A Novel Approach to Identify Technological Innovation Opportunities Using Patent Mining for Floating Liquefied Natural Gas Systems

Abstract

The floating liquefied natural gas (FLNG) system is an offshore facility that floats above a natural gas field, directly liquefying natural gas without the need for subsea pipelines. In recent years, there has been growing interest in exploring technological innovations for FLNG systems. As such, advancements could lead to breakthroughs in optimizing layout and operations within the limited space of these platforms. To address this, we first apply a patent mining method to cluster FLNG-related patent texts, identifying the key technological components. We then conduct a morphological analysis to pinpoint potential technological opportunities. In our case study, we identify seven such opportunities, which include a combination of plate-fin heat exchangers, horizontal LNG storage tanks, flexible flowlines, and tail loading methods. These findings offer valuable insights and directions for the future development of FLNG systems.

1. Introduction

The floating liquefied natural gas (FLNG) system is a highly integrated offshore natural gas processing platform. It is capable of liquefying, storing and offloading natural gas directly from deepwater regions. The FLNG system is typically designed as a ship-shaped structure and is equipped with facilities such as natural gas liquefaction units and liquefied natural gas storage tanks. The detailed structure of the FLNG system is shown in Figure 1. This design enables the use of FLNG systems in deepwater areas, overcoming the technical challenges associated with laying subsea pipelines, while providing a cost-effective solution for developing marginal gas fields [1,2]. With the continuously changing global energy demands, FLNG technology is garnering increasing attention and offers broad development prospects [3,4].

The FLNG system shows great advantages owing to the integration of multiple processes (i.e., natural gas extraction, processing, liquefaction and storage). However, the space of the ship’s hull is limited, so it is a technical challenge for us to achieve a reasonable layout of all these functions. Therefore, it calls for novel methods to identify technological opportunities and thereby for the redesign of the size of the process units for a better layout.

To address these challenges, this study aims to explore innovative approaches to optimize the design and functionality of FLNG systems. By leveraging advanced technologies and methodologies, we seek to enhance the efficiency and feasibility of FLNG operations. Specifically, our research focuses on identifying technological gaps and opportunities within the current FLNG framework, proposing solutions that can lead to more compact and efficient designs. This will not only improve the operational performance of FLNG systems but also reduce costs and the environmental impact, making them more viable for future energy projects.

The remainder of this paper is organized as follows: Section 2 reviews previous studies related to patent mining and morphological analysis. Section 3 provides the main framework designed for identifying opportunities for FLNG. Section 4 presents a case study to validate the effectiveness of the proposed framework and conducts a comprehensive analysis of the results. Finally, Section 5 outlines the conclusions.

2. Literature Review

In Section 2.1, we analyze the existing technological gaps in current FLNG technologies and we have listed the discussed content in

Table 1. Current research on FLNG systems and identifying technological opportunities.

References	Kind	Main Contributions
Harada et al. [5]	FLNG System	Examining the environmental conditions limits for side-by-side offloading operations
Zhang et al. [6]	FLNG System	Considering the fully nonlinear simulation of liquid sloshing in FLNG tanks
Jin et al. [7]	FLNG System	Using the potential flow solver AQWA to study the hydrodynamics of FLNG-liquefied natural gas (LNG) offloading system
Jin et al. [8]	FLNG System	Using an unsteady Reynolds-Averaged Navier–Stokes solver, numerical investigations of hydrodynamic interactions of a conceptual FLNG-LNG offloading system in regular head sea waves are presented.
Chun & Kim [9]	FLNG System	Concluding that DWC can save steam costs while also reducing overall utility costs by comparing the simulation results of DWC using HYSYS with those of conventional distillation systems
Xu et al. [10]	FLNG System	investigating the influence of yaw motion on the hydrodynamic response of FLNG systems during side-by-side operations
Yang et al. [11]	Identifying Technological Opportunities	Proposing a new method based on International Patent Classification (IPC) for research institutions
Wang et al. [12]	Identifying Technological Opportunities	Applying text mining and an algorithm capable of clustering high-dimensional data objects to papers and patents on microalgae biofuels
Lee et al. [13]	Identifying Technological Opportunities	Customizing existing technologies and technical capabilities for small and medium enterprises through a two-stage patent analysis
Ma et al. [14]	Identifying Technological Opportunities	Proposing a hybrid method combining topic modeling, SAO semantic analysis, machine learning, and expert judgment
Hu et al. [15]	Identifying Technological Opportunities	Proposing a systematic approach to identify potential product opportunities by reflecting the target firm’s internal capabilities

. In Section 2.2 and Section 2.3, we discover that patent mining and morphological analysis are effective tools to identify technological innovation opportunities for FLNG systems. Therefore, we briefly review recent studies in the domain of patent mining and morphological analysis, and we summarize their main contributions in Table 2.

2.1. Previous Studies for FLNG Systems

Previous studies have explored various aspects of FLNG systems and technological opportunity identification. Representative works are summarized in Table 1. Harada et al. [5] investigated two-dimensional and three-dimensional gap resonances as well as the motion of floating bodies using numerical methods based on viscous and potential flows. They also examined the environmental condition limits for side-by-side offloading operations. Zhang et al. [6] conducted fully nonlinear simulations of liquid sloshing in FLNG tanks, demonstrating a high probability of accurately predicting the damping coefficient for nonlinear sloshing behavior. Jin et al. [7,8] used the potential flow solver AQWA to study the hydrodynamics of an FLNG and LNG offloading system in a side-by-side configuration. Their analysis included relative motions of the FLNG and LNG on the horizontal plane, as well as mechanical loads on mooring lines, fenders, and berthing equipment. Chun and Kim [9] compared simulation results of Dividing Wall Column (DWC) systems using HYSYS with those of conventional distillation systems, concluding that DWC could save steam costs while reducing overall utility expenses. Finally, Xu et al. [10] studied the influence of yaw motion on the hydrodynamic response of FLNG systems during side-by-side operations. Their findings showed that under liquid loading conditions, yaw motion can positively affect the motion of connecting systems and reduce induced loads.

As a prerequisite for technological innovation, the identification of technological opportunities is increasingly attracting widespread attention from experts and scholars both domestically and internationally. Yang et al. [11] proposed a new method based on the International Patent Classification (IPC) for research institutions to identify technological opportunities, validated through a case study on drones. However, their approach yields approximate results when analyzing sophisticated cutting-edge technologies, highlighting a gap in precision for advanced technological analysis. Wang et al. [12] applied text mining and an algorithm capable of clustering high-dimensional data objects on microalgae biofuel papers and patents. Although their analytical results provide an intellectual basis for constructing R&D strategies, they fall short of offering actionable strategies themselves, indicating a need for more practical and implementable outcomes. Lee et al. [13] customized existing technologies and technical capabilities for small and medium enterprises through a two-stage patent analysis. Nonetheless, their research subjects lack sufficient expertise, suggesting a limitation in the generalizability and depth of their findings. Ma et al. [14] proposed a hybrid method combining topic modeling, SAO semantic analysis, machine learning, and expert judgment to identify technological themes and potential development opportunities. However, their determination of the weight ratio between the title and the abstract as 2:1 based on experience, without analyzing the impact of different weight ratios, points to a methodological gap that could affect the robustness of their conclusions. Seo et al. [16] proposed a systematic approach to identify potential product opportunities by reflecting the target firm’s internal capabilities and evaluated the potential value of these opportunities. However, their focus solely on the technical aspect of product opportunities overlooks other critical factors such as market demand and competitive landscape, indicating a need for a more holistic approach.

In summary, existing studies have explored technological opportunities in various fields. However, most of them use descriptive methods that suffer from over-reliance on human experience. Moreover, they tended to focus more on the technical aspects, overlooking the constraints in practical applications. Therefore, it is particularly important to develop quantitative methods to explore technological opportunities for FLNG systems and assess whether they align with actual production.

Table 2. Literature summary for patent mining and morphological analysis.

References	Kind	Main Contributions
Hu et al. [15]	Patent Mining	Describing the patent mining problem of automatically discovering core patents
Han & Sohn [17]	Patent Mining	Enhancing patent valuation considering the risk of patent infringement
Milanez et al. [18]	Patent Mining	Proposing a method to classify nanomaterials into primary types and developed advanced patent indicators
Liu et al. [19]	Patent Mining	Identifying relevant technical phrases to summarize and represent patents from a technological perspective
Trappey, Charles V. et al. [20]	Patent Mining	Proposing a non-exhaustive overlapping clustering algorithm and applied it to clustering RFID (Radio Frequency Identification) patents
Xiao et al. [21]	Patent Mining	Providing a comprehensive study on the application of digital technology in decarbonizing shipping from 2005 to 2024 by analyzing 201 articles from the SCI-EXPANDED and SSCI databases.
Feng and Fu [22]	Morphological Analysis	Developing a patent text mining and informatics-based patent technological morphology analysis technique, validated through an empirical study on patents
Feng et al. [23]	Morphological Analysis	Proposing a hybrid method based on Morphological Analysis (MA) and Unified Structured Inventive Thinking (USIT)
Feng et al. [24]	Morphological Analysis	Suggesting a semi-autonomous and systematic procedure to extend the existing MA-based TRM and simplifying TRIZ application according to the occurrence frequency of the keywords.
Choi and Hwang [25]	Morphological Analysis	Analyzing patents in the fields of Light Emitting Diodes (LEDs) and Wireless Broadband
Guo et al. [26]	Morphological Analysis	Enhancing the performance of morphological analysis through integration with Subject–Action–Object (SAO) semantic analysis

Table 2. Literature summary for patent mining and morphological analysis.

References	Kind	Main Contributions
Hu et al. [15]	Patent Mining	Describing the patent mining problem of automatically discovering core patents
Han & Sohn [17]	Patent Mining	Enhancing patent valuation considering the risk of patent infringement
Milanez et al. [18]	Patent Mining	Proposing a method to classify nanomaterials into primary types and developed advanced patent indicators
Liu et al. [19]	Patent Mining	Identifying relevant technical phrases to summarize and represent patents from a technological perspective
Trappey, Charles V. et al. [20]	Patent Mining	Proposing a non-exhaustive overlapping clustering algorithm and applied it to clustering RFID (Radio Frequency Identification) patents
Xiao et al. [21]	Patent Mining	Providing a comprehensive study on the application of digital technology in decarbonizing shipping from 2005 to 2024 by analyzing 201 articles from the SCI-EXPANDED and SSCI databases.
Feng and Fu [22]	Morphological Analysis	Developing a patent text mining and informatics-based patent technological morphology analysis technique, validated through an empirical study on patents
Feng et al. [23]	Morphological Analysis	Proposing a hybrid method based on Morphological Analysis (MA) and Unified Structured Inventive Thinking (USIT)
Feng et al. [24]	Morphological Analysis	Suggesting a semi-autonomous and systematic procedure to extend the existing MA-based TRM and simplifying TRIZ application according to the occurrence frequency of the keywords.
Choi and Hwang [25]	Morphological Analysis	Analyzing patents in the fields of Light Emitting Diodes (LEDs) and Wireless Broadband
Guo et al. [26]	Morphological Analysis	Enhancing the performance of morphological analysis through integration with Subject–Action–Object (SAO) semantic analysis

2.2. Patent Mining

Patent identification is an effective approach for uncovering potential technological innovations. The rapid advancement of FLNG technology is driven by the support of core patents, which span various technical areas, including liquefaction, storage, and transportation. By conducting patent mining, it becomes possible to identify key technology patents related to FLNG and assess the technological strategies of companies within the FLNG industry. This process primarily involves analyzing and processing patent information to identify key or core patents from a vast dataset. Hu et al. [15] described the patent mining challenge of automatically discovering core patents, defined as novel and influential patents in a specific field. The significance of core patent mining is further highlighted by its ability to reveal potential competitive relationships between companies based on their core patents. For instance, Han and Sohn [17] utilized text mining to identify critical factors influencing patent value, such as lifespan, to improve patent valuation while accounting for the risk of infringement. Similarly, Milanez et al. [18] proposed a method to classify nanomaterials into primary categories and develop advanced patent indicators using text mining techniques to map technological developments. Liu et al. [19] studied the construction of technical profiles for patents by identifying relevant technical phrases to effectively summarize and represent patents from a technological standpoint. FLNG, as a global technology, involves both collaboration and competition among technology companies across various countries and regions. Using patent data clustering analysis (e.g., Trappey et al. 2010 [20]), FLNG-related patents from different regions can be grouped to uncover the competitive dynamics of global FLNG technology. For instance, analyzing the patent portfolios of different countries in the FLNG sector can help identify which nations or companies lead in FLNG technology and which are driving technological innovation. Such analysis is essential for shaping research and development strategies and gaining insights into industry technology trends. Xiao et al. [21] analyzed 201 articles from the SCI-EXPANDED and SSCI databases, providing a comprehensive study on the application of digital technology in decarbonizing shipping from 2005 to 2024. Their work outlines the current state, challenges, and future prospects of digital technology in this field.

In summary, existing research has explored various aspects of patent mining. Building on this foundation, we advance patent mining methods by performing cluster analyses on patent texts and presenting the findings through a morphological matrix. This innovative approach aims to accurately and intuitively capture key technological elements, thereby enhancing the efficiency of leveraging information to identify opportunities for technological innovation. These studies highlight the significant advantages of text mining, showcasing its ability to efficiently analyze vast amounts of data, uncover hidden patterns, and provide deeper insights that might otherwise remain unnoticed through traditional research methods.

2.3. Morphological Analysis

The morphological analysis method, which focuses on analyzing problems through morphological perspectives, has garnered extensive attention from experts and scholars worldwide since its inception. Feng and Fu [22] developed a patent text mining and informatics-based morphological analysis technique, validated through an empirical study of patents related to liquid crystal displays. The differences among results means it is useful to compare evaluation results and this provides more comprehensive information for decision-makers to aid the development of future patent technology development strategy. Feng et al. [23,24] proposed a hybrid method combining morphological analysis (MA) and Unified Structured Inventive Thinking (USIT). This approach was validated using coalbed methane (CBM) extraction technology, where a morphological matrix was employed to construct existing Technology Roadmaps (TRMs) by calculating correlations between various technology and product nodes. Additionally, two Sparse Generative Topographic Mapping (SGTM)-based maps were created to identify new technology and product opportunities by analyzing sparse areas to uncover development trends and innovation elements. However, the above-mentioned literature lacks a clear connection between opportunities and practical applications. Choi and Hwang [25] combined network-based and keyword-based patent analysis methods to study patents in the fields of light-emitting diodes (LEDs) and wireless broadband. However, due to the ambiguous boundaries between technical fields, it is not possible to collect patents from all relevant fields. Therefore, this problem suggests that a more refined patent extraction and classification process is necessary to ensure comprehensive data analysis, particularly in industries like FLNG, where technological convergence occurs frequently across various domains. Guo et al. [26] further enhanced the performance of morphological analysis by integrating it with semantic Subject–Action–Object (SAO) analysis, demonstrating the method’s potential to improve the depth and precision of patent analysis. This expansion is particularly relevant for the FLNG sector, where technological advancements are often interconnected and rapidly evolving. However, they may overlook certain technical details. Therefore, it is necessary to appropriately expand the morphology of the elements and introduce a method to calculate the weight of each element in order to analyze its importance in a specific technological opportunity.

In summary, previous studies have applied morphological analysis across various domains, including liquid crystal displays, CBM extraction technology, and text mining. Over time, continuous refinements have enhanced its effectiveness, solidifying its central role in identifying technological opportunities. Based on these, we introduce the integration of the entropy weight method into morphological analysis to analyze the obtained data, aiming to improve the reliability and precision of the results.

3. Methodology

We utilize patent mining to extract key elements in the target domain, such as gas processing, gas storage, gas transportation, and gas loading and unloading. These findings are organized into a morphological matrix and further analyzed using the entropy weight method. This approach enables the identification of novel solution combinations, which are then subjected to in-depth analysis to determine the optimal technological approach. Based on these findings, we develop a pathway for identifying technological opportunities within the innovation process, as illustrated in Figure 2.

3.1. Morphological Analysis Stage

The steps in this stage are as follows: ➀ The first step is to clearly delineate the scope of the research by focusing on the FLNG systems as the primary area of investigation. This ensures that the subsequent steps are aligned with the specific technological domain of FLNG. ➁ Based on the defined research area, targeted search expressions are formulated to retrieve relevant patents from patent databases. These search expressions are carefully designed to capture key technological concepts and innovations related to FLNG systems. ➂ The retrieved patents are subjected to clustering analysis to group them based on shared technological themes or features. This step helps in organizing the vast amount of patent data into manageable and meaningful categories. ➃ The clustering results are further refined to identify and extract key technological elements that are critical to the FLNG system. This step ensures that the most relevant and impactful aspects of the patents are highlighted for further analysis.

3.2. Patent Mining Stage

This stage involves the following steps: ➀ The extracted key elements from the patent mining stage are analyzed to categorize them into distinct groups. This categorization is based on their functional or technological characteristics within the FLNG system. ➁ A morphological matrix is developed to systematically organize the identified categories and their corresponding elements. This matrix serves as a structured framework for evaluating potential technological combinations. ➂ Factors such as technological feasibility, cost, and operational efficiency are incorporated to create a comprehensive evaluation index system. This system is customized to address the specific requirements of each element category. ➃ The entropy weighting method is employed to objectively determine the relative importance of each evaluation index within the element categories. This ensures a balanced and data-driven approach to weighting.

3.3. Evaluation Analysis Stage

This stage mainly involves the following steps: ➀ The weights assigned to each element category in the previous stage are organized and combined. Then, we list selectable options ranked in descending order of their total weight, providing a prioritized view of potential technological configurations. ➁ Each option in the morphological table is evaluated by comparing it with existing technologies. If a comparable technology is found, it is classified as a current trending technology. If no comparable technology exists, the feasibility of the combination is assessed. Feasible combinations are identified as new technological opportunities, while infeasible ones are discarded. This step ensures that only viable and innovative solutions are considered for further development.

4. Case Study

4.1. Patent Mining

With the rapid development of FLNG systems, an abundance of patent information resources has become available in this field, reflecting the growing technological advancements and innovations. For this study, we source data from the Patent Services Information Platform (http://vip.cnipr.com, URL (accessed on 17 July 2024)), a comprehensive and reliable database that provides access to a wide range of patent documents. By performing cluster analysis on these data, we systematically identify and categorize the constituent elements of FLNG systems, enabling a structured and in-depth exploration of the technological landscape. This approach ensures that the analysis is grounded in robust and relevant data, facilitating the identification of key technological trends and opportunities within the FLNG domain.

4.1.1. Define the Target Domain and Patent Retrieval

To ensure the universality of the experimental results, we retrieve all relevant patents across the entire timeframe. The final retrieval scheme is shown in

Table 3. FLNG patent retrieval scheme.

Search Platform	Patent Information Service Platform (http://vip.cnipr.com, URL (accessed on 17 July 2024))
Search Date	July 2024
Database Coverage	All patents
Timeframe	10 December 1902–14 December 2023 (By application date)
Search Query	Title, Abstract, Claims = ‘FLNG’

.

4.1.2. Patent Text Clustering

After patent retrieval, we obtain a total of 333 patents related to FLNG (including 31 utility model patents and 302 other patents). All retrieved keywords are listed in Figure 3. Since the retrieved patents contain multiple languages, in order to make the clustering results more accurate and effective, we decided to separately extract keywords from patents in different languages, select the most frequent keywords for each language, and finally perform clustering. Ultimately, by considering the proportion of patents in each language and the frequency of each keyword, we extracted 15 Chinese keywords, 15 English keywords, and 16 Korean keywords. The clustering result of these keywords is detailed in

Table 4. Keywords cluster obtained from FLNG-related patents.

Group	Clustering Results	Quantity	Proportion
1	‘外输设备’, ‘换热筒体’, ‘换热空腔’, ‘控制室’, ‘换热芯体’, ‘转盘式软管’, ‘浮式液化天然气’, ‘核动力’, ‘轻烃回收’, ‘转盘式刚性管’, ‘伸缩式刚性管’, ‘外输软管支撑’, ‘乙烷冷剂’, ‘转动装置’, ‘浮体’, ‘FLNG 换热器’, ‘offshore production’, ‘FLNG 船’, ‘Offshore production’, ‘flat wire’, ‘액화가스 운반선’, ‘구명정(10)’, ‘멤브레인형 저장탱크’, ‘천연가스’, ‘배출 드레인’, ‘fluid contact tray’, ‘외기 유입구’, ‘분류 칼럼 (fractionating column)’, ‘분류 칼럼’	30	63.04%
2	‘FLNG’, ‘FLNG facility’, ‘FLNG vessel’, ‘FPSO’, ‘ LNG-FPSO’, ‘Very Large Ore Carrier (VLOC)’, ‘liquefied gas carrier’, ‘부유식 액화천연가스 설비’, ‘부유식 생산설비’, ‘부유식 채취선’, ‘부유식 액화천연가스 생산설비’	11	23.92%
3	‘LNG’, ‘부유식 LNG 생산 저장 하역 설비’, ‘액화된 LNG 저장 제어 시스템’, ‘부유식 LNG’, ‘turret mooring system’, ‘Turret mooring system’	6	13.05%

.

4.1.3. Distinct Elements

From Table 4, it can be seen that the three most frequently appearing keywords in FLNG patents are LNG, FLNG facilities, and FLNG vessels. LNG refers to the gas stored in floating facilities rather than equipment, which is not in the aim range of our consideration. The concept of FLNG vessels is included in the keyword “FLNG facilities”. Therefore, we take FLNG facilities as the final considered keyword. In the clustering results, among the keywords related to floating liquefied natural gas facilities, considering the working steps of FLNG and the frequency of keyword appearances, four key elements are selected: FLNG heat exchangers, storage tanks, export equipment, and floating LNG handling systems. These correspond to the gas processing, storage, transportation and handling aspects of FLNG, respectively. Hence, we determine the final key elements as FLNG heat exchangers, storage tanks, export equipment, and floating LNG handling systems.

From Table 4, it is evident that the three most frequently appearing keywords in FLNG patents are “LNG”, “FLNG facilities”, and “FLNG vessels”. The term “LNG” refers to the gas stored in floating facilities rather than the equipment itself, which falls outside the scope of our focus. Additionally, the concept of “FLNG vessels” is encompassed within the broader term FLNG facilities. Therefore, we adopt FLNG facilities as the final keyword for consideration.

In the clustering results, keywords related to FLNG facilities were analyzed based on the working processes of FLNG and the frequency of keyword occurrences. From this analysis, four key elements were identified: “FLNG heat exchangers”, “storage tanks”, “export equipment”, and “floating LNG handling systems”. These elements correspond to the primary operational aspects of FLNG, encompassing gas processing, storage, transportation, and handling, respectively. Thus, the final key elements are determined to be “FLNG heat exchangers”, “storage tanks”, “export equipment”, and “floating LNG handling systems”.

4.2. Morphological Analysis

4.2.1. Explicit Element Morphology and Construction of Morphology Matrix Table

Before constructing the morphology matrix, it is essential to first clarify the morphology of each element. Through an analysis of the obtained patents, the morphology of each element is determined, and the morphology matrix is constructed accordingly. Refer to

Table 5. Element morphology based on four elements of a patent search.

Elements	FLNG Heat Exchanger	Storage Tank	Export Equipment	Floating LNG Handling System
Form	Plate-fin heat exchanger A1	Vertical LNG storage tank B3	Flexible loading hose C13	Tail loading method D18
	Shell-and-tube heat exchanger A2	Vertical LNG bullet tank B4	Static loading arm C14	Side-by-side mooring method D19
		LNG sphere tank B5	Loading arm C15
		Cylindrical LNG single-containment tank B6	Fixed transfer pip C16
		Cylindrical LNG double-containment tank B7	Flexible flowline C17
		Cylindrical LNG full-containment tank B8
		Cylindrical LNG membrane tank B9
		Cylindrical LNG triple-layer tank B10
		Atmospheric storage tank B11
		Horizontal LNG storage tank B12

.

4.2.2. Morphological Evaluation and Weighting

We then use the entropy weight method to process the indicator data of various morphological elements of FLNG.

Through an expert evaluation of element categories, four evaluation indicators can be determined: research input, technological advancement, economic benefits, and performance reliability. Subsequently, the expert scoring method is employed to rate each category on a scale of 1 to 10. Let $X_{ij}$ ($i=1,2,\dots ,m;j=1,2,\dots ,n$) denote the score of the jth indicator in the ith category, as shown in

Table 6. Expert scores of element categories.

	Indicator	Research and Development Investment	Technological Advancement	Economic Benefits	Performance Reliability
Form
Plate-fin heat exchanger A1		8.50	8.90	8.00	8.30
Shell-and-tube heat exchanger A2		7.50	8.00	8.70	8.30
Vertical LNG storage tank B3		9.00	8.90	8.50	8.00
Vertical LNG bullet tank B4		8.00	8.50	9.50	8.50
LNG sphere tank B5		8.50	7.80	8.00	8.50
Cylindrical LNG single-containment tank B6		7.00	7.50	6.50	7.00
Cylindrical LNG double-containment tank B7		7.50	5.50	7.00	7.50
Cylindrical LNG full-containment tank B8		5.00	7.00	7.60	8.00
Cylindrical LNG membrane tank B9		7.00	3.70	7.50	7.00
Cylindrical LNG triple-layer tank B10		7.90	4.50	7.60	5.20
Atmospheric storage tank B11		5.60	6.60	7.20	5.60
Horizontal LNG storage tank B12		9.20	8.70	8.80	9.00
Flexible loading hose C13		8.70	5.50	8.90	8.80
Static loading arm C14		6.70	5.60	9.00	8.40
Loading arm C15		8.70	7.80	9.20	8.40
Fixed transfer pip C16		7.00	9.00	9.40	8.20
Flexible flowline C17		9.00	8.40	7.90	9.60
Tail loading method D18		8.90	8.40	9.00	7.80
Side-by-side mooring method D19		9.00	7.80	8.20	9.10

.

Since data normalization plays a crucial role in ensuring the stability and accuracy of the results in the entropy weight method calculation, we normalized the data to eliminate the impact of different units and scales, thereby avoiding the potential distortion of weight calculations due to these differences.

By applying normalization, the influence of extreme values or outliers is minimized, leading to more reliable results. This enhances the reproducibility of the research method, as the results are not overly sensitive to variations in the original scale of the data. This step ensures that all variables contribute equally to the entropy calculation, allowing for a more balanced assessment. Additionally, it improves the scientific rigor of the method, as it can be consistently applied to different datasets without introducing bias due to varying data ranges.

Normalize the data using Formulas (1)–(3) as follows:

$$X_{ij}^{\prime }={\displaystyle \frac{Min{X}_i}{X_i}}$$

(1)

where $X_i$ refers to the original dataset of the $i_{th}$ evaluation indicator, while $Min{X}_i$ represents the minimum value of this indicator across all samples.

$$Y_{ij}={\displaystyle \frac{X_{ij}^{\prime }}{{\sum }_{i=1}^m{X}_{ij}^{\prime }}}$$

(2)

$$h_i={\displaystyle \frac{-{\sum }_{j=1}^n{Y}_{ij}ln{Y}_{ij}}{lnn}}$$

(3)

Then, compute the indicator weights $Y_{ij}$ and subsequently the weights based on $h_i$.

The calculation formula for weight $G{L}_i$ is

$$G{L}_i={\displaystyle \frac{1-h_i}{{\sum }_{i=1}^m(1-h_i)}}$$

(4)

Moreover,

$$0\le G{L}_i\le 1$$

(5)

$$\sum _{i=1}^{m}G{L}_i=1$$

(6)

After this process, we can obtain the weights for each morphology, as shown in

Table 7. Morphology weights of each morphology element.

GLA1	GLA2	GLB3	GLB4	GLB5	GLB6	GLB7	GLB8	GLB9	GLB10
0.055168	0.054223	0.055711	0.055799	0.054494	0.049985	0.049309	0.049004	0.045528	0.045836
GLB11	GLB12	GLC13	GLC14	GLC15	GLC16	GLC17	GLD18	GLD19
0.046450	0.056737	0.053117	0.051196	0.055521	0.054898	0.056043	0.055491	0.055489

.

Through the above analysis process, the proportions of FLNG heat exchangers, storage tanks, export equipment and floating LNG loading and unloading systems are determined, with $G{L}_i=1(i=1,2,\dots ,19)$ for each morphology.

4.3. Evaluation and Analysis

Using the data from Table 4, combinations of various forms of FLNG were generated, resulting in $2\times 10\times 5\times 2=200$ possible combinations. The weights ranged from a maximum of $0.2234$ to a minimum of $0.2064$. The top ten combinations were selected based on their total weights, as shown in

Table 8. Top ten combinations based on total weights.

Ranking	Form				Total GL
	FLNG Heat Exchanger	Storage Tank	Export Equipment	Floating LNG Handling System
1	Plate-fin heat exchange	Horizontal LNG storage tank	Flexible flowline	Tail loading method	0.223439
2	Plate-fin heat exchange	Horizontal LNG storage tank	Flexible flowline	Side-by-side mooring method	0.223437
3	Plate-fin heat exchange	Horizontal LNG storage tank	Loading arm	Tail loading method	0.222917
4	Plate-fin heat exchange	Horizontal LNG storage tank	Loading arm	Side-by-side mooring method	0.222915
5	Plate-fin heat exchange	Vertical LNG bullet tank	Flexible flowline	Tail loading method	0.222502
6	Plate-fin heat exchange	Vertical LNG bullet tank	Flexible flowline	Side-by-side mooring method	0.222500
7	Shell-and-tube heat exchanger	Horizontal LNG storage tank	Flexible flowline	Tail loading method	0.222494
8	Shell-and-tube heat exchanger	Horizontal LNG storage tank	Flexible flowline	Side-by-side mooring method	0.222492
9	Plate-fin heat exchange	Vertical LNG storage tank	Flexible flowline	Tail loading method	0.222414
10	Plate-fin heat exchange	Vertical LNG storage tank	Flexible flowline	Side-by-side mooring method	0.222412

.

According to Table 8, the combination with the highest weight consists of plate-fin heat exchanger, horizontal LNG storage tank, flexible flowline and tail loading method. Additionally, flexible loading hose and flexible flowlines can be used for the tail loading method or side-by-side mooring method, whereas static loading arm and loading arms are primarily used for the side-by-side mooring method. Vertical LNG tanks are typically used for onshore natural gas storage; their vertical structure makes them less stable under excessive bottom pressure, rendering them unsuitable for onboard transportation. Therefore, Options $1,2,3,5,6,7$ and 8 are feasible, while Options $4,9$ and 10 are not feasible. For these results, to verify whether the technological opportunities derived from this paper can be applied to actual production, we invited five experts from the product department of COSCO Shipping (Shanghai Port Group) to evaluate the outcomes. To ensure the accuracy of the evaluation, the invited experts were all engaged in the design of equipment related to heat exchangers, storage tanks, pipelines, and similar technologies. Analysis results can be seen in

Table 9. Evaluation analysis of each combination among the top ten schemes.

Option	Whether the Scheme Exists in Patents	Whether the Scheme Is Feasible	Reasons
1	No	Yes	Plate-fin heat exchangers and horizontal LNG tanks both exhibit excellent resistance to sway, while flexible flowlines are highly adaptable, capable of handling fluctuating offshore weather conditions. Additionally, they are suitable for gas transportation using tail loading method. The advantage of this method is that it has good seismic resistance and adaptability, allowing it to cope with more complex marine environments. However, the downside is that it requires a clean working environment; otherwise, more maintenance costs will be incurred. Additionally, the tail loading method is not yet technically mature.
2	No	Yes	Plate-fin heat exchangers and horizontal LNG tanks both exhibit excellent resistance to sway, while flexible flowlines are highly adaptable, capable of handling fluctuating offshore weather conditions. Additionally, they are suitable for gas transportation using side-by-side mooring method. The advantage of this method is that it has good seismic resistance and adaptability, enabling it to handle more complex marine environments. However, the downside is that using flexible flowlines may increase costs in the side-by-side mooring method.
3	No	Yes	Plate-fin heat exchangers have excellent sway resistance. Vertical LNG bullet tanks have large storage capacity, small volume, facilitating transportation and loading arm designs enable precise control over the transfer of liquids or gases, making them suitable for scenarios requiring accurate loading and unloading. They are also compatible with tail loading method. The advantage of this method is that it has lower costs and is technically mature. However, the downside is that it cannot adapt to harsher marine environments.
4	No	No	Loading arms are typically designed as single-point connections for transport and are generally not used for side-by-side mooring method.
5	No	Yes	Plate-fin heat exchangers have excellent sway resistance. Vertical LNG bullet tanks have large storage capacity, small volume, facilitating transportation, and flexible flowlines are highly adaptable, capable of handling variable offshore weather conditions, and suitable for gas transportation using tail loading method. The advantage of this method is that it can transport larger quantities of natural gas and that it is suitable for typical marine environments. However, the downside is that it has a high cost and the technology is not yet mature.
6	No	Yes	Plate-fin heat exchangers have excellent sway resistance. Vertical LNG bullet tanks have a large storage capacity and small volume, facilitating transportation, and flexible flowlines are highly adaptable, capable of handling variable offshore weather conditions, and suitable for gas transportation using the side-by-side mooring method. The advantage of this method is that it can transport larger quantities of natural gas and is suitable for typical marine environments. However, the downside is that using flexible flowlines may increase costs in the side-by-side mooring method.
7	No	Yes	Shell-and-tube heat exchangers can withstand high pressure, while horizontal LNG tanks exhibit excellent sway resistance. Flexible flowlines are highly adaptable, capable of handling variable offshore weather conditions, and they are suitable for gas transportation using the tail loading method. The advantage of this method is that it has a long service life, making it cheaper, and it can adapt to typical marine environments. However, the downside is that the tail-end delivery method is not yet mature, and it cannot adapt to harsher marine environments.
8	No	Yes	Shell-and-tube heat exchangers can withstand high pressure, while horizontal LNG tanks exhibit excellent sway resistance. Flexible flowlines are highly adaptable, capable of handling variable offshore weather conditions, and they are suitable for gas transportation using the side-by-side mooring method. The advantage of this method is that it has a long service life, making it cheaper, and it can adapt to typical marine environments. However, the downside is that it cannot adapt to more harsh marine environments.
9	No	No	Vertical LNG tanks are typically used for onshore natural gas storage. Their vertical structure makes them less stable under excessive bottom pressure, rendering them unsuitable for onboard transportation.
10	No	No	Vertical LNG tanks are typically used for onshore natural gas storage. Their vertical structure makes them less stable under excessive bottom pressure, rendering them unsuitable for onboard transportation.

.

In summary, a total of seven technological opportunities were identified. Regarding plate-fin and shell-and-tube heat exchangers, they both exhibited excellent heat transfer efficiency. Plate-fin heat exchangers offer superior sway resistance but are prone to blockages and have high manufacturing costs. Conversely, shell-and-tube exchangers are suitable for high-pressure conditions and have longer lifespans. However, their structure is relatively complex, making maintenance and cleaning more difficult. They are typically used in processes with cleaner mediums.

The design of the bullet tanks allows one or more auxiliary tanks to be installed above the main tank, enabling flexible adjustment of storage capacity. This design facilitates easier transportation, on-site installation, and accommodates various site conditions. However, the multi-tank structure of bullet tanks increases the complexity of connections and operations, resulting in higher manufacturing and installation costs. Horizontal LNG storage tanks, on the other hand, store LNG at the bottom of the tank, distributing pressure more evenly across the ground. These tanks are suitable for flat terrain and meet the stability requirements of FLNG installations.

Flexible flowlines can accommodate changes in terrain and irregularities on the seabed, making them more adaptable for underwater layout and connections. However, they are more vulnerable to physical and chemical impacts from the environment, requiring more frequent inspections and maintenance. Loading arms, which can move along multiple axes and angles, offer flexibility in adapting to various positions and angles of ships or equipment. However, the many moving parts in loading arms can lead to higher maintenance and inspection costs.

Based on the analysis above, the optimal technological opportunities involve a combination of plate-fin heat exchangers, horizontal LNG storage tanks, flexible flowlines, and the tail loading method, as illustrated in Figure 4. The next options to be considered, in a sequential order, are Options 2, 4, 5, 6, 7, and 8.

5. Conclusions

This paper proposes a pathway for identifying technological opportunities in the innovation process, utilizing patent mining and morphological analysis. The approach is applied to the FLNG system for practical analysis. Through a search and clustering analysis of FLNG patents, we identify key attributes, including “FLNG heat exchangers”, “storage tanks”, “export equipment”, and “floating LNG handling systems”. These attributes are then used to construct a morphological matrix. The entropy weight method, combined with expert scoring and calculations, is applied to determine the entropy weight values for each morphological structure. By organizing and combining these structures, this study generates new proposals, validating the feasibility of this approach. The combination of plate-fin heat exchangers, horizontal LNG storage tanks, flexible flowlines, and the tail loading method is identified as the optimal technological opportunity.

Based on the results of the case study, we drew the following conclusions: First, during the patent mining process, clustering analysis is performed on the retrieved patents, enabling a more comprehensive and efficient screening of attribute factors in the innovation process, such as “FLNG heat exchangers”, “storage tanks”, “export equipment”, and “floating LNG handling systems”. In the form analysis phase, the results of patent mining are presented in the form of a morphological matrix, and the entropy weight method is incorporated into the calculations. Finally, during the analysis phase, this study determines whether the new combination of solutions already exists, and a feasibility analysis is conducted to identify the optimal solution. This approach, combining patent mining, morphological analysis, and the entropy weighting method, not only enhances the comprehensiveness of data collection but also supports the objective and scientific prediction of technological opportunities, demonstrating its versatility and applicability.

Based on the above discussion, this paper proposes a method for identifying technological opportunities in FLNG based on patent mining and morphological analysis. The main contributions are as follows:

➀ We introduce quantitative analysis methods into the process of identifying technological opportunities in FLNG systems to handle relevant data information. This approach enables the prediction of technology development trends, making the results more objective and scientifically grounded.

➁ We establish a technological opportunity identification framework that combines patent mining with morphological analysis. During the patent mining process, we perform cluster analysis on patent texts, which enhances the comprehensiveness of data collection and makes the presentation clearer and more intuitive.

➂ We identify potential technological opportunities in FLNG that enable the efficient allocation of heat exchangers, storage tanks, export equipment, and handling systems within the limited space available. Meanwhile, we discussed their potential challenges in practical application scenarios.

Although the pathway for identifying technological opportunities improves methods regarding technological innovations, it also has some limitations. First, it requires manual screening of a large number of patents. Second, the selection of morphological structures must be complete; otherwise, the conclusions drawn may lack a referential value. Finally, the issue of “time lag” in patent citations is difficult to avoid. To address these challenges, future research should focus on adopting more advanced methods for data mining and processing to reduce the need for manual screening and enhance the efficiency and accuracy of predicting technological innovation opportunities. Additionally, as transportation methods and equipment continue to improve, the above opportunities will become more cost-effective and easier to integrate into enterprise production and usage. Moreover, this will enable our approach to be expanded to other application fields. Lastly, exploring more advanced patent identification technologies to enable dynamically updating technological opportunity identification will better support technological innovation activities in a scientific, objective, and comprehensive manner, thereby enriching the system of technological innovation methods. Future research can explore how the identified technological innovations will influence the FLNG industry’s future development. By further examining the practical applications of these innovations and their wide-reaching impacts on the industry—particularly on key stakeholders, such as FLNG operators, technology suppliers, and policymakers—a more comprehensive understanding of the sector can be achieved. This will highlight how new technologies can improve efficiency, reduce costs, and foster sustainable development, thereby establishing a strong foundation for the FLNG industry’s long-term competitiveness. Additionally, the industry faces several practical challenges, including the complexity of technology implementation, financial investment, regulatory hurdles, and market acceptance. These factors could affect the smooth adoption of new technologies and the industry’s sustainable growth.