Optimizing Built-in Refrigerator Integration: BEHAVIOR Model for Evaluating Kitchen Workflow and Spatial Adaptability

Abstract

As ergonomic and user-centered kitchen design gains importance, integrating built-in appliances such as refrigerators has become common in modern households. However, spatial misalignment and circulation conflicts often disrupt kitchen routines. This study introduces the BEHAVIOR model (Behavioral Embeddedness Evaluation for Appliance-Versatile Integrated Operation Routing), a multidimensional framework for evaluating the movement path adaptability of embedded refrigerators in integrated kitchen–dining environments. The model identifies eight behavioral dimensions: Body Clearance, Embedded Compatibility, Handling Logic, Accessibility, Visual Feedback, Interaction Conflict, Operating Time, and Routing Simplicity, from a user–space–product coordination perspective. Expert-based AHP weighting and user entropy methods were combined to construct adaptability scores across five kitchen layouts (L-shaped, U-shaped, single-line, G-shaped, and island). The findings indicate that Routing Simplicity and Accessibility are the core determinants of layout adaptability, while Operating Time and Body Space show layout-dependent variations. Interaction Conflict and Embedded Compatibility are significantly influenced by spatial compactness. This research identifies key behavioral bottlenecks in kitchen workflows and presents a scalable model for appliance–space compatibility analysis, contributing to behavioral product evaluation and highlighting the role of user dynamics in design decisions.

1. Introduction

1.1. Research Background and Significance

With the shrinking of urban residential spaces, integrated kitchen–dining designs have emerged as a dominant trend in modern households. Built-in refrigerators, which are central to these layouts, are expected to grow significantly from a global market size of USD 8.3 billion in 2024 to USD 14.6 billion by 2032, at a compound annual growth rate (CAGR) of 6.5% [1]. In China, offline retail sales of built-in refrigerators have risen from 19.1% in 2023 to 23.1%, with projections indicating full-year penetration will exceed 30% by 2025 [2]. As a response to this trend, standardization systems are evolving, such as Midea’s 2025 refrigerator standard emphasizing “scenario-defined categorization” [3].

However, despite the growing popularity of built-in refrigerators, spatial incompatibilities in real-life installations, such as restricted door openings, user standing conflicts, and inefficient circulation paths, continue to disrupt daily kitchen routines. These issues lead to negative user experiences and installation failures [4]. Previous studies have demonstrated that even minor differences in kitchen layout can significantly alter pollutant dispersion and ventilation effectiveness, indirectly impacting users’ movement paths and spatial usage patterns [5]. However, current research mainly focuses on size standardization and structural parameters and lacks a behavioral systems perspective that considers the interaction between users, space, and products. As a result, dynamic misalignments in everyday kitchen use remain unresolved.

This study aims to construct and validate the BEHAVIOR model (Behavioral Embeddedness Evaluation for Appliance-Versatile Integrated Operation Routing), a multi-dimensional framework designed to assess the circulation adaptability of embedded refrigerators. The model integrates user movement, spatial constraints, and behavioral interactions into a unified framework. It addresses the theoretical gap in dynamic adaptability analysis by introducing a human-centered, real-time evaluation approach. Ultimately, this study contributes actionable metrics and methodological tools for optimizing appliance integration and layout planning in future kitchen designs.

1.2. Literature Review

1.2.1. Evolution of Embedded Refrigerators and Adaptation Challenges

The evolution of embedded refrigerators has progressed from standalone units to fully integrated “zero-protrusion” models. Milestones include the invention of the household refrigerator in 1918, the introduction of integrated kitchens in Frankfurt in 1926, and the later emergence of flush-panel, customizable, and structurally harmonized designs. Today’s models emphasize seamless aesthetics, compact depth, and standardized ventilation [6].

Based on installation style, embedded refrigerators fall into three categories: zero-clearance, semi-embedded, and reserved-space types. Zero-clearance models demand strict control of depth, hinge design, and bottom-ventilation clearances (≥4 mm side/top, ≥50 mm rear) [7]. Semi-embedded products allow door protrusion and offer greater flexibility (e.g., CN106369919A patent) [8]. Reserved-space models require precise cabinet design, including adequate clearance for opening angles and heat dissipation [9]. Frequent failures include insufficient cabinet space, obstructed drawers, and ventilation issues—often due to mismatched cabinet planning or neglected spatial standards.

1.2.2. User Movement and Kitchen Interaction Patterns

User circulation behavior plays a critical role in refrigerator placement and kitchen layout design. Early studies used observational and interview-based methods to analyze movement paths and task sequences, such as the classic “kitchen work triangle” (refrigerator–sink–stove) model that prioritizes shortest distance [10]. The refrigerator, sink, and stove are considered the three key nodes in the kitchen. These nodes should be connected by direct, unobstructed, and appropriately sized paths (the “kitchen work triangle”), which should neither be excessively long nor intersect with the primary circulation path. This triangle corresponds to the typical sequence of kitchen tasks: sourcing (refrigerator), washing (sink), preparation (countertop), and cooking (stove), with the dish being returned to the refrigerator after cooling. Therefore, in open-plan kitchens, the refrigerator should be placed near the entrance for quick access to groceries, with a nearby surface or buffer countertop on one side for immediate placement and sorting of items. Recent studies have highlighted practical design issues such as door interference and retrieval sequence mismatches, emphasizing that refrigerator design for elderly users must consider cognitive and physical differences that affect safe and efficient use [11,12].

With the rise of perceptual technologies, multimodal behavior tracking (e.g., eye tracking, video logging, heatmaps) is now used to model dynamic use. For instance, the EPIC-KITCHENS dataset [13] provides large-scale egocentric recordings of kitchen operations. Wearables such as the Smart-Badge [14] combine IMU, audio, and touch sensors to recognize actions, while real-time trajectory modeling using floor pressure and wrist sensors enhances layout optimization [15].

Artificial intelligence further advances behavior modeling. Multi-dimensional trajectory clustering [16] enables the prediction of personalized routines, while systems like CloudFridge [17] utilize sensor fusion to forecast user intentions and optimize interaction interfaces. Kiran et al. [18] emphasize how smart compact kitchen layouts can enhance space utilization and kitchen efficiency by using modular designs that allow for flexible and efficient appliance and storage placement, particularly in small living spaces.

1.2.3. Spatial–Product–Behavior Adaptation Frameworks

Traditional appliance evaluation focuses on performance and structure, neglecting user behavior and spatial adaptability. Emerging three-dimensional frameworks evaluate built-in refrigerators across space, product, and behavioral dimensions. The spatial dimension considers cabinet fit, ventilation design, and door swing angles—requiring adherence to guidelines such as 10–50 mm clearances and “work triangle” distances between appliances (1.2–2.7 m) [19,20].

The product dimension evaluates structural integrity, energy efficiency, and interface usability. Analytical methods such as the Analytic Hierarchy Process (AHP) [20] and the Function–Behavior–Structure (FBS) model [21,22] and semantic differential analysis [23] support multi-criteria evaluation.

The behavioral dimension emphasizes ease of access, error prevention, and habit matching. Studies highlight the need for streamlined interaction paths [24] and recommend adapting designs to user lifestyles via AI and IoT technologies [25]. Together, these findings emphasize the need for behavior-sensitive, scenario-specific appliance design frameworks.

2. Materials and Methods

To systematically evaluate the movement path adaptability of embedded refrigerators in integrated kitchen–dining spaces, this study developed the BEHAVIOR model—a multidimensional behavioral evaluation framework. The methodology includes three key parts: model development, weighting strategy, and experimental validation (see Figure 1).

2.1. BEHAVIOR Model Development

The BEHAVIOR model (Behavioral Embeddedness Evaluation for Appliance-Versatile Integrated Operation Routing) identifies eight behavioral dimensions, grouped into four core categories. Spatial Adaptability includes Body Clearance (B), which refers to the minimum radius required for movements like door opening, bending, and item retrieval, impacting ergonomic feasibility [26], and Embedded Compatibility (E), which assesses the spatial fit between the refrigerator and surrounding cabinetry, walls, or islands [6,7]. Operational Efficiency includes Handling Logic (H), which reflects the cognitive alignment between task sequence and user expectations, enhancing usability [27], and Accessibility (A), which measures how easily users reach handles, drawers, and control panels [28]. Routing Clarity covers Visual Feedback (V), evaluating the intuitiveness of operational cues and interface prompts [29], and Interaction Conflict (I), which captures spatial interference in multi-user scenarios, such as congestion or overlap [30]. Interaction Control includes Operating Time (O), the time needed to complete refrigerator-related tasks like food storage or retrieval [31], and Routing Simplicity (R), which assesses the clarity and efficiency of user movement paths within the kitchen [32]. These dimensions together enable a comprehensive evaluation of the embedded refrigerator’s performance in integrated kitchen layouts, addressing user-product-environment interaction from multiple perspectives.

Each dimension addresses a specific aspect of human-product-environment interaction (see Figure 2), which provides a comprehensive view of how embedded refrigerators perform within integrated kitchen layouts.

In open-plan kitchens, the workflow typically includes five functional zones: storage, food storage, cleaning, preparation, and cooking, with the refrigerator serving the storage function. Following this zoning sequence: B/E focuses on the storage area (refrigerator location) and its interface with the cleaning/preparation zones (clearance, ventilation, and door opening envelope); H/A evaluates the transitions and reachability between “storage → cleaning → preparation → cooking” (including the work triangle distance); V/I locates shared passages and evaluates the guidance and conflicts at door/drawer openings; O measures the time taken for placing and retrieving items starting from storage, transitioning to cleaning/preparation; and R assesses the simplicity of cross-zone paths from the entrance through storage to subsequent zones.

At the equipment level, embedded refrigerators and standalone refrigerators have significant differences in installation, ventilation, and clearance (see Figure 3). Embedded refrigerators require standard cabinet openings and must be flush with the surface, using front/bottom ventilation. They have lower requirements for the rear gap and side clearance but are sensitive to installation precision and constrained by the opening size for future replacements. In contrast, standalone refrigerators do not require openings and offer more flexible placement. Their bodies often protrude and rely on rear/side ventilation, requiring larger side clearance and door swing radius. They are prone to interference with adjacent appliances (e.g., dishwashers, drawers) during door operations and can create bottlenecks in narrow passages. Since the scale and schematic of this study are based on the assumption of standard cabinet openings and flush installation, this study focuses only on embedded refrigerators.

2.2. Evaluation Weighting Strategy

To ensure robustness and sensitivity to user preferences, the study employs a hybrid weighting method [33]. This combines expert-based Analytic Hierarchy Process (AHP) results with user entropy data. The expert panel consists of five professionals from relevant fields (kitchen planning, ergonomics, home appliance engineering, interior architecture, usability, etc.), with an average of 20 years of experience, drawn from both academia and industry. Before formal scoring, a brief calibration session was conducted to standardize the operational definitions of the eight dimensions and provide an AHP example. Each expert performed pairwise comparisons of the eight dimensions using the 1–9 Saaty scale, with individual judgment matrices meeting the CR ≤ 0.10 condition. Subsequently, the matrices were aggregated using the geometric mean of the elements, and the eigenvector corresponding to the maximum eigenvalue was extracted and normalized to obtain the expert weight vector. In the expert weighting process, five domain experts conducted pairwise comparisons of the eight dimensions using a 1–9 Saaty scale.

In the AHP method, the pairwise comparison matrix A of the experts is as follows:

$$A=\left(\begin{array}{cccc}{a}_{11}& a_{12}& \dots & a_{1n}\\ a_{21}& a_{22}& \dots & a_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ a_{n1}& a_{n2}& \dots & a_{nn}\end{array}\right)$$

(1)

Here, a_ij represents the relative importance of dimension i compared to dimension j. Next, the normalization process of the matrix is performed:

$${\prime }_{ij}={\displaystyle \frac{x_{ij}}{\sum _{j=1}^n{x}_{ij}}}$$

(2)

The maximum eigenvalue λ_max is calculated, and the consistency ratio (CR) is computed:

$$CI={\displaystyle \frac{{\lambda }_{\mathrm{m}\mathrm{a}\mathrm{x}}-n}{n-1}}CR={\displaystyle \frac{CI}{RI}}$$

(3)

The expert’s weight vector Wexpert is obtained by normalizing the eigenvector. The results showed the highest weights for Routing Simplicity, Accessibility, and Handling Logic.

For the user side (entropy method), the survey scores were collected using an online questionnaire platform (Wenjuanxing). After standardizing the scores to a 0–1 scale, user weights for each layout type were calculated separately. Thus, each layout type received a weight vector containing 8 dimensions and their respective rankings. The entropy-based method evaluates the discriminability of each dimension within the same layout type’s respondents: the higher the discriminability, the larger the weight; the more consistent the scores, the smaller the weight. The weight vectors for each layout type were then normalized so that their sum equals 1.

After normalizing the user scores, the entropy value E_j for each dimension j is calculated using the following formula:

$$p_{ij}={\displaystyle \frac{x{\prime }_{ij}}{\sum _{i=1}^{n}x{\prime }_{ij}}}$$

(4)

$$E_j=-{\displaystyle \frac{1}{\ln (n)}}\sum _{i=1}^n{p}_{ij}\ln (p_{ij})$$

(5)

Here, p_ij is the probability of each score, and n is the number of respondents. The weight calculation formula based on entropy is:

$$W_j={\displaystyle \frac{1-E_j}{\sum _{j=1}^m(1-E_j)}}$$

(6)

Finally, a composite weighting model was constructed by combining expert and user weights using linear fusion:

$$W_{\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}=0.6\cdot W_{\mathrm{e}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{t}}+0.4\cdot W_{\mathrm{u}\mathrm{s}\mathrm{e}\mathrm{r}}$$

(7)

This model balances theoretical rigor with empirical user feedback.

2.3. Experimental Design and Data Collection

Adults who use kitchens regularly were recruited through universities, communities, and social media. The inclusion criteria were ≥18 years of age and using a home kitchen at least once a week in the past year. Participants gave informed consent online. Before formal scoring, participants were informed about the “ergonomic hierarchy” in the kitchen (high-frequency: eye–chest; common-use: chest–waist; low-frequency: below waist/top of head) as a vertical reference for storage and reachability. Based on this hierarchy, subjective ratings for accessibility (A) and body clearance (B) were completed. This hierarchy served solely as a scoring anchor and calibration reference and was not included as an independent metric in the calculations and reporting. Attention checks and minimum response time thresholds were set for quality control, and responses that did not meet the criteria were excluded. A total of 165 valid responses were obtained, with 56.7% females and 43.3% males; the average age was 30 ± 5 years; most participants were primarily composed of corporate employees and full-time homemakers and other professional jobs; 90.0% of the participants were right-handed. Five typical kitchen layouts—L-shaped, U-shaped, single-wall, G-shaped, and island—were selected for evaluation (see Figure 4). All five evaluated layouts are specified as open kitchens (integrated with dining/living areas), ensuring comparisons under a consistent spatial paradigm.

To collect data, three complementary methods were used: path sketches (participants marked their movement paths on a floor plan to extract objective indicators such as path length, frequency of turns, backtracking ratio, and door interference); behavioral task lists (participants performed a simplified meal preparation task in their own kitchen, recording collisions/obstructions and difficulties with access/clearance in real time); and retrospective surveys (collecting qualitative feedback on user experiences, such as uncomfortable postures, appliance proximity, and perceived distances in the work triangle). Subsequent processing followed: the scale scores were directly used as input for calculating the user-side weights, while path sketches and behavioral records were solely used for visualizing results, identifying bottlenecks, and validating the work triangle. When necessary, indicators from different sources were standardized to the same scale for descriptive comparison. The kitchen setup in the path sketches included a refrigerator, sink, and induction cooktop, with standard modular dimensions commonly used in residential design (refrigerator 600 mm, sink ≈ 800 mm, countertop 600 mm). The minimum work-aisle clearance referred to the International Plumbing Code (IPC 2024) [34], which specifies at least 1015 mm (40 in.) between opposing cabinets or appliances. A 1200 mm aisle was therefore adopted as the baseline for path mapping and behavioral sequence analysis to test its adequacy under weekend-intensity conditions.

We conducted ergonomic baseline checks using human body envelopes to verify circulation/maneuvering and operation/reach: circulation width and turning were assessed using 95th-percentile male body dimensions, while reach was checked across 5th-percentile female to 95th-percentile male envelopes. These checks were used only as rating anchors/calibration for A (Accessibility) and B (Body Clearance) and were not included as additional metrics in the evaluation.

3. Results

3.1. Evaluation Results of Kitchen Layouts

The adaptability scores of five kitchen layouts, based on the eight dimensions of the BEHAVIOR model, are shown in the radar chart (see Figure 5). In the L-shaped layout, path simplicity (R) is a key focus, though movement efficiency is a shortcoming. The U-shaped layout excels in accessibility (A) and operational time (O), with a high-efficiency work area, though bodily space (B) is limited. The single-wall layout stands out in operational time (O) due to its linear flow, but interaction conflicts (I) are significant. The G-shaped layout is constrained by its dense structure, limiting bodily space (B) and negatively affecting interaction conflicts (I) and embedding compatibility (E). The island layout scores highest in operation logic (H), optimizing workflow, but path simplicity (R) is compromised due to detours.

The IPA (Importance–Performance Analysis) decision matrix was used to cross-analyze user satisfaction with the importance of each dimension [35]. This analysis identified priority areas for improvement, strength maintenance, low-priority areas, and over-attention areas (see Figure 6). The priority improvement area (I) includes accessibility (A) and path simplicity (R) in the L-shaped layout, and accessibility (A) and operational time (O) in the U-shaped layout. The strength maintenance area (II) includes operational time (O) in the L-shaped layout and bodily space (B) and accessibility (A) in the G-shaped layout. The low-priority area (III) includes embedding compatibility (E) and visual feedback (V) across various layouts, while the over-attention area (IV) includes bodily space (B) and interaction conflicts (I) in the island layout.

Follow-up data and interviews further validated the findings from the IPA decision matrix. Some dimensions, despite having high scores in the questionnaire, still exhibited issues. For example, operational time (O) in the L-shaped layout and path simplicity (R) in the one-wall layout were reported as cumbersome. Users also noted that body space (B) and accessibility (A) in the G-shaped layout were insufficient during high-intensity use, leading to inefficiency. These issues confirmed the areas identified for improvement in the IPA matrix.

3.2. Weight Distribution and Key Influencing Factors

3.2.1. Expert Subjective Weighting

The AHP method was used to assign weights to the criteria of embedded refrigerator circulation adaptability. Significant differences were observed in the weight distribution of the criteria across different kitchen layouts (see

Table 1. Comparison of Five Layout Criteria Layer Weights.

Layout Type	Spatial Adaptability	Operational Adaptability	Path Clarity	Interference Control
L-shaped	0.133	0.273	0.145	0.4489
U-shaped	0.1167	0.4359	0.1403	0.3071
Single-wall	0.149	0.2933	0.1333	0.4243
G-shaped	0.4488	0.2244	0.191	0.1357
Island	0.1411	0.3298	0.1411	0.388

). The L-shaped and linear layouts prioritized interference control, while the U-shaped layout emphasized operational adaptability. The G-shaped layout focused on spatial adaptability. These findings indicate that the layout form directly determines design priorities, with linear layouts needing to address operational conflicts, while surrounding layouts should focus on efficiency and space optimization.

The comprehensive weight indicator ranking of the five kitchen layouts revealed both common needs and distinctive characteristics (see

Table 2. Comprehensive weight of evaluation indicators for the BEHAVIOR model under AHP.

Sort	L-Shaped		U-Shaped		Single-Wall		G-Shaped		Island
1	R	0.2992	A	0.3295	R	0.2735	B	0.402	O	0.2301
2	A	0.1594	O	0.1947	A	0.2008	I	0.1644	H	0.2199
3	O	0.1496	R	0.1124	O	0.1509	A	0.1496	R	0.1579
4	H	0.1136	H	0.1064	B	0.1178	R	0.086	A	0.1099
5	B	0.087	B	0.0998	H	0.0925	H	0.0748	B	0.0941
6	I	0.0826	I	0.0961	V	0.0791	O	0.0497	I	0.0894
7	V	0.0624	V	0.0443	I	0.0542	E	0.0468	V	0.0516
8	E	0.046	E	0.0169	E	0.0312	V	0.0266	E	0.047

). Path simplicity (R) was consistently ranked among the top three for the L-shaped, linear, and island layouts, while accessibility (A) was ranked in the top two for the U-shaped, linear, and island layouts. Layout-specific needs showed notable differentiation: the U-shaped layout requires improved accessibility, the G-shaped layout needs sufficient body space, and the island layout needs to reduce operational time.

3.2.2. User Objective Weight

To evaluate the performance of embedded refrigerators across different kitchen layouts, a user-centered evaluation system was developed [36]. A multi-dimensional scoring matrix for the five kitchen layouts across the eight BEHAVIOR dimensions was calculated (see

Table 3. Average rating matrix of five kitchen layouts in eight dimensions.

Dimension	B	E	H	A	V	I	O	R
L-shaped	6.273	7.394	6.424	6.924	6.970	5.561	7.682	7.409
U-shaped	5.273	7.439	5.439	6.030	7.470	3.439	6.424	6.030
Single-wall	8.409	7.636	5.697	9.091	6.652	5.348	4.788	6.288
G-shaped	6.485	7.318	8.576	7.015	8.379	4.439	7.500	7.485
Island	9.273	7.379	8.394	9.742	6.409	8.364	8.015	7.697

). The island layout demonstrated the best performance in spatial utilization and operational smoothness, with higher scores in operational space, accessibility, and reduction of interaction conflicts. The G-shaped layout excels in operational logic and visual feedback but requires optimization of interaction conflicts. The single-wall layout excels in spatial efficiency but needs further improvement in operational efficiency. The U-shaped layout shows weaknesses in interaction conflicts and operational logic. The L-shaped layout is well-balanced but lacks dominant dimensions.

The entropy method was applied to calculate the weights for each indicator in the BEHAVIOR model. The weights were derived based on entropy values and standardized data (see

Table 4. Weight of evaluation indicators for BEHAVIOR model using the entropy method.

Sort		B	E	H	A	V	I	O	R
L-shaped	Information entropy value	0.939	0.965	0.962	0.943	0.979	0.918	0.968	0.936
	Information utility value	0.0619	0.035	0.038	0.057	0.021	0.082	0.032	0.064
	weight coefficient	0.1579	0.090	0.099	0.146	0.053	0.210	0.081	0.163
U-shaped	Information entropy value	0.958	0.960	0.971	0.959	0.973	0.977	0.944	0.975
	Information utility value	0.042	0.040	0.029	0.041	0.027	0.023	0.056	0.024
	weight coefficient	0.150	0.142	0.103	0.144	0.095	0.080	0.199	0.087
Single-wall	Information entropy value	0.928	0.975	0.965	0.946	0.974	0.945	0.956	0.971
	Information utility value	0.072	0.025	0.035	0.055	0.026	0.055	0.044	0.029
	weight coefficient	0.212	0.073	0.102	0.161	0.076	0.161	0.130	0.084
G-shaped	Information entropy value	0.968	0.973	0.961	0.956	0.935	0.912	0.967	0.968
	Information utility value	0.032	0.027	0.039	0.044	0.065	0.088	0.033	0.032
	weight coefficient	0.088	0.076	0.108	0.123	0.180	0.244	0.092	0.088
Island	Information entropy value	0.897	0.943	0.937	0.716	0.939	0.968	0.946	0.962
	Information utility value	0.103	0.057	0.063	0.285	0.060	0.032	0.054	0.038
	weight coefficient	0.148	0.082	0.091	0.4115	0.0875	0.046	0.079	0.055

). This method minimizes subjective influences and ensures that the evaluation results are objective and reflective of the actual conditions.

To assess the accuracy of the scale in measuring its intended objectives, validity analysis was conducted. Prior to performing factor analysis, the KMO (Kaiser–Meyer–Olkin) and Bartlett’s tests were applied. A KMO value greater than 0.5 and a significance level (p-value) of Bartlett’s test less than 0.05 indicate that the questionnaire data are reliable and suitable for factor analysis [37]. In this study, the KMO value was 0.882, and the significance of Bartlett’s test was 0.000, confirming the suitability of the data for further analysis (see

Table 5. KMO and Bartlett’s Test.

Test	Value
KMO (Kaiser-Meyer-Olkin) Measure of Sampling Adequacy	0.882
Bartlett’s Test of Sphericity	Approximate Chi-Square	1673.782
	Degrees of Freedom	120
	Significance	0.000

).

A one-way analysis of variance (ANOVA) revealed significant differences (p-values < 0.001) across all 16 descriptions within the 8 dimensions of the BEHAVIOR model, indicating that these dimensions have a significant impact on user operations and experience (see

Table 6. One-Way ANOVA Results for Each Dimension of the BEHAVIOR Model.

Dimension	Describe	F	Significance
Body Clearance	Sufficient space when opening the refrigerator door	78.992	<0.001
	Sufficient space for body movement when retrieving items	95.590	<0.001
Embedded Compatibility	The refrigerator dimensions are well-matched with the cabinet/wall dimensions	95.431	<0.001
	The drawer and refrigerator door can be fully opened	112.605	<0.001
Handling Logic	The layout of storage compartments and control panel aligns with user habits	85.886	<0.001
	The control panel position is convenient for operation	58.225	<0.001
Accessibility	The position of the drawer and door handles is convenient	149.910	<0.001
	The internal height of the refrigerator is reasonable, allowing for smooth retrieval of items	95.214	<0.001
Visual Feedback	The refrigerator interface is clear and easy to recognize	25.298	<0.001
	The refrigeration/freezing zones are clearly defined, making items easy to locate	26.351	<0.001
Interaction Conflict	The usage path does not overlap with that of others	113.874	<0.001
	Simultaneous use by multiple people does not cause confusion	132.962	<0.001
Operating Time	The operation process is smooth, with a reasonable duration	96.591	<0.001
	The operation does not require repeated adjustments	37.310	<0.001
Requirement Indicator Layer	The refrigerator workflow is simple and direct, without detours	170.445	<0.001
	The retrieval path is clear and uncomplicated	59.944	<0.001

).

3.3. Bottlenecks and Scenario-Specific Findings

3.3.1. Spatial Circulation Path Visualization Analysis

Using path drawing and behavioral task forms, users marked their movement paths on floor plans and simulated performing standard tasks, thereby summarizing and categorizing user movement paths for five different layouts on both workdays and weekends (see Figure 7).

The results revealed that the L-shaped layout features relatively simple movement paths, but path conflicts occur at the corners. The U-shaped layout offers smoother movement paths; however, due to the concentration of work areas, users experience longer travel times between different workstations, and on weekends, the paths become more complex with intersections. The island layout is ideal for larger spaces, but in smaller spaces, the island creates path crossings, especially when multiple users are present. The single-wall layout has straightforward movement paths, but path overlap becomes significant with multiple users, reducing operational efficiency. The G-shaped layout provides more workspace, but its more complex movement paths require users to spend more time completing tasks.

Moreover, during weekends, the number of kitchen users increases, leading to more congested movement paths and longer travel times. This is particularly evident in U-shaped and G-shaped layouts, where path intersections become more complex. In contrast, on weekdays, fewer users result in simpler movement paths and fewer conflicts. These findings highlight the necessity of considering kitchen usage intensity and the number of users in layout design, as these factors significantly affect path efficiency and overall usability.

In summary, each layout presents movement bottlenecks and conflicts in different scenarios. When designing, it is crucial to consider the space size and number of users, optimize movement paths, and enhance operational efficiency.

3.3.2. Comprehensive Weighting and Evaluation Analysis

Based on expert subjective weight assignment using AHP and user objective weight calculation through the entropy method under the BEHAVIOR model, a linear aggregation method was applied to construct a unified dimension weight coefficient. By combining both subjective and objective weights, the combined weights for each evaluation index were calculated, and the ranking of these combined weights was determined (see

Table 7. Comparison of Eight Dimensions in the BEHAVIOR Model with Comprehensive Weights.

Sort	L-Shaped		U-Shaped		Single-Wall		G-Shaped		Island
1	R	0.2448	A	0.2840	R	0.2146	B	0.2943	A	0.3128
2	A	0.1540	O	0.2359	B	0.1979	I	0.2451	H	0.1862
3	I	0.1337	B	0.1500	O	0.1685	A	0.1636	O	0.1853
4	O	0.1223	H	0.1255	A	0.1437	V	0.1242	B	0.1455
5	B	0.1151	R	0.1198	I	0.1294	H	0.1098	R	0.1278
6	H	0.1076	I	0.1059	H	0.1169	R	0.1044	V	0.0835
7	E	0.0638	E	0.0955	V	0.0931	O	0.0848	I	0.0812
8	V	0.0586	V	0.0835	E	0.0627	E	0.0738	E	0.0777

).

The combined weight analysis based on the BEHAVIOR model reveals that different kitchen layouts exhibit specific design flaws and common bottlenecks. In L-shaped kitchens, path simplicity is a prominent issue, with complex traffic flow in corner areas reducing operational efficiency. G-shaped kitchens are constrained by limited bodily space, restricting physical movement. On a broader level, operational efficiency issues are common: U-shaped and island kitchens rank highest in accessibility, while U-shaped, single-wall, and island layouts have the highest operation times.

The spatial and logical dimensions highlight deeper needs: operation logic is more pronounced in island kitchens, while body space in single-wall and G-shaped kitchens ranks higher, indicating that the rationality of operational processes and activity space is a core requirement across all layouts. Visual feedback and embedding compatibility receive relatively lower weights, suggesting that users pay less attention to interactive visual information and equipment compatibility. Therefore, design optimization should focus on improving corner path efficiency, simplifying complex layouts, and addressing operational conflicts and ergonomic design flaws across kitchen types.

4. Discussion

4.1. Model Validity and Contribution

The BEHAVIOR model addresses the gap in the study of dynamic coordination between behavior and spatial constraints in kitchen appliances, providing a system-level framework that integrates human behavior and space. Unlike existing research, which focuses primarily on static parameters or generalized guidelines, this model offers a real-time, behavior-based evaluation. By proposing eight quantifiable dimensions, the model introduces new perspectives, such as Interaction Conflict (I), which quantifies spatiotemporal interference among multiple users, and Operation Duration (O), which uses task simulation to capture dynamic efficiency. These innovations provide a deeper ergonomic insight beyond traditional methods.

At the methodological level, the hybrid weighting approach, combining expert-derived Analytic Hierarchy Process (AHP) and user-based entropy weighting, improves the robustness and relevance of the evaluation. Results indicate that users are most sensitive to Path Simplicity (R), and that the Island layout receives higher weighting for Interaction Conflict (I) due to frequent path intersections. This dual-source data fusion bridges the gap between subjective preferences and empirical validation, enhancing the practical applicability of the model.

Furthermore, the full-process validation system (path drawing, behavioral simulation, and surveys) improves the identification of behavioral bottlenecks. The modular design allows the framework to be extended for evaluating other embedded appliances, thus advancing ergonomics research in residential equipment and offering actionable guidance for industry practices.

4.2. Design Implications and Practical Optimization

Based on the BEHAVIOR model’s multi-dimensional evaluation, this study proposes core strategies to optimize door opening and circulation path design. In L-shaped and Island layouts, high-frequency path conflicts (Interaction Conflict (I)) are addressed by implementing double-axis hinge designs and drawer-style doors. These solutions minimize the door’s rotation radius and reduce Operation Time (O), preventing collision risks in multi-user scenarios.

To systematically optimize kitchen circulation paths, this study incorporates universal ergonomic strategies—such as upward-opening cabinet doors and full-extension pull-out drawers—as the foundational design language. Building upon this foundation and guided by an in-depth BEHAVIOR model analysis of each layout, a set of customized design strategies is proposed (see Figure 8). For instance, supplementing the universal strategies, an integrated corner system in L-shaped kitchens specifically enhances Body Clearance (B) and Routing Simplicity (R); while for island kitchens dealing with path intersections, embedding an auxiliary refrigerator restructures the flow to improve Routing Simplicity (R) and reduce Interaction Conflict (I). Collectively, these strategies transform inherent layout disadvantages into advantages for behavioral adaptation, achieving a paradigm shift from static planning to dynamic behavioral optimization.

Additionally, the BEHAVIOR model drives three key ergonomic improvements that shift the industry from static compliance to dynamic, user-centered design. These include layout-responsive embedded solutions, such as asymmetric heat dissipation designs in L-shaped kitchens and shallow-deep cabinet systems in single-wall layouts, which address conflicts between Embedded Compatibility (E) and Body Clearance (B). Furthermore, task guidance light strips and voice priority systems, based on Visual Feedback (V) and Operation Logic (H), help reduce error rates. Lastly, the integration of pre-installed multifunctional sliding track interfaces enhances compatibility between dishwashers and ovens, fulfilling the modular requirements of “Scenario-defined Appliances”.

4.3. Limitations and Future Research

While this study provides an effective evaluation framework and design recommendations, there are limitations. The age distribution of the sample is imbalanced, with 80% of participants being middle-aged or younger, limiting the generalizability of the results, particularly for elderly users. Additionally, the simulated path method, relying on manually marked paths and task forms, struggles to capture spontaneous behavior conflicts in real-world scenarios, leading to limited sensitivity to low-frequency, abnormal events. The simulation path method based on user-drawn movement trajectories and task charts has limitations in capturing spontaneous conflicts and low-frequency abnormal events in real-world scenarios, with limited sensitivity to such phenomena. This study does not quantify the “ergonomic hierarchy” as an independent metric; future research could integrate reach zones/line of sight and usage frequency to construct measurable hierarchy compliance and examine its impact on efficiency and comfort. This study did not perform full digital human/kinematic simulations; human body envelopes were used for qualitative checks instead. Future research will incorporate digital human models to quantitatively assess maneuvering and reach margins.

Future research will focus on three directions:

Developing a VR-IoT integrated platform to simulate multi-user interaction conflicts in virtual kitchen environments, with dynamic adjustments for door opening angles and pressure-sensitive flooring to quantify the impact of Body Clearance (B) on microclimate factors such as ventilation efficiency.

Implementing an AI-based behavioral prediction system, trained on user data, to model habitual usage patterns (e.g., high-frequency retrieval zones), enabling pre-adjustment of refrigerator compartments and early warnings for conflicts, thus enhancing Operation Logic (H).

Expanding the BEHAVIOR model to other domains, such as bathroom HVAC systems and smart wardrobes, will facilitate a dynamic, pre-adaptive behavioral assessment.

5. Conclusions

This study constructs and validates the BEHAVIOR model to systematically uncover the behavioral adaptability mechanisms of embedded refrigerators within integrated kitchen environments. The empirical findings highlight that Path Simplicity (R) and Accessibility (A) are key determinants of spatial fluency in combined dining–kitchen spaces. The comparative analysis of layout types reveals that island kitchens exhibit heightened sensitivity to Interaction Conflict (I) due to intersecting movement paths, while U-shaped kitchens rely more on optimized Operation Logic (H).

Several practical design parameters were derived based on the BEHAVIOR model: optimization of door operation mechanisms to reduce Interaction Conflict (I), the formulation of a three-tier circulation buffer zone standard (core zone: 1.2 m, secondary path: 0.9 m, appliance clearance: 0.15 m) based on Path Simplicity (R), integration of task-guiding light strips to improve Visual Feedback (V), and the implementation of layout-responsive modular systems, such as asymmetrical ventilation schemes for L-shaped kitchens. These strategies provide a human-factor-oriented design framework that aligns with the “scenario-defined appliances” industry standard, facilitating a shift in residential kitchen design from static installation compliance to dynamic behavioral adaptability, ultimately promoting the co-evolution of user operational efficiency and spatial performance.