.78 ± 1.46 bcde	18.49 ± 1.05 abcd	13.88 ± 1.07 a	2.58 ± 0.24 defghijk	0.39 ± 0.04 a	3.57 ± 0.09 a	1.62 ± 0.06 a	1.95 ± 0.03 ab
		45	24.28 ± 0.87 abcde	18.43 ± 0.54 abcd	15.63 ± 0.72 a	2.27 ± 0.14 abcdefghi	0.44 ± 0.03 a	3.65 ± 0.28 a	1.83 ± 0.13 a	1.82 ± 0.19 ab
		60	23.35 ± 0.52 abcde	18.17 ± 0.66 abc	15.45 ± 0.23 a	2.75 ± 0.2 ghijk	0.36 ± 0.03 a	4.01 ± 0.27 a	1.76 ± 0.25 a	2.25 ± 0.17 ab
Rimski okrugli	100	15	21.83 ± 2.22 abcde	21.89 ± 1.87 bcde	16.01 ± 3.48 a	2.24 ± 0.05 abcdefgh	0.45 ± 0.01 a	3.89 ± 0.45 a	1.76 ± 0.05 a	2.13 ± 0.46 ab
		30	22.04 ± 3.37 abcde	20.88 ± 0.98 abcde	16.03 ± 3.08 a	1.78 ± 0.25 ab	0.57 ± 0.07 a	3.56 ± 0.14 a	1.68 ± 0.06 a	1.88 ± 0.19 ab
		45	19.58 ± 0.71 a	22.38 ± 1.02 de	14.07 ± 1.04 a	1.89 ± 0.16 abcd	0.53 ± 0.05 a	3.39 ± 0.16 a	1.51 ± 0.11 a	1.88 ± 0.1 ab
		60	19.52 ± 1.25 a	22.25 ± 1.08 de	14.12 ± 1.48 a	2.41 ± 0.22 bcdefghij	0.42 ± 0.04 a	3.99 ± 0.32 a	1.63 ± 0.21 a	2.37 ± 0.36 ab
	120	15	19.52 ± 0.91 a	22.66 ± 1.19 e	13.5 ± 1.67 a	2.74 ± 0.17 fghijk	0.37 ± 0.02 a	3.52 ± 0.22 a	1.66 ± 0.09 a	1.86 ± 0.14 ab
		30	20.94 ± 0.45 abc	21.09 ± 2.95 abcde	13.38 ± 2.03 a	2.16 ± 0.43 abcdefg	0.47 ± 0.09 a	4.09 ± 0.94 a	1.8 ± 0.28 a	2.29 ± 0.68 ab
		45	21.7 ± 2.11 abcde	21.66 ± 1.7 abcde	14.83 ± 2.28 a	1.8 ± 0.17 abc	0.56 ± 0.05 a	4.3 ± 0.19 a	1.83 ± 0.19 a	2.48 ± 0.13 ab
		60	19.43 ± 2.38 a	22.09 ± 1.62 cde	14.93 ± 1.44 a	2.29 ± 0.13 abcdefghi	0.44 ± 0.03 a	3.7 ± 0.59 a	1.69 ± 0.21 a	2.01 ± 0.44 ab
	140	15	20.59 ± 2.24 ab	20.67 ± 1 abcde	13.82 ± 1.42 a	2.92 ± 0.42 hijk	0.35 ± 0.05 a	4.38 ± 0.62 a	1.89 ± 0.27 a	2.49 ± 0.53 ab
		30	21.04 ± 3.14 abc	21.47 ± 0.88 abcde	14.3 ± 0.65 a	2.04 ± 0.15 abcdefg	0.49 ± 0.03 a	3.49 ± 0.45 a	1.55 ± 0.26 a	1.94 ± 0.19 ab
		45	22.32 ± 2.88 abcde	21.06 ± 2.27 abcde	16.43 ± 2.19 a	2.7 ± 0.2 efghijk	0.37 ± 0.03 a	3.78 ± 0.29 a	1.78 ± 0.1 a	2.01 ± 0.21 ab
		60	21.91 ± 1.88 abcde	22.2 ± 1.21 de	15.53 ± 1.82 a	2.13 ± 0.16 abcdefg	0.47 ± 0.04 a	3.62 ± 0.59 a	1.62 ± 0.13 a	2 ± 0.47 ab
	160	15	21.48 ± 2.59 abcd	21.18 ± 0.24 abcde	15.52 ± 1.94 a	2.35 ± 0.09 abcdefghi	0.43 ± 0.02 a	4.1 ± 0.19 a	1.69 ± 0.12 a	2.41 ± 0.18 ab
		30	20.75 ± 0.44 abc	21.13 ± 0.13 abcde	13.11 ± 1.03 a	2.42 ± 0.05 bcdefghij	0.41 ± 0.01 a	3.86 ± 0.33 a	1.81 ± 0.17 a	2.05 ± 0.2 ab
		45	21.43 ± 1.08 abcd	21.67 ± 2.52 abcde	13.04 ± 1.64 a	2.34 ± 0.15 abcdefghi	0.43 ± 0.03 a	4.2 ± 0.92 a	1.96 ± 0.41 a	2.24 ± 0.75 ab
		60	19.34 ± 1.05 a	20.02 ± 0.16 abcde	12.79 ± 0.77 a	2.98 ± 0.06 ijk	0.34 ± 0.01 a	3.85 ± 0.25 a	1.5 ± 0.31 a	2.35 ± 0.09 ab
Statistical significance			*	*	n.s.	*	*	**	n.s.	**

Len—Length; Wid—Width; Thick—Thickness; Den—Density; Vol—Volume; FM—Fruit mass; KM—Kernel mass; SM—Shell mass. Tukey’s HSD test identified statistically significant differences (p < 0.05) between means in the same column with distinct lowercase superscripts. Statistical significance is denoted as follows: * p ≤ 0.01, ** p ≤ 0.05, while the abbreviation n.s. signifies non-significance.

shows the physical properties of hazelnuts of the varieties Istarski duguljasti and Rimski okrugli after treatment by drying at different temperatures (100, 120, 140, and 160 °C) and for different durations of the process (15, 30, 45, and 60 min).

The Istarski duguljasti variety exhibited the shortest average length (23.35 mm) when roasted at 160 °C for 60 min, whereas the longest average length (26.97 mm) was observed at 100 °C for 30 min. The width of the variety had the highest average value (19.79 mm) at 100 °C for 15 min, while the lowest value (17.87 mm) was noted at 140 °C for 30 min. The highest average values for thickness and density of the observed Istarski duguljasti variety were obtained when drying at 120 °C for a duration of 60 min (15.95 mm and 3.31 g/cm³). The lowest average value for thickness was observed at a temperature of 120 °C and a drying time of 30 min (13.24 mm), while density was the lowest at 100 °C and a drying time of 60 min (1.65 g/cm³). The lowest average values for volume, fruit mass, kernel mass, and shell mass were observed at 120 °C and 60 min in duration (0.30 cm³, 3.00 g, 1.43 g and 1.57 g respectively). The highest values of the above parameters were obtained at 100 and 120 °C.

For the Rimski okrugli variety, the lowest average values for length (19.34 mm), width (20.02 mm), thickness (12.79 mm), volume (0.34 cm³), and kernel mass (1.50 g) were obtained at 160 °C and 60 min in process duration. The lowest average value for density (1.78 g/cm³) was noted at 100 °C and 30 min, while fruit mass (3.39 g) was measured at the same temperature and for 45 min in duration.

The mass of the shell had the lowest value at 120 °C and 15 min: 1.86 g. The highest values for length (22.32 mm) and thickness (16.43 mm) were obtained at 140 °C and 45 min, while the weight of the fruit (4.38 g) and shells (2.49 g) were at the same temperature and 15 min in duration. At a temperature of 160 °C for 60 and 45 min, the highest average values were found for density (2.98 g/cm³) and kernel mass (1.96 g). After applying the drying treatment, the variables of length, width, density, and volume showed a significant difference, with a coefficient of p ≤ 0.01, while fruit and shell mass showed a coefficient of p ≤ 0.05. The differences between the observed samples with respect to the variables of thickness and kernel mass were not significant.

Hazelnuts of the varieties Istarski duguljasti and Rimski okrugli were dried at different temperatures and durations, and the resulting proximate analysis is shown in

Table 4. Proximate analysis of observed hazelnut varieties after drying treatment.

Sample	Temp. (°C)	Time (min)	Moisture (%)	DM (%)	Ash (%)	Coke (%)	FC (%)	VM (%)
Istarski duguljasti	100	15	9.29 ± 0.23 p	90.71 ± 0.23 a	1.06 ± 0.12 cdefgh	23.57 ± 0.21 cdef	20.32 ± 0.37 a	69.34 ± 0.02 a
		30	7.71 ± 0.24 no	92.29 ± 0.24 bc	1.03 ± 0.05 cdefg	23.63 ± 0.25 cdefg	20.78 ± 0.34 ab	70.48 ± 0.04 bcde
		45	8.03 ± 0.01 o	91.97 ± 0.01 b	1.15 ± 0.07 efghi	23.55 ± 0.87 cdef	20.51 ± 0.86 bc	70.31 ± 0.81 abcd
		60	7.67 ± 0.01 n	92.33 ± 0.01 c	0.99 ± 0.07 abcde	23.13 ± 0.29 abc	20.36 ± 0.34 bcd	70.98 ± 0.26 defgh
	120	15	7.79 ± 0.18 no	92.21 ± 0.18 bc	0.93 ± 0.04 abcd	24.2 ± 0.07 efghij	21.38 ± 0.06 bcd	70.43 ± 0.99 bcde
		30	6.95 ± 0.04 klm	93.05 ± 0.04 def	1.01 ± 0.02 bcdef	24.94 ± 0.14 jk	22.19 ± 0.14 bcde	69.84 ± 0.16 abc
		45	6.67 ± 0.03 ijkl	93.33 ± 0.03 efgh	1.01 ± 0.04 bcdef	24.47 ± 0.35 ghij	21.83 ± 0.38 bcdef	70.49 ± 0.31 bcde
		60	6.9 ± 0.19 jklm	93.1 ± 0.19 defg	0.84 ± 0.02 a	23.81 ± 0.23 cdefghi	21.32 ± 0.16 bcdef	70.94 ± 0.36 defgh
	140	15	7 ± 0.04 lm	93 ± 0.04 de	0.92 ± 0.05 abc	24.08 ± 0.17 defghij	21.47 ± 0.1 bcdefg	70.61 ± 0.19 bcdef
		30	6.61 ± 0.22 ij	93.39 ± 0.22 gh	1.01 ± 0 bcdef	24.14 ± 0.4 efghij	21.54 ± 0.43 bcdefgh	70.84 ± 0.2 cdefg
		45	6.64 ± 0.02 ijk	93.36 ± 0.02 fgh	1.03 ± 0 cdefg	24.1 ± 0.03 defghij	21.47 ± 0.04 cdefghi	70.86 ± 0.02 cdefgh
		60	5.96 ± 0.15 def	94.04 ± 0.15 klm	1.2 ± 0.01 hi	23.84 ± 0.26 cdefghi	21.22 ± 0.22 cdefghij	71.62 ± 0.36 fghi
	160	15	6.27 ± 0.07 fgh	93.73 ± 0.07 ijk	1.16 ± 0.06 fghi	25.56 ± 0.17 k	22.79 ± 0.11 cdefghij	69.77 ± 0.1 ab
		30	5.91 ± 0.08 cde	94.09 ± 0.08 lmn	0.99 ± 0.03 abcd	24.16 ± 0.26 efghij	21.74 ± 0.29 cdefghij	71.36 ± 0.18 efghi
		45	5.36 ± 0.03 ab	94.64 ± 0.03 op	1.09 ± 0.03 defghi	24.04 ± 0.27 defghi	21.66 ± 0.22 defghijk	71.89 ± 0.28 hij
		60	5.58 ± 0.06 bc	94.42 ± 0.06 no	1.02 ± 0.01 cdef	23.93 ± 0.3 cdefghi	21.57 ± 0.25 efghijkl	71.82 ± 0.33 ghij
Rimski okrugli	100	15	9.11 ± 0.1 p	90.89 ± 0.1 a	0.97 ± 0.03 abcd	23.71 ± 0.31 cdefgh	20.58 ± 0.27 efghijkl	69.34 ± 0.21 a
		30	8.01 ± 0.04 no	91.99 ± 0.04 bc	1.24 ± 0.09 i	22.38 ± 0.38 a	19.34 ± 0.26 fghijkl	71.41 ± 0.39 efghi
		45	7.71 ± 0 no	92.29 ± 0 bc	0.91 ± 0.06 abc	22.64 ± 0.29 ab	19.99 ± 0.21 ghijkl	71.4 ± 0.26 efghi
		60	7.81 ± 0.08 no	92.19 ± 0.08 bc	0.98 ± 0.06 abcd	23.8 ± 0.15 cdefgh	20.96 ± 0.1 ghijkl	70.25 ± 0.08 abcd
	120	15	7.69 ± 0.03 n	92.31 ± 0.03 c	0.98 ± 0.03 abcd	23.4 ± 0.14 bcde	20.62 ± 0.15 hijkl	70.71 ± 0.15 bcdef
		30	7.71 ± 0.07 no	92.29 ± 0.07 bc	0.86 ± 0.01 ab	23.59 ± 0.1 cdefg	20.91 ± 0.07 hijkl	70.52 ± 0.14 bcde
		45	6.58 ± 0.04 hij	93.42 ± 0.04 ghi	1.05 ± 0.04 cdefgh	23.06 ± 0.1 abc	20.5 ± 0.13 hijkl	71.87 ± 0.13 ghij
		60	6.96 ± 0.04 klm	93.04 ± 0.04 def	1.01 ± 0.03 bcdef	24.7 ± 0.2 ijk	21.97 ± 0.15 hijkl	70.06 ± 0.22 abcd
	140	15	7.02 ± 0 m	92.98 ± 0 d	1.01 ± 0.04 bcdef	24.08 ± 0.04 defghij	21.38 ± 0 ijkl	70.59 ± 0.03 bcdef
		30	6.47 ± 0.02 ghi	93.53 ± 0.02 hij	1.18 ± 0.06 ghi	24.32 ± 0.22 fghij	21.57 ± 0.27 ijkl	70.79 ± 0.19 bcdef
		45	6.14 ± 0.02 efg	93.86 ± 0.02 jkl	1.06 ± 0.07 cdefgh	24.35 ± 0.14 fghij	21.79 ± 0.21 ijkl	71.01 ± 0.12 defgh
		60	5.98 ± 0.15 def	94.02 ± 0.15 klm	0.84 ± 0.06 a	23.87 ± 0.46 cdefghi	21.61 ± 0.34 jkl	71.58 ± 0.54 fghi
	160	15	5.93 ± 0.12 de	94.07 ± 0.12 lm	1.16 ± 0.06 fghi	24.53 ± 0.28 hij	21.92 ± 0.3 klm	70.99 ± 0.36 defgh
		30	5.74 ± 0.03 cd	94.26 ± 0.03 mn	1.05 ± 0.06 cdefgh	23.41 ± 0.1 bcde	21.02 ± 0.16 klm	72.19 ± 0.07 ij
		45	5.24 ± 0.11 a	94.76 ± 0.11 p	1.01 ± 0.01 bcdef	23.22 ± 0.1 abcd	21 ± 0.06 lm	72.75 ± 0.18 jk
		60	5.02 ± 0.02 a	94.98 ± 0.02 p	1.06 ± 0.01 cdefgh	22.58 ± 0.04 ab	20.39 ± 0.05 m	73.54 ± 0.05 k
Statistical significance			n.s.	n.s.	n.s.	n.s.	n.s.	n.s.

DM—Dry matter; FC—Fixed carbon; VM—Volatile matter. Tukey’s HSD test identified statistically significant differences (p < 0.05) between means in the same column with distinct lowercase superscripts.

.

After the statistical analysis, the changes in proximate analysis of the observed hazelnut samples were determined. For the Istarski duguljasti variety, the highest moisture content (9.29%) was measured at 100 °C and 15 min, while the lowest content (5.36%) was measured at 160 °C and 45 min. The lowest values of DM (90.71%), FC (20.32%), and VM (69.34%) were observed at 100 °C and 15 min of process time, while the highest values of the mentioned variables (94.64%, 22.79%, and 71.89%) were recorded at 160 °C for 15 or 45 min. The ash and coke values increased with the application of higher process temperatures. For example, the lowest ash value (0.84%) was noted at 120 °C and 60 min of process time, and the highest (1.20%) was recorded at 140 °C and 60 min. The lowest average coke content was 23.13% (100 °C and 60 min) and the highest was 25.26% (160 °C and 15 min).

In addition, for the tested sample, the moisture content of the Rimski okrugli was highest at 100 °C and 15 min (9.11%), while the lowest value decreased with increasing temperature (5.02% at 160 °C and 60 min). The lowest average percentages of DM (90.89%), coke (22.38%), FC (19.34%). and VM (69.33%) were noted at a temperature of 100 °C for a duration of 100 min, and the highest percentages measured were recorded at 120 °C and 160 °C for 60 min. The average ash content was highest at 100 °C and 30 min in process duration (1.24%) and lowest at 140 °C and 60 min in process duration (0.84%). In the analysis of variance (ANOVA), the differences between the samples were not statistically significant.

Hazelnuts of the varieties Istarski duguljasti and Rimski okrugli were dried at different temperatures and durations, and the resulting ultimate analysis is shown in

Table 5. Ultimate and calorimetric analysis of observed hazelnut varieties after drying treatment.

Sample	Temp. (°C)	Time (min)	N (%)	C (%)	S (%)	O (%)	H (%)	HHV (MJ kg−1)
Istarski duguljasti	100	15	0.31 ± 0.05 a	54.88 ± 0.91 a	0.07 ± 0.01 a	37.73 ± 0.98 def	7 ± 0.02 jklmn	20.4 ± 0.19 fghi
		30	0.39 ± 0 a	54.86 ± 0.51 a	0.07 ± 0 a	37.98 ± 0.45 def	6.7 ± 0.05 efghijklm	20.29 ± 0.02 defgh
		45	0.31 ± 0.02 a	55.21 ± 0.37 ab	0.07 ± 0 a	37.69 ± 0.35 def	6.73 ± 0 efghijklm	20.75 ± 0.01 jkl
		60	0.19 ± 0.03 a	55.01 ± 0.21 a	0.06 ± 0 a	38.92 ± 0.19 f	5.81 ± 0.05 a	20.59 ± 0.02 hijkl
	120	15	0.22 ± 0 a	55.5 ± 0.23 ab	0.06 ± 0 a	37.92 ± 0.68 def	6.29 ± 0.46 bcde	20.36 ± 0.06 efghi
		30	0.33 ± 0.12 a	56.18 ± 0.02 abcd	0.09 ± 0 a	36.81 ± 0.08 bcdef	6.59 ± 0.03 defghij	20.82 ± 0.06 kl
		45	0.26 ± 0.01 a	56.16 ± 0.19 abcd	0.72 ± 0.09 b	36.25 ± 0.3 abcd	6.61 ± 0.01 defghijk	20.57 ± 0.02 hijkl
		60	0.29 ± 0.01 a	55.76 ± 0.12 abc	0.07 ± 0 a	36.83 ± 0.08 bcdef	7.06 ± 0.21 klmn	20.75 ± 0.01 jkl
	140	15	0.23 ± 0 a	55.42 ± 0.01 ab	0.07 ± 0.01 a	37.48 ± 0.02 def	6.8 ± 0.02 fghijklm	20.49 ± 0 ghij
		30	0.29 ± 0.01 a	56.47 ± 0.06 abcd	0.07 ± 0 a	36.26 ± 0.04 abcd	6.91 ± 0.02 hijklmn	20.81 ± 0.08 kl
		45	0.33 ± 0 a	56.05 ± 0 abcd	0.07 ± 0 a	37.1 ± 0.11 cdef	6.45 ± 0.11 cdefg	20.88 ± 0.2 l
		60	0.33 ± 0 a	56.3 ± 3.13 abcd	0.07 ± 0 a	36.74 ± 3.15 bcde	6.55 ± 0.01 cdefghi	20.47 ± 0.01 ghij
	160	15	0.35 ± 0.01 a	56.61 ± 0.28 abcd	0.09 ± 0 a	36.53 ± 0.28 abcde	6.42 ± 0.01 cdef	19.72 ± 0.09 ab
		30	0.36 ± 0 a	56.65 ± 0.1 abcd	0.07 ± 0 a	36.09 ± 0.1 abcd	6.83 ± 0.01 fghijklmn	19.54 ± 0.06 a
		45	0.42 ± 0.01 a	57.11 ± 0.12 bcd	0.08 ± 0 a	35.33 ± 0.12 abc	7.07 ± 0.01 lmn	20.13 ± 0.06 cdef
		60	0.34 ± 0.02 a	56.64 ± 0.01 abcd	0.08 ± 0 a	35.88 ± 0.01 abcd	7.06 ± 0 lmn	20.18 ± 0.01 cdefg
Rimski okrugli	100	15	0.2 ± 0 a	55.1 ± 0.05 a	0.07 ± 0.01 a	37.64 ± 0.46 def	6.99 ± 0.41 ijklmn	20.52 ± 0.17 hijk
		30	0.5 ± 0.03 a	55.54 ± 0.03 ab	0.07 ± 0 a	36.94 ± 0.03 bcdef	6.96 ± 0.03 hijklmn	20.63 ± 0.09 ijkl
		45	0.21 ± 0 a	55.2 ± 0 ab	0.06 ± 0 a	38.59 ± 0.22 ef	5.94 ± 0.22 ab	20.52 ± 0.02 hijk
		60	0.29 ± 0.01 a	56.19 ± 0.26 abcd	0.07 ± 0.01 a	36.4 ± 0.57 abcd	7.06 ± 0.33 lmn	20.66 ± 0.01 ijkl
	120	15	0.22 ± 0 a	56.22 ± 0.02 abcd	0.06 ± 0 a	36.73 ± 0.08 bcde	6.77 ± 0.06 fghijklm	20.73 ± 0.04 jkl
		30	0.23 ± 0.01 a	55.63 ± 0.06 abc	0.08 ± 0 a	37.53 ± 0.08 def	6.53 ± 0 cdefgh	20.66 ± 0.06 ijkl
		45	0.22 ± 0.01 a	56.27 ± 0.11 abcd	0.07 ± 0.01 a	37.21 ± 0.22 cdef	6.23 ± 0.11 abcd	20.73 ± 0.04 jkl
		60	1.23 ± 1.07 b	55.54 ± 0.09 ab	0.07 ± 0 a	37.02 ± 1.18 bcdef	6.14 ± 0.02 abc	20.6 ± 0.01 hijkl
	140	15	0.3 ± 0.01 a	55.6 ± 0.09 ab	0.07 ± 0 a	37.19 ± 0.13 cdef	6.84 ± 0.02 fghijklmn	20.65 ± 0.08 ijkl
		30	0.28 ± 0.01 a	55.99 ± 0.02 abcd	0.07 ± 0 a	37.09 ± 0.02 cdef	6.56 ± 0.01 cdefghij	20.46 ± 0.03 ghij
		45	0.27 ± 0.02 a	56.02 ± 0.02 abcd	0.07 ± 0 a	36.76 ± 0.04 bcde	6.88 ± 0.03 ghijklmn	20.87 ± 0.25 l
		60	0.24 ± 0 a	55.69 ± 0.08 abc	0.07 ± 0 a	37.34 ± 0.05 cdef	6.66 ± 0.02 defghijkl	20.75 ± 0.02 jkl
	160	15	0.3 ± 0.02 a	56.23 ± 0.36 abcd	0.07 ± 0.01 a	36.69 ± 0.37 bcde	6.72 ± 0.01 efghijklm	19.88 ± 0.02 bc
		30	0.25 ± 0 a	56.63 ± 0.03 abcd	0.07 ± 0.01 a	36.23 ± 0.04 abcd	6.82 ± 0 fghijklmn	20.45 ± 0.3 fghij
		45	0.33 ± 0 a	57.56 ± 0.12 cd	0.07 ± 0 a	34.93 ± 0.12 ab	7.11 ± 0 mn	20.06 ± 0.04 cde
		60	0.42 ± 0 a	57.84 ± 0.03 d	0.07 ± 0 a	34.42 ± 0.04 a	7.25 ± 0.02 n	20.02 ± 0.03 bcd
Statistical significance			n.s.	n.s.	n.s.	n.s.	**	n.s.

N—Content of nitrogen; C—Content of carbon; S—Content of sulfur; O—Content of oxygen; H—Content of hydrogen; HHV—Higher heating value. Tukey’s HSD test identified statistically significant differences (p < 0.05) between means in the same column with distinct lowercase superscripts. Statistical significance is denoted as follows: ** p ≤ 0.05, while the abbreviation n.s. signifies non-significance.

.

For the Istarski duguljasti variety, the highest average values of N (0.42%), C (57.11%), and H (7.07%) were recorded at 160 °C and 45 min, while the lowest percentages (0.19%, 54.86%, and 5.81%) were observed at a temperature of 100 °C for durations of 30 and 60 min, respectively. The value of S varied between 0.06 and 0.71%, with the lowest value noted at 120 °C and 15 min and the highest at the same temperature and a duration of 45 min. The percentage of O decreased on average with increasing temperature and duration of the process, so that the highest value was 38.92% (100 °C and 60 min) and the lowest was 35.33% (160 °C and 45 min). The heating value (HHV) was highest at 140 °C and 45 min (20.88 MJ kg⁻¹) and lowest at 160 °C and 30 min (19.54 MJ kg⁻¹).

The variables of the ultimate analysis of the Rimski okrugli variety changed similarly when different drying treatments were applied. The lowest percentage of concentration of N (0.20%) and C (55.10%) was found at 100 °C and 15 min. With increasing temperature and duration of drying, the percentage of N (1.23%) increased at 120 °C and 60 min and of C (57.84%) at 160 °C and 60 min. The percentage of sulfur varied between 0.05–0.08%, while the percentage of O decreased with increasing temperature and time (38.59% at 100 °C and 45 min; 34.42% at 160 °C and 60 min). The percentage of H increased with the application of higher temperature and longer duration of the process (5.94% at 100 °C and 45 min; 7.25% at 160 °C and 60 min). As with the previous sample, the highest HHV was found at 140 °C and 45 min (20.87 MJ kg⁻¹) and the lowest at 160 °C and 15 min (19.88 MJ kg⁻¹).

RSM (Response Surface Methodology) is a set of statistical and mathematical techniques that can be used to develop, improve, and optimize a process based on multiple independent variables that influence a particular response [21]. Figure 1 and Figure 2 illustrate the results of a response surface methodology (RSM) analysis that evaluates the relationship between the independent variables of drying time and temperature (plotted on the x and y axes) and the changes in the dependent variables of the ultimate and proximate analyses (plotted on the z axis). The RSM plots visually illustrate how variations in drying time and temperature affect the results of the ultimate and proximate analyses, with the colors of the response surface changing gradually from blue to yellow, indicating an increase or decrease in the proportion of a particular component. In this way, changes in process conditions can be tracked.

The changes in physical properties of hazelnuts, caused by variety, temperature, drying time, and their interactions, were described using univariate analysis, and the results are presented in

Table 6. Univariate analysis of changes in physical properties of hazelnuts as a function of sample parameters, drying time, and temperature.

Effect	DF	SS
		Len. (mm)	Wid. (mm)	Thick. (mm)	Den. (g/cm3)	Vol. (cm3)	FM (g)	KM (g)	SM (g)
Sample	1	478.07 *	188.47 *	3.69	0.01	0.11 *	0.00	0.10	0.06
Temp.	3	10.18	11.68	12.53	1.05 *	0.53 *	0.11	0.00	0.10
Time	3	11.49	4.04	10.48	1.17 *	0.19 *	1.58	0.38 **	0.64
Sample × Temp.	3	4.84	0.95	6.20	0.71 *	0.17 *	0.76	0.16	0.31
Sample × Time	3	1.77	0.19	6.20	1.61 *	0.37 *	0.28	0.01	0.24
Temp. × Time	9	46.44	9.04	27.70	5.98 *	1.55 *	3.28	0.60	1.74
Sample × Temp. × Time	9	15.31	8.27	28.30	5.52 *	1.84 *	4.25 *	0.80 **	1.97 **
Error	64	184.63	97.88	147.19	3.33	0.71	12.31	2.78	6.33

DF—Degrees of freedom; SS—Sum of squares; Len—Length; Wid—Width; Thick—Thickness; Den—Density; Vol—Volume; FM—Fruit mass; CM—Kernel mass; SM—Shell mass. Statistical significance: * p ≤ 0.01; ** p ≤ 0.05.

.

Significant differences in length, width, and volume variables are influenced by the Sample parameter (different samples), with a coefficient of p ≤ 0.01. Temperature significantly affects the differences in density and volume, with a coefficient of p ≤ 0.01. The parameter Time of the drying process significantly affects changes in density and volume, with a coefficient of p ≤ 0.01, while the change in kernel mass is significantly affected, with a coefficient of p ≤ 0.05. The interactions of the parameters Sample × Temp., Sample × Time, and Temp. × Time significantly affect the change in density and volume of the studied samples, with a coefficient of p ≤ 0.01. The change in density, volume, and fruit mass of hazelnuts is significantly affected by the interaction between sample, temperature, and time (p ≤ 0.01), while the change in kernel and shell mass is affected, with a coefficient of p ≤ 0.05. Other parameters do not have a significant influence on the physical properties of hazelnuts.

The changes in proximate analysis of hazelnuts, caused by variety, temperature, drying time, and their interactions, were described using univariate analysis, and the results are presented in

Table 7. Univariate analysis of the change in proximate analysis of hazelnuts in relation to sample parameters, drying time, and temperature.

Effect	DF	SS
		Moisture (%)	DM (%)	Ash (%)	Coke (%)	FC (%)	VM (%)
Sample	1	0.15 *	0.15 *	0.00	5.25 *	4.11 *	5.15 *
Temp.	3	82.85 *	82.85 *	0.14 *	9.56 *	22.13 *	26.09 *
Time	3	16.02 *	16.02 *	0.04 *	3.21 *	0.60 *	19.85 *
Sample × Temp.	3	0.68 *	0.68 *	0.01	3.99 *	3.76	4.28 *
Sample × Time	3	0.63 *	0.63 *	0.07 *	2.70 *	2.92 *	2.33 *
Temp. × Time	9	3.00 *	3.00 *	0.20 *	11.19 *	8.74 *	15.04 *
Sample × Temp. × Time	9	0.92 *	0.92 *	0.44 *	6.50 *	5.06 *	5.71 *
Error	64	0.70	0.70	0.16	4.95	4.74	6.57

DF—Degrees of freedom; SS—Sum of squares; DM—Dry matter; FC—Fixed carbon; VM—Volatile matter; Statistical significance: * p ≤ 0.01.

.

At p ≤ 0.01, each parameter shows a significant influence on the change in the variables of the proximate analysis, namely moisture, DM, ash, coke, FC, and VM. However, for the hazelnut samples studied, the sample parameter shows no significant influence on the change in ash. The changes in ultimate analysis of hazelnuts, caused by variety, temperature, drying time, and their interactions, were described using univariate analysis, and the results are presented in

Table 8. Univariate analysis of the change in ultimate and calorimetric analysis of hazelnuts in relation to sample parameters, duration, and drying temperature.

Effect	DF	SS
		N (%)	C (%)	S (%)	O (%)	H (%)	HHV (MJ kg−1)
Sample	1	0.03	0.56	0.05 *	0.76	0.03	0.20 *
Temp.	3	0.13	33.68 *	0.12 *	48.62 *	1.86 *	7.23 *
Time	3	0.31 **	3.50 **	0.11 *	4.46 **	0.18 **	0.63 *
Sample × Temp.	3	0.23	1.96	0.12 *	3.42	0.56 *	0.13 *
Sample × Time	3	0.40 **	0.70	0.12 *	4.47 **	0.55 *	0.27 *
Temp. × Time	9	1.08 *	3.96	0.35 *	9.55 **	3.13 *	0.35 *
Sample × Temp. × Time	9	0.84 **	4.58	0.35 *	13.54 *	4.75 *	1.69 *
Error	64	2.34	23.06	0.02	28.25	1.22	0.62

N—Content of nitrogen; C—Content of carbon; S—Content of sulfur; O—Content of oxygen; H—Content of hydrogen; HHV—Higher heating value. Statistical significance: * p ≤ 0.01; ** p ≤ 0.05.

.

In the ultimate analysis, the change in N content is significantly affected by the Time of the process and by the Sample × Time and Sample × Temp. × Time interactions, with a coefficient of p ≤ 0.05. The change in carbon content in the observed hazelnut samples is significantly influenced by the application of Temperature (p ≤ 0.01) as well as the Time of the process (p ≤ 0.05). All observed parameters significantly affect the change in the percentage of S and HHV (p ≤ 0.01). The parameters Temp. and Sample × Temp. × Time (p ≤ 0.01), Time, and Temp. × Time (p ≤ 0.05) have a significant effect on the change in the percentage of O. The value of H is most affected by the parameter Time (p ≤ 0.05), Temp., and the interactions of Sample × Temp., Sample × Time, Temp. × Time, and Sample × Temp. × Time (p ≤ 0.01).

4. Discussion

Before drying, the average values of physical properties of the Istarski duguljasti variety were analyzed: length (25.93 mm), width (19.93 mm), thickness (16.36 mm), density (1.05 g/cm³), volume (0.96 cm³), fruit mass (3.85 g), kernel mass (1.78 g), and shell mass (2.07 g). With increasing temperature and drying time, the length value decreased so that the minimum value was 23.35 mm (at 160 °C and 60 min), while the maximum value was measured at 100 °C and 30 min (26.97 mm). Both the width and thickness of the sample also decreased by applying drying for different time periods. When analyzing the volume variables, it was found that the value decreased compared to the natural sample, except for drying at 100 °C and 60 min, where there was an increase in volume (1.65 cm³). The minimum values for volume, fruit, kernel, and peel mass were obtained at a temperature of 120 °C and a process time of 60 min. For the variety Rimski okrugli, the values of physical characteristics before treatment were as follows: length (21.41 mm), width (22.79 mm), thickness (15.92 mm), density (1.30 mm), volume (0.80 cm³), fruit mass (3.85 g), kernel mass (1.75 g), and shell mass (2.10 g). As in the previous variety, the minimum values of the variables of length, width, thickness, volume, and kernel mass were present when the temperature and the duration of the drying process were increased (160 °C at 60 min). In contrast to the Istarski duguljasti variety, the greatest reduction in fruit mass and skin mass was observed at temperatures of 100 and 120 °C for 15 and 45 min, respectively, while an increase was observed at 140 °C and 15 min. The physical properties of the hazelnut samples were significantly affected by sample, temperature, time, and their interactions, as shown in the study. Length, width, and volume were significantly affected by these factors, with temperature having a significant effect on density and volume and time having a significant effect on changes in density, volume, and kernel mass. The interactions between sample, temperature, and time also had a significant influence on the changes in density and volume. All three factors had a significant influence on the changes in density, volume, fruit mass, kernel mass, and skin mass. Moisture content decreased, while dry matter, coke, and volatile content increased with increasing temperature and drying time for both varieties studied. No other parameter was found to have a significant influence on the changes in physical properties of the hazelnut samples. The percentage of moisture content is gradually reduced by increasing the drying time and temperature [36,37]. The drying rate indicates the amount of water that is removed from the dry substance per unit area within a certain time [38]. For the Istarski duguljasti variety, the ash content was lowest at 120 °C and 60 min (0.84%), while the highest value (1.20%) increased with the application of a higher temperature (140 °C and 60 min). In contrast, for the Rimski okrugli variety, the highest ash content was obtained at 100 °C and 30 min (1.24%), while the lowest value (0.84%) was observed at a higher temperature (140 °C and 60 min). The study found that all parameters had a significant impact on the change in proximate analysis variables, such as moisture, dry matter, ash, coke, fixed carbon, and volatile matter. However, the sample parameter was not found to have a significant influence on the change in ash content of the observed hazelnut samples. The values of the ultimate analysis variables of nitrogen, carbon, sulfur, and hydrogen increased in parallel with the increase in temperature (maximum values were recorded at 160 °C and 120 °C at 45 and 60 min) and the duration of the process for both varieties tested. The percentage of oxygen and the upper heating value decreased with increasing temperature. The change in nitrogen content in the ultimate analysis was affected by time, and by the interactions of sample and time and sample, temperature, and time. The change in carbon content was influenced by the application of temperature and the time of the process. All parameters were found to have a significant impact on the changes in sulfur content and higher heating value. The parameters of temperature and time and the interactions of sample, temperature, and time had a significant effect on the change in oxygen content. The value of hydrogen was most affected by time, temperature, and their interactions with sample.

Planning for energy optimization in the drying process is critical to reducing costs and saving time. An effective energy optimization plan can help minimize energy consumption and reduce operating costs [39,40,41]. In addition, optimizing energy consumption can also improve the efficiency and speed of the drying process, which reduces the overall time to complete the task and increases productivity. Effective planning and implementation of energy optimization measures can result in significant cost savings and improved operational efficiency in the drying process.

5. Conclusions

In the study, the physical properties of two hazelnut varieties, Istarski duguljasti and Rimski okrugli, were analyzed before and after drying. Before drying, the Istarski duguljasti variety had an average length of 25.93 mm, a width of 19.93 mm, a thickness of 16.36 mm, a density of 1.05 g/cm³, a volume of 0.96 cm³, a fruit mass of 3.85 g, a kernel mass of 1.78 g, and a shell mass of 2.07 g. The variety Rimski okrugli had an average length of 21.41 mm, a width of 22.79 mm, a thickness of 15.92 mm, a density of 1.30 g/cm³, a volume of 0.80 cm³, a fruit mass of 3.85 g, a kernel mass of 1.75 g, and a shell mass of 2.10 g. With increasing temperature and drying time, the length of both varieties decreased, with a minimum value of 23.35 mm for Istarski duguljasti at 160 °C and 60 min and a minimum value of 18.98 mm for Rimski okrugli at 140 °C and 60 min. Both the width and thickness of the sample also decreased with increasing temperature and time. The volume decreased compared to the natural sample for both varieties, except when dried at 100 °C and 60 min for Istarski duguljasti and 140 °C and 15 min for Rimski okrugli, where an increase in volume was observed. The physical properties of the hazelnut samples were significantly influenced by several factors, including sample, temperature, time, and their interactions. These factors had a significant influence on the differences in length, width, and volume of the samples. In addition, temperature had a significant effect on the variations in sample density and volume, and the time parameter had a significant effect on the changes in sample density, volume, and core mass. Interactions between the sample, temperature, and time also had a significant effect on changes in density and volume. In addition, the interaction of all three factors had a significant effect on changes in density, volume, fruit mass, kernel mass, and skin mass. An increase in temperature and drying time led to an increase in dry matter, coke, and volatile content and a decrease in moisture content for both varieties. It is noteworthy that the sample parameter had no significant influence on the change in ash content in the observed hazelnut samples. The percentage of moisture content gradually decreased with increasing drying time and temperature. Ash content was affected by temperature and time, with the lowest value of 0.84% for Istarski duguljasti at 120 °C and 60 min and the highest value of 1.24% for Rimski okrugli at 100 °C and 30 min. The values of the ultimate analysis variables of nitrogen, carbon, sulfur, and hydrogen increased with the increase in temperature and duration of the process for both varieties tested. The percentage of oxygen and the higher heating value decreased with increasing temperature.