Research on Cooperative Innovation Network Structure and Evolution Characteristics of Electric Vehicle Industry

Abstract

The electric vehicle industry involves various technological cooperative innovations, meaning that electric vehicle companies must not only cooperate and innovate with similar enterprises in the same industry, but also collaborate with enterprises across the industry and research institutions (RIs). This paper empirically analyzed the structure and evolution characteristics of cooperative innovation networks, based on the patent application data in the field of electric vehicles from 2006 to 2021. It was found that the firm–firm intra-industry and inter-industry cooperative innovation networks, and firm–RI cooperative innovation networks have the characteristics of phased evolution in structure, and the network structure in the evolution process has similarities and differences. Furthermore, during the industrial formation period, inter-industry cooperative innovation focuses on the cooperation between the midstream and downstream industries of the industrial chain; but in the industrial growth period, inter-industry cooperative innovation has been widely extended to multiple industries in the upper, middle and lower reaches of the industrial chain. While the scale of intra-industry and industry–research institution cooperative innovation continues to expand with the development of the industry, the collaboration is concentrated in the five industries in the middle and downstream of the industrial chain. The research conclusion can provide a reference for different cooperative innovation partners of electric vehicles and other emerging industries to formulate differentiated policies.

1. Introduction and Literature Review

Vehicle electrification is the key to reducing mobile carbon emission sources. According to the International Energy Agency, the number of electric vehicles in China in 2020 was 4.92 million, accounting for 49% of the world’s 10 million in that year [1]. In 2021, China added 2.95 million electric vehicles [2]. The electric vehicle industry has complex technology, high uncertainty, high cost and high risk in the innovation process. Technical cooperation and innovation are some of the trends in the network era. It is conducive to all R & D bodies to become embedded in the innovation network to give full play to their advantages in the industrial chain or knowledge chain, and to obtain complementary or substitutionary knowledge through the flow of information and knowledge, so as to form resource optimization and efficiency improvement. The present literature mainly studies the evolution law of the innovation network from the two dimensions of time and space, lacking the comparative analysis of the heterogeneity of the cooperative innovation partners. This paper constructs cooperative innovation networks of multiple partners through patent cooperative innovation relationships, and explores the structure and evolution characteristics of firm–firm intra-industry and inter-industry cooperative innovation networks, and firm–research institution (RI) cooperative innovation networks.

Through review of the relevant literature, one stream of studies concentrates on the structure of aggregated innovation and cooperation network. Newman (2003) and Barrat et al. (2004) used the network average shortest distance, vertex degree distribution, aggregation degree, degree correlation and other indicators to measure the structural characteristics of cooperative networks, defined the cooperative network matrix in combination with weight and topology, and studied the statistical characteristics of complex networks and the actual strength difference between vertices and edges [3,4]. Some scholars also use network topology, network clustering, network structure entropy, network node degree, network density, centrality and other indicators to analyze the cooperative innovation networks of the software industry cluster, electric vehicle industry innovation network, patent weighted cooperation network and industry–university research cooperation networks, so as to deeply excavate the characteristics and evolution law of network structure [5,6,7]. By analyzing the structure of China’s solar photovoltaic power technology innovation network and the characteristics of patent applicants, Lacasa et al. (2018) found that technology catch-up is driven by more and more Chinese patent applicants gathering in isolated communities, while the role of foreign participants in joint patent activities is decreasing [8].

Researchers further explored the spatial evolution characteristics and driving mechanisms of the regional innovation network from the spatial dimension. Xia et al. (2019) and Yaqin et al. (2019) analyzed the spatial structure, agglomeration mode and distribution evolution law of the regional innovation network of the electric vehicle industry [9,10]. Chunxiao et al. (2021) constructed China’s urban logistics innovation spatial network based on patent right transfer data, and explored the spatial network diamond structure, diffusion mode and connection situation formed by the core urban agglomeration [11]. Yike et al. (2021) analyzed the cooperation tightness, knowledge absorption ability, knowledge control ability and cluster development ability of the industry–university research cooperative innovation network of listed companies from three aspects: enterprise factor density, life cycle and regional characteristics [12].

Another stream of studies contributes to research on the mode of knowledge flow and technology diffusion in the innovation network. Soh (2010) and Choe et al. (2013) found that the average path length and clustering coefficient of the patent citation network, the degree centrality of nodes, the degree of intermediary centrality, the degree of strategic coordination and knowledge sharing affect the knowledge diffusion and innovation performance in the network [13,14]. In addition, the patent citation network in line with the characteristics of the small world has good overall connectivity, fast information dissemination, and nodes with high intermediary centrality dominating knowledge diffusion [15,16]. Costa Campi et al. (2014) studied the patent citation network in the field of oil and gas, and found that when the cited patent involves multiple industries, intersectoral spillover is likely to happen, while when the cited patent is limited to the industry, there is intra-sectoral spillover [17].

The present literature studied the structural evolution characteristics of the cooperative innovation network from many perspectives, and most of them regard the interaction of cooperative bodies within the evolution of the innovation network as the same. Few articles dynamically analyze the evolution characteristics of the innovation network of the new energy vehicle industry from the perspective of the different types of cooperation partners. However, enterprises can not only establish cooperative relations with partners in the same industry, but also with partners in different industries or research institutions. Although these relationships may affect innovation through similar mechanisms, such as increasing resource acquisition and knowledge spillover [18], due to the different nature of shared resources, inter-industry cooperation, intra-industry cooperation and industry–university research cooperation may have different effects on enterprise innovation [19]. Since the product lines, technologies, processes, specifications and management practices in the same industry are more similar, the knowledge of enterprises in the same industry is easier to absorb [20] and provides more advantages for knowledge assimilation and reorganization [18]. Inter-industry cooperation can expose enterprises to more heterogeneous information, which increases the possibility of exploratory innovation. Cooperation between companies and research institutions can often solve the technical problems of enterprises, improve innovation efficiency in the short term and promote technological upgrading in the long term [21].

Since the electric vehicle industry involves the comprehensive application of machinery, energy, control, safety and artificial intelligence, the firm–firm intra-industry cooperation, firm–firm inter-industry cooperation, and firm–research institution cooperation may have different impacts on innovation. By studying these three types of cooperative innovation networks, this paper can further reveal the characteristics of the R & D cooperative innovation mechanism and its dynamic evolution of the electric vehicle industry.

2. Construction of Cooperative Innovation Network for Electric Vehicle Industry

2.1. Sample Processing and Construction of Cooperative Innovation Network

The patent data comes from the patent database service platform of the China National Intellectual Property Administration. We collected the patent application data of invention and utility models in the field of electric vehicles in China from 2006 to 2021. Data processing methods included merging renamed enterprises, cleaning up problems such as name entry errors, and eliminating patents that were obviously irrelevant to the electric vehicle industry; confirming the nature and industry of the applicants from the national enterprise credit information publicity system, the company’s official website and other network resources, and corresponding the industry to the major categories in the national economic industry classification.

If two or more applicants belonged to the same patent, it was considered that these applicants had a cooperative innovation relationship. According to whether the applicants of the joint patent belonged to the same industry or research institution, the patent was divided into inter-industry cooperation, intra-industry cooperation and firm–research institution cooperation. According to the collaboration relationship of the applicants, this paper first constructed firm–firm inter-industry cooperative innovation networks, firm–firm intra-industry cooperative innovation networks, and firm–research institution (RI) cooperative innovation networks. In addition, based on the industry or research institution of the applicants, this paper further built inter-industry cooperative innovation networks, intra-industry cooperative innovation networks, and industry–research institution cooperative innovation networks.

2.2. Characteristic Indicators of Innovation Network

2.2.1. Average Degree Centrality ($\mathrm{ADC}$)

The ratio of the sum of degree centrality of all nodes to the number of nodes [22], represents the connection breadth of applicants. The calculation formula is as follows:

$$\mathrm{ADC}=\frac{\sum {\mathrm{a}}_{\mathrm{ij}}}{\mathrm{N}}$$

(1)

where ${\mathrm{a}}_{\mathrm{ij}}$ is the centrality of each node, $\mathrm{N}$ is the number is nodes.

2.2.2. Average Connection Ties ($\mathrm{ACT}$)

The ratio of the total number of ties to the degrees of all nodes [9], represents the connection depth of applicants. The calculation formula is as follows:

$$\mathrm{ACT}=\frac{2\mathrm{T}}{\sum {\mathrm{a}}_{\mathrm{ij}}}$$

(2)

where T is the actual number of ties in the network, ${\mathrm{a}}_{\mathrm{ij}}$ is the centrality of each node.

2.2.3. Network Density

The ratio of the actual number of edges in the network to the upper limit of the number of edges that can be accommodated [8]. With the increase of network density, the degree of knowledge sharing and cooperation among bodies in the network also increases accordingly [23], which is conducive to the development of emerging technologies. The calculation formula is:

$$\mathrm{Density}=\frac{2\mathrm{E}}{\mathrm{N}\left(\mathrm{N}-1\right)}$$

(3)

where $\mathrm{E}$ is the actual number of edges in the network, $\mathrm{N}$ is the number of nodes.

2.2.4. Average Clustering Coefficient (ACC)

The average value of clustering coefficients of all nodes in the network. The average clustering coefficient can reflect the local characteristics and collectivization degree of the cooperative networks [24]. The calculation formula is:

$${\mathrm{C}}_{\mathrm{a}}=\frac{2{\mathrm{e}}_{\mathrm{a}}}{{\mathrm{k}}_{\mathrm{a}}\left({\mathrm{k}}_{\mathrm{a}}-1\right)}$$

(4)

$$\mathrm{ACC}=\frac{{{\displaystyle \sum }}_{\mathrm{a}=1}^{\mathrm{N}}{\mathrm{C}}_{\mathrm{a}}}{\mathrm{N}}$$

(5)

where ${\mathrm{C}}_{\mathrm{a}}$ is the node clustering coefficient, ${\mathrm{k}}_{\mathrm{a}}$ is the number of nodes jointly connected with node $\mathrm{a}$, ${\mathrm{e}}_{\mathrm{a}}$ is the actual number of edges between adjacent nodes of node $\mathrm{a}$. $\mathrm{ACC}$ is the network average clustering coefficient, $\mathrm{N}$ is the number of nodes.

2.2.5. Average Path Length (APL)

The average value of the shortest path for any two nodes in the network to connect [25]. The larger the average distance, the more difficult the communication between nodes in the cooperative network is and the lower the efficiency of information transmission. The calculation formula is as follows:

$$\mathrm{APL}=\frac{\sum {\mathrm{d}}_{\mathrm{ab}}}{\mathrm{N}\left(\mathrm{N}-1\right)}$$

(6)

where ${\mathrm{d}}_{\mathrm{ab}}$ is the shortest path length between nodes $\mathrm{a}$ and $\mathrm{b}$, $\mathrm{a}\ne \mathrm{b}$.

2.2.6. Distance-Based Cohesion Index (DCI)

It reflects the network cohesion. The higher the cohesion index, the stronger the cohesion of the cooperative network [26].

3. Analysis of the Evolution Characteristics of Firm–Firm and Firm–RI Cooperative Innovation Networks of Electric Vehicle Industry

3.1. Life Cycle Identification of Electric Vehicle Industry

Figure 1 shows the number of patent applications in the electric vehicle industry from 2006 to 2021. It can be seen that the number of patent applications in the electric vehicle industry from 2006 to 2010 is small; Since 2010, China has issued subsidy policies for electric vehicles to promote the formation of the industry, and the number of patents has gradually increased after 2011; Since 2016, relevant national policies have been adjusted to promote the improvement of the performance of electric vehicles and their core components, as well as the orderly development of the industrial scale. After that, the number of patent applications has further increased and the patent scale has significantly expanded. From 2019 to 2021, the number of patent applications fluctuated slightly due to the subsidy decline policy and the impact of the epidemic. In order to compare the characteristics of the innovation network structure in different development stages, this paper selected five years as a life cycle stage, that is, 2006–2010 is the R & D stage of the electric vehicle industry, 2011–2015 is the industrial formation period, and 2016–2020 is the industrial growth period.

3.2. Measurement of Structural Characteristics of Firm–Firm and Firm–RI Innovation Network

This paper uses UCINET software to measure the structural characteristics of firm–firm inter-industry and firm–firm intra-industry cooperative innovation networks, and firm–research institution cooperative innovation networks of the electric vehicle industry in different life cycle stages.

Table 1. Structural characteristics of firm–firm and firm–RI innovation networks in the electric vehicle industry.

	Network Indicators	2006–2010	2011–2015	2016–2020
Firm–firm inter-industry cooperative innovation network	Number of nodes	22	128	465
	Number of ties	40	324	1468
	Average connection ties	3.64	4.05	5.05
	Average degree	1	1.25	1.25
	Density	0.048	0.010	0.003
	Avg. clustering coefficient	0	0	0.08
	Average path length	1	2.035	3.177
	Distance-based cohesion	0.048	0.021	0.006
Firm–firm intra-industry cooperative innovation network	Number of nodes	15	108	325
	Number of ties	17	271	1108
	Average connection ties	2.27	3.58	4.88
	Average degree	1	1.4	1.4
	Density	0.071	0.013	0.004
	Avg. clustering coefficient	0	0	0.124
	Average path length	1.111	1.969	1.99
	Distance-based cohesion	0.081	0.515	0.505
Firm–research institution cooperative innovation network	Number of nodes	21	174	630
	Number of ties	25	487	1477
	Average connection ties	2.26	3.88	2.76
	Average degree	1.05	1.44	1.7
	Density	0.053	0.008	0.003
	Avg. clustering coefficient	0	0.046	0.082
	Average path length	1.083	1.969	1.991
	Distance-based cohesion	0.055	0.516	0.505

is the measurement results of the structural characteristics of the innovation network.

3.3. Structure Map of Firm–Firm and Firm–RI Innovation Networks

This paper further uses the visualization software Netdraw to draw the evolution of the firm–firm inter-industry, firm–firm intra-industry, and firm–research institution cooperative innovation networks in different life-cycle stages of the electric vehicle industry. Figure 2 is the phased evolution map of the three types of cooperative innovation networks. The node in Figure 2 represents the applicant (firm or research institution). The size of the node depends on the degree centrality of the applicant, that is, the more cooperation with other applicants, the larger the node. The edge between nodes indicates that there is a cooperative innovation relationship between the two applicants. The thicker the edge, the deeper the cooperation.

3.4. Structural Evolution Trend of Firm–Firm and Firm–RI Innovation Networks

In general, from 2006 to 2020, the firm–firm inter-industry, firm–firm intra-industry, and firm–research institution cooperative innovation networks showed an upward trend in the network scale, cooperation depth and breadth, average clustering coefficient and average path length, and a downward trend in network density and distance-based cohesion.

3.4.1. The Scale of Innovation Network Has Grown Considerably

In the R & D stage, the scale of the three types of cooperative innovation networks was quite low, the number of nodes and connection ties in the network were small, and the applicants were mainly in pairwise cooperation. With the formation and rapid development of the industry, more and more companies and research institutions poured into the field of electric vehicles for cooperative technological innovation, and the scale of the three types of cooperative innovation networks expanded rapidly. After entering the growth period, the scale of the three types of innovation networks was: firm–RI cooperation > firm–firm inter-industry cooperation > firm–firm intra-industry cooperation.

3.4.2. The Breadth and Depth of Firm–Firm Cooperation and Firm–Research Institution Cooperation Have Different Trends

The average degree and average connection ties represent the breadth and depth of R & D cooperation, respectively. Between 2006 and 2010, the cooperation breadth and depth of the three types of cooperative innovation networks were relatively low, and the applicants mainly cooperated in pairs. With the expansion of the industry, the breadth of firm–firm cooperation basically remained stable after the rise in the formation period, while the depth of firm–firm cooperation was constantly improving.

The breadth of firm–research institution cooperation was gradually increasing, while the general level of cooperation depth showed an inverted U-shaped trend that increased in the formation period, but decreased in the growth period. Furthermore, in the growth period, the general level of cooperation depth between the firms and research institutions (ranking in the top 10) was 27 times, which was much higher than 19 times in the formation period; in the growth period, the general level of cooperation depth between the firms and research institutions (ranking in the top 20) was 21 times, which was much higher than the 13 times in the formation period. It shows that the cooperation between large automobile companies, power companies and strong scientific research institutions was gradually deepening in the growth period. The general level of the depth of firm–research institution cooperation rose during the formation period, but decreased during the growth period, which was mainly because of the influx of a large number of companies and research institutions for cooperative R & D during the growth period with a relatively low degree of cooperation, which made the general level of the depth of firm–research institution cooperation decline.

3.4.3. The Density of Innovation Networks Gradually Decreased

With the development of the industry, the number of applicants of patent and cooperation ties in the three types of cooperative innovation networks increased, while the network density decreased significantly. This is mainly because the innovation network continued to expand in the evolution process, more and more R & D subjects participated in cooperative innovation, and the upper limit of the number of connection edges that could be accommodated in the network was much larger than the actual number of connection edges in the network due to the increase of nodes in the network. Therefore, the network density shows a sparse trend.

3.4.4. The Average Clustering Coefficient of Innovation Network Increased Continuously

In the R & D stage of the electric vehicle industry, the average clustering coefficient of the three types of innovation networks was zero. With the rapid development of the industry, the average clustering coefficient of the firm–research institution cooperative innovation network increased gradually, while the average clustering coefficient of the firm–firm cooperative innovation network experienced a significant rise in the growth period.

The improvement of the average clustering coefficient reflected that strong cooperative groups had appeared in the growth period of the three types of cooperative innovation networks. In the firm–RI cooperative innovation network, companies cooperated more closely with powerful large scientific research institutes and universities, and the agglomeration effect of scientific and technological innovation resources was gradually strengthened. The firm–firm cooperative innovation network mainly focused on the technological interactions among large automobile companies, large power companies and their related enterprises.

3.4.5. The Average Path Length and Distance-Based Cohesion of Innovation Networks Had Different Trends

The path length reflects the shortest distance between nodes in the network. With the development of the industrial life cycle, the average path length of the firm–firm inter-industry cooperative innovation network gradually increased, while the firm–firm intra-industry and firm–RI cooperative innovation network basically remained stable after experiencing a significant increase in the average path length. From the distance-based cohesion, the index of the firm–firm inter-industry cooperative innovation network gradually decreased, while the cohesion of the firm–firm intra-industry and firm–RIs cooperative innovation network basically remained stable after experiencing rapid growth in the formation period.

It can be seen that with the rapid expansion of the scale of the innovation network, the tightness of the company’s cross industry R & D cooperation decreased, resulting in the decline of technology diffusion, information transmission and cooperation efficiency. However, after continuous expansion, the firm–firm intra-industry and firm–RI cooperative innovation network still maintained a low average distance and high cohesion. The cooperation among applicants was relatively close, which reduced the cost and loss of resources and information flow among innovation subjects, and the efficiency of knowledge dissemination in the network was high.

4. Analysis of the Structural Evolution Characteristics of Inter-Industry Cooperative Innovation Network

According to the industry of the applicant of the firm–firm inter-industry cooperative patent, this paper further built the inter-industry cooperative innovation network of the electric vehicle industry in 2006–2010, 2011–2015 and 2016–2020, respectively. The node represents the industry of the applicant. The size of the node depends on the degree of centrality of the industry, that is, the more cooperation with other industries, the larger the node. The edge between nodes indicates that there is a cooperative innovation relationship between the two industries. The thicker the edge, the deeper the cooperation.

Table 2. Structural characteristics of inter-industry cooperative innovation network of the electric vehicle industry from 2006 to 2020.

Network Indicators	2006–2010	2011–2015	2016–2020
Number of nodes	8	21	35
Number of ties	40	324	1468
Avg. connection ties	4.44	8.12	15.48
Average degree	2.25	3.8	5.42
Density	0.32	0.19	0.16

is the values of the indicators of the inter-industry cooperative innovation network in the three stages.

4.1. Structural Evolution Characteristics of Inter-Industry Cooperative Innovation Network in the R & D Period

Table 2 shows that in the industrial R & D stage from 2006 to 2010, only eight industries participated in the inter-industry cooperative innovation network, and the amount of cooperation was low. The average connection ties and average degree suggest that the depth and breadth of inter-industry cooperation at this stage were quite low, and each industry cooperated with only two other industries on average. While the network density in this stage was high, indicating that the cooperative network structure was relatively tight.

Figure 3 is the visual map of the inter-industry cooperative innovation network of the electric vehicle industry from 2006 to 2010 with Netdraw software. It can be seen that the technical cooperative innovation between the downstream and midstream links of the industrial chain had just started in the R & D period.

Table 3. Inter-industry cooperation portfolio of the electric vehicle industry from 2006 to 2010.

Inter-Industry Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Automobile Manufacturing–research and test development	16	40%
Automobile manufacturing–electrical machinery and equipment manufacturing	12	30%
Power and heat production and supply–software and information technology services	4	10%
Power and heat production and supply–professional technical service	3	7.5%

shows the inter-industry cooperation portfolio of the electric vehicle industry from 2006 to 2010 (with more than one occurrence of collaboration). It can be seen that this stage focuses on collaboration between automobile manufacturing, electrical machinery and equipment manufacturing, electric power and heat production and supply, scientific research and computing services industry. It reflects that the R & D period began with the research on the power battery system, the core link of the industrial chain, and less in other areas.

4.2. Structural Evolution Characteristics of Inter-Industry Cooperative Innovation Network in the Formation Period

As the electric vehicle industry entered the formation period, the number of nodes and ties in the inter-industry cooperative innovation network increased to 21 and 324, respectively (Table 2). At this stage, the average degree increased to 3.8, indicating the improvement of the breadth of inter-industry cooperation and each industry cooperated with 3–4 other industries on average; the average connection ties were doubled, suggesting that the inter-industry collaboration was more frequent; the network density decreased, mainly due to the rapid expansion of the scale of the network and many new inter-industry cooperation relationships were established. Figure 4 is a visual map of the inter-industry cooperative innovation network of the electric vehicle industry from 2011 to 2015 constructed by Netdraw software. It can be seen that the collaboration between the midstream and downstream links of the industrial chain was further deepened, and more midstream industries entered this field for cooperative technological innovation, such as special equipment manufacturing industry and general equipment manufacturing industry.

Table 4. Inter-industry cooperation portfolio of the electric vehicle industry from 2011 to 2015.

Inter-Industry Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Automobile manufacturing–research and test development	69	21%
Automobile manufacturing–electrical machinery and equipment manufacturing	54	17%
Power and heat production and supply–science and technology promotion and application service	41	13%
Power and heat production and supply–software and information technology services	32	10%
Electrical machinery and equipment manufacturing–science and technology promotion and application services	21	6%
Automobile manufacturing–technology promotion and application service	13	4%
Electrical machinery and equipment manufacturing–power and heat production and supply	12	4%
Comprehensive utilization of waste resources–science and technology promotion and application service	11	3%
Automobile Manufacturing–software and information technology services	10	3%

is the inter-industry cooperation portfolio of the electric vehicle industry from 2011 to 2015 (with more than 10 occurrences of cooperation). The first is the cooperation between the automobile manufacturing industry with the scientific research and computing service, electrical machinery and equipment manufacturing industries, accounting for the highest proportion, suggesting that the innovation of electric vehicles and power batteries was the emphasis of the inter-industry cooperative innovation; The second is the cooperation between the power and heat production and supply industry with the software and information technology service, science and technology promotion and application service industries, accounting for 23%, mainly focusing on the research and development of energy storage systems, energy management systems and battery management systems. Compared with the first stage, the cooperation between the comprehensive utilization of waste resources industry with the science and technology promotion and application service industry also appeared in the formation period, used for research into the recycling and reuse technology of waste power batteries.

4.3. Structural Evolution Characteristics of Inter-Industry Cooperative Innovation Network in the Growth Period

In the rapid development stage of 2016 to 2020, the scale of inter-industry cooperative innovation network continued to expand in terms of nodes and ties, which had increased to 35 and 1468, respectively (Table 2). In addition, the average connection ties and average degree improved gradually, so that the inter-industry collaboration was strengthened. Besides, the network density decreased only slightly, indicating that the inter-industry cooperation has maintained a relatively close structure, and the diffusion of information or resources among the industries were stable. Figure 5 is a visual map of the inter-industry cooperative innovation network of the electric vehicle industry from 2016 to 2020, constructed by Netdraw software, showing that the network is still dominated by collaborations between the midstream and downstream industries. Meanwhile, more upstream and downstream industries entered the network, further broadening the cooperation between the upstream, midstream and downstream of the electric vehicle industrial chain.

Table 5. Inter-industry cooperation portfolio of the electric vehicle industry from 2016 to 2020.

Inter-Industry Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Automobile manufacturing industry–technology promotion and application service	466	31.7%
Professional and technical services–science and technology promotion and application services	194	13.2%
Automobile manufacturing–research and test development	164	11.1%
Automobile manufacturing–electrical machinery and equipment manufacturing	73	4.9%
Power and heat production and supply–software and information technology services	46	3.1%
Power and heat production and supply–science and technology promotion and application service	33	2.2%
Electrical machinery and equipment manufacturing–research and experimental development	29	2%
Software and information technology services–science and technology promotion and application services	28	1.9%
Automobile manufacturing–special equipment manufacturing	27	1.8%
Research and experimental development–science and technology promotion and application service	25	1.7%
Electrical machinery and equipment manufacturing–comprehensive utilization of waste resources	23	1.5%
Electrical machinery and equipment manufacturing–chemical raw materials and chemical products manufacturing	20	1.4%
Electrical machinery and equipment manufacturing–science and technology promotion and application services	19	1.3%
Electrical machinery and equipment manufacturing–computer, communication and other electronic equipment manufacturing	17	1.2%
Electrical machinery and equipment manufacturing–software and information technology services	17	1.2%
Non-ferrous metal smelting and rolling processing–special equipment manufacturing	13	0.8%
Instrument manufacturing–science and technology promotion and application service	13	0.8%

is the inter-industry cooperation portfolio of the electric vehicle industry (with more than 10 occurrences of cooperation) from 2016 to 2020, showing that the cooperation between downstream and midstream industries had further deepened. Cooperative innovation between automobile manufacturing industry, electrical machinery and equipment manufacturing industry, special equipment manufacturing industry, scientific research and computing service industry was still the focus, the proportion of which had further expanded to 50%. Except for the in-depth exploration of the performance of electric vehicles, relevant parts and power batteries, the research into charging piles and charging and replacement power stations was also on the rise. Moreover, the depth of cooperation between power and heat production and supply industries, software and information technology services, science and technology promotion and application services had improved for the research of energy storage systems and facilities, charging and exchange systems and corresponding facilities.

The cooperation between downstream and midstream industries had also been broadened (Table 5), including the cooperation between the electrical machinery and equipment manufacturing industry with the computer, communication and other electronic equipment manufacturing industry, software and information technology service industry, which were mainly used for the research of charging devices and equipment, charging management system, etc. There was also cooperation between the automobile manufacturing, rubber and plastic products, and metal products industries, etc.

Moreover, the network newly increased the cooperation between the midstream and upstream industries (Table 5). For example, the innovation between the electrical machinery and equipment manufacturing, the chemical raw materials and chemical products manufacturing, and the research and experimental development industries, which improved the research into cathode materials, battery refractories and rare earth materials for the motors of lithiumion power batteries, so as to enhance the performance of batteries and motors.

5. Analysis on the Structural Evolution Characteristics of Intra-Industry Cooperative Innovation Network

Based on the industry of the applicants of the firm–firm intra-industry cooperative patents, in this paper, we further constructed the intra-industry cooperative innovation networks of the electric vehicle industry in 2006–2010, 2011–2015 and 2016–2020, respectively. The nodes represent the industry of the applicant. The edge indicates the cooperative innovation relationship within the industry.

Table 6. Structural characteristics of intra-industry cooperative innovation network of the electric vehicle industry from 2006 to 2020.

Network Indicators	2006–2010	2011–2015	2016–2020
Number of nodes	3	9	18
Number of ties	14	271	1108
Average connection ties	4.67	30.11	61.56

is the values of the indicators of the intra-industry cooperative innovation network in the three stages.

5.1. Structural Evolution Characteristics of Intra-Industry Cooperative Innovation Network in the Industrial R & D Period

In the R & D stage from 2006 to 2010, the scale of intra-industry cooperative innovation network was small, with only 14 instances of R & D collaboration carried out within three industries (Table 6).

Table 7. Intra-industry cooperation portfolio of the electric vehicle industry from 2006 to 2010.

Intra-Industry Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Automobile manufacturing industry	15	88%
Electrical machinery and equipment manufacturing	1	6%
Power and heat production and supply industry	1	6%

is the collaboration portfolio of the intra-industry cooperation from 2006 to 2010. At this stage, foreign automobile companies took the lead in carrying out intra-industry cooperation 15 times, and only one instance of intra-industry cooperation in the other two industries.

5.2. Structural Evolution Characteristics of Intra-Industry Cooperative Innovation Network in the Industrial Formation period

In the industry formation stage from 2011 to 2015, the scale of the intra-industry cooperative innovation network was gradually expanding (Table 6), the number of industries cooperating with the same industry had increased to 9, and the depth of cooperation had also improved significantly. On average, each industry had cooperated with the same industry 30 times.

Table 8. Intra-industry cooperation portfolio of the electric vehicle industry from 2011 to 2015.

Intra-Industry Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Power and heat production and supply industry	141	52%
Electrical machinery and equipment manufacturing	92	34%
Automobile manufacturing industry	22	8%

is the intra-industry cooperation portfolio of the electric vehicle industry from 2011 to 2015 (with more than 10 occurrences of cooperation). It can be seen that with the start of the electric vehicle industry, the number of cooperative innovations within the power, heat production and supply, electrical machinery and equipment manufacturing, and automobile manufacturing industries had increased dramatically, aimed at the research of battery and electronic control technology of pure electric vehicles and hybrid electric vehicles. The top three industries accounted for 94% of all intra-industry cooperation in this stage, and other industries conducted little intra-industry cooperation.

5.3. Structural Evolution Characteristics of Intra-Industry Cooperative Innovation Network in the Industrial Growth Period

During the industrial growth period from 2016 to 2020 (Table 6), the scale further rose to 18 industries, newly increased civil engineering construction industry, special equipment manufacturing industry, etc. And the intra-industry cooperation degree had also been deepened, reaching an average of 61 times for each industry.

Table 9. Intra-industry cooperation portfolio of the electric vehicle industry from 2016 to 2020.

Intra-Industry Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Automobile manufacturing industry	494	45%
Power and heat production and supply industry	256	23%
Electrical machinery and equipment manufacturing	186	17%
Technology promotion and application services	59	5%
Research and experimental development	53	5%
Software and information technology services	16	1%
Civil engineering and construction	15	1%
Special equipment manufacturing	10	1%

is the intra-industry cooperation portfolio of the electric vehicle industry from 2016 to 2020 (with more than 10 occurrences of cooperation). It shows that the automobile manufacturing industry took the lead in the intra-industry cooperative innovation, accounting for 45%, which has continuously deepened the research on electric vehicle batteries, motors, electronic controls, powertrain systems, body frames, auto parts, etc. The second and third ranked actors are production and supply of electric power and heat and the manufacturing of electrical machinery and equipment, which is used for more in-depth and diversified exploration of energy storage, charging and power distribution equipment. The number of intra-industry collaboration within science and technology promotion and application service industry, research and experimental development industry has also increased, mainly for the research of auxiliary power battery system and charging system.

The cooperative innovation in the above five industries accounted for 95% of the total cooperation, and the other 13 industries accounted for no more than 1%. In conclusion, although the scale of intra-industry cooperative innovation is gradually increasing, it is mainly concentrated in the five industries in the middle and downstream of the electric vehicle industry chain.

6. Analysis of the Structural Evolution Characteristics of Industry–Research Institution Cooperative Innovation Network

Based on the industry or research institution of the applicant for the firm–research institution’s cooperative patent, in this paper we further constructed the industry–research institution cooperative innovation networks of the electric vehicle industry in 2006–2010, 2011–2015 and 2016–2020, respectively. The nodes represent the industry or research institution of the applicant. The edges between nodes indicate that there is a cooperative innovation relationship between the industry and the research institution.

Table 10. Structural characteristics of industry–research institution cooperative innovation networks of the electric vehicle industry from 2006 to 2020.

Network Indicators	2006–2010	2011–2015	2016–2020
Number of nodes	4	10	28
Number of ties	25	420	1255
Average connection ties	8.33	46.67	46.48

shows the values of the indicators of the industry–research institution cooperative innovation networks in the three stages.

6.1. Structural Evolution Characteristics of Industry–Research Institution Cooperative Innovation Networks in the Industrial R & D Period

In the industrial R & D stage from 2006 to 2010 (Table 10), the scale of the industry–research institution cooperative innovation network was small, with only four industries participating in the collaboration. In addition, the cooperation degree was low, each industry only cooperated with research institutions 4–5 times on average. This might be because the core technology path of the electric vehicle industry at this stage was not clear, mainly focussing on basic research, so the application research conducted by the company cooperating with the research institutions was less.

Table 11. Industry–research institution cooperation portfolio of the electric vehicle industry from 2006 to 2010.

Industry–Research Institution Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Automobile manufacturing industry–research institutions	12	48%
Power and heat production and supply industry–research institutions	8	32%
Electrical machinery and equipment manufacturing–research institutions	4	16%
Research and experimental development–research institutions	1	4%

shows the industry–research institution cooperation portfolio of the electric vehicle industry from 2006 to 2010, showing that mainly automobile manufacturing, power and heat production and supply, and electrical machinery and equipment manufacturing cooperated with research organizations at this stage.

6.2. Structural Evolution Characteristics of Industry–Research Institution Cooperative Innovation Networks in the Industrial Formation Period

During 2011–2015, the requirements for technology gradually improved, the number of industries cooperating with research institutions for innovation increased to 10, and the depth of cooperation also rose significantly (Table 10), with an average of 46 times for each industry cooperating with research institutions.

Table 12. Industry–research institution cooperation portfolio of the electric vehicle industry from 2011 to 2015.

Industry–Research Institution Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Power and heat production and supply industry–research institutions	168	40%
Automobile manufacturing industry–research institutions	93	22%
Research and experimental development–research institutions	63	15%
Technology promotion and application services–research institutions	33	8%
General equipment manufacturing industry–research institutions	19	5%
Software and information technology services–research institutions	16	4%
Professional and technical services–research institutions	12	3%
Electrical machinery and equipment manufacturing–research institutions	11	3%

shows the industry–research institution cooperation portfolio of the electric vehicle industry from 2011 to 2015 (with more than 10 instances of cooperation). Specifically, the power and heat production and supply industry cooperated with research institutions 168 times, accounting for 40%, mainly for the research and development of electric vehicle energy storage devices, charging and changing devices and equipment, etc.; the automobile manufacturing, research and test development, science and technology promotion and application service industries cooperated with research institutions 189 times, accounting for 45%, in the study of the power module, drive system and lightweight body of electric vehicles. It can be seen that the top four collaboration combinations accounted for 82%, and there was less cooperation between other industries and research institutions.

6.3. Structural Evolution Characteristics of Industry–Research Institution Cooperative Innovation Networks in the Industrial Growth Period

With the electric vehicle industry entering the fast lane from 2016 to 2020, the requirements for the technical performance of electric vehicles were becoming more and more strict. The scale of the industry–research institution cooperative innovation networks had expanded to 28 industries (Table 10). There were more industries seeking to cooperate and innovate with research institutions, such as instrument manufacturing, metal products, computer, communication and other electronic equipment manufacturing.

Table 13. Industry–research institution cooperation portfolio of the electric vehicle industry from 2016 to 2020.

Industry–Research Institution Cooperation Portfolio	No. of Instances of Cooperation	Ratio of Cooperation
Power and heat production and supply industry–research institutions	407	32%
Automobile manufacturing industry–research institutions	307	24%
Technology promotion and application services–research institutions	144	11%
Electrical machinery and equipment manufacturing–research institutions	98	8%
Professional and technical services–research institutions	94	7%
Research and experimental development–research institutions	69	5%
Software and information technology services–research institutions	40	3%
Computer, communication and other electronic equipment manufacturing–research institutions	17	1%
General equipment manufacturing industry–research institutions	15	1%
Instrument manufacturing industry–research institutions	11	1%

shows the industry–research institution cooperation portfolio of the electric vehicle industry from 2016 to 2020 (with more than 10 instances of cooperation), presenting that the five industries in the middle and lower reaches of the industrial chain cooperate most with research institutions, accounting for 92%. In addition, the cooperation between other industries and research institutions accounts for less than 1%. It can be judged from Table 10 that although the average cooperation degree of the overall industry–research institution cooperative innovation network basically remained unchanged, the cooperation degree of the five major industries and research institutions had increased significantly.

Specifically, the cooperation between power and heat production and supply industries and research institutions was still the focus of the network, accounting for 32%, used to study the optimal configuration of charging devices, charging cluster systems, intelligent charging and power generation systems and devices, charging access to distribution networks, etc. The second rank was the cooperation between automobile manufacturing industry and research institutions, aimed at optimizing electric vehicle power management systems, control systems, heat dissipation systems, fault diagnosis systems, etc. The scientific research and technology services industry and research institutions cooperated 307 times, accounting for 24%, mainly focusing on the setting, detection and evaluation of electronic control systems and charge discharge management systems.

7. Conclusions

Based on the patent cooperation application data of the electric vehicle industry, combined with the social network analysis method, this paper analyzed the structure and evolution characteristics of the firm–firm inter-industry, firm–firm intra-industry, and firm–research institution cooperative innovation networks. As the industry entered the R & D period, formation period and growth period, the structure and changes of the three types of cooperative innovation networks in the electric vehicle industry were as follows:

(1) The scale and quality of the three types of cooperative innovation networks show an overall increasing trend.

The values of nodes, ties, average connection ties, average degree, average clustering coefficient and distance-based cohesion of the three types of networks have increased in varying degrees or increased first and then leveled off, reflecting that as an emerging technology industry, more and more firms and research institutions have entered the field of electric vehicles for innovation, and their cooperative innovation development and sustainability have formed a support for the industry; The increase of the average clustering coefficient indicates that cooperative innovation tends to gather to institutions with strong scientific research strength. In addition, the average path length and distance-based cohesion of firm–firm intra-industry and firm–research institution cooperative innovation networks remain stable after the significant expansion of the scale of cooperation, which can be considered as a relative growth.

(2) The three types of network cooperative innovation gradually show some selective characteristics.

It was shown that the density of the three types of network cooperative innovation decreased, the cohesion index of the cross industry cooperative innovation declined, and the average path length increased. First, the decrease in the density of cooperative innovation networks is most likely due to the rapid expansion of the scale of cooperation and the significant increase in selectivity, which makes the cooperation density tend to be scattered; secondly, the closeness of the company’s inter-industry cooperation and connection continues to reduce, so that the average path length is significantly improved and the efficiency of information transmission and cooperation is low; third, the general level of the depth of cooperation between firms and research institutions shows an inverted U-shaped trend, but the collaboration between large enterprises and strong scientific research institutes is becoming closer and more stable.

This is mainly because in the early stage of the formation of an emerging industry, a large number of new enterprises need to seek extensive basic cooperation. The cooperation density is high, but the cooperation quality is quite low. With the expansion of the industry, a few highly competitive enterprises stand out and lead to the improvement of industrial concentration. The firm–research institution cooperation has evolved more into in-depth collaboration between large enterprises and strong scientific research institutions. The general level of cooperation density of the whole industry shows a downward trend, but the cooperation quality has improved significantly. After entering the growth period, the industry concentration has further improved, large enterprises have obtained considerable market share and excess returns, have more ability and motivation to carry out more R & D cooperation, and further promote the depth of cooperation between such enterprises and research institutions.

In the actual life cycle of the electric vehicle industry, in the R & D stage and formation stage, inter-industry cooperation is mainly between the midstream and downstream links of the industrial chain; in the industrial growth period, inter-industry cooperative innovation is widely extended to multiple industries in the upstream, middle and downstream links of the industrial chain. The cooperative innovation between the upstream, middle and downstream links of the industrial chain is basically connected, and the breadth and depth of cooperation have been improved. However, there are still some difficulties in cooperation with some downstream industries (such as charging and power exchange). For intra-industry cooperation and industry–research institution cooperation, as the industry enters the formation and growth period, the cooperation is highly concentrated in three to five industries in the middle and downstream of the electric vehicle industry chain.