Core ideas
Products and technologies: CAE industrial software has a high entry threshold, the product development cycle is short but the product improvement cycle is long, and product research and development requires a lot of upfront funds and support from compound discipline talents. At the same time, in view of the current pain points on the market, CAE software is also constantly iterating, with the change of the underlying software and hardware environment, the current CAE software has begun to develop gradually in the direction of cloud (SaaS), CAX integration, and AI intelligence;
Industry and market: At present, the domestic and international CAE markets are facing a multi-head monopoly situation, and the three giants have a domestic market share of more than 95%. Domestic and international manufacturers are slightly worse than international manufacturers in terms of product maturity, commercialization, reliability, stability, etc., but with the national policy support and domestic manufacturing upgrades, the window of opportunity left for domestic CAE is gradually opening;
Track financing: From the perspective of capital, it can be seen that the foreign leading giants have achieved a vertical product layout of the whole chain of industrial software and horizontal to various industrial segments through a large number of mergers and acquisitions and self-research due to the first-mover advantage. Domestic CAE manufacturers have begun to exert efforts in the past 5 years, compared with the traditional software model of old companies, new startups have more choices to develop in the direction of cloud collaboration. During this period, the head institutions and industrial CVC also successively laid out with the trend of policies;
Automotive applications: Although the automotive industry is the industry with the most CAE applications, especially power simulation and NVH simulation are indispensable in the process of automobile manufacturing, but the current automotive companies use, whether it is CAE software or front and rear processors, overseas manufacturers occupy an indispensable dominant position, domestic simulation wants to "overtake in curves", it is best to start from emerging scenarios, such as automatic driving.
This research report will be divided into four parts - product and technology, industry market, track financing, automotive scenarios to introduce you to the simulation test of autonomous driving.
01
Products & Technologies
CAE concept
CAE (computer aid engineer) is computer-aided engineering design. It mainly refers to the use of computers to analyze the performance and safety reliability of engineering and products, simulate their future working conditions and operating behaviors, detect design defects early, and confirm the availability and reliability of future engineering, product functions and performance.
CAE is a broad concept that can literally encompass all aspects of engineering and manufacturing informatization. With the development of industry segments, such as EDA and CFD develop independent branches, the concept of CAE is gradually weakening and abstracting. CAE in a broad sense mainly refers to the use of computer simulation analysis to help or solve problems such as inspection, design, verification, and optimization in practical engineering applications. When talking about CAE, it may refer to CAD/EDA/CFD/TCAD, or structural dynamic analysis, thermal analysis, electromagnetic analysis, coupling analysis, and so on. CAE in the narrow sense refers more to the use of CAE software for simulation optimization analysis.
The industrial value of CAE
The biggest industrial value of CAE tools is that they can simplify the product design process, improve trial and error, and reduce costs.
Reduce costs: Help enterprises find problems as early as possible in the production chain, so as to reduce the number of iterations and experiments of actual products, optimize the entire design process, and achieve the effect of reducing design costs;
Improve efficiency: help enterprises to achieve some difficult or difficult to complete experimental steps in reality, while helping enterprises reduce design risks in actual production;
Accumulated experience: CAE tools not only provide software tools, but more importantly, integrate industry standards and accumulated experience that are critical to design, improving the efficiency and success rate of R&D design.
Product structure
The main components of the basic structure of CAE software include: user interface, data management system, database, expert system and knowledge base.
The data management system is the core component used in performance analysis or simulation using CAE software, on the one hand, it realizes the input of CAD, CAM and other format files through the interface, and on the other hand, it provides three processes of pre-processing, solution analysis (mostly finite element analysis) and post-processing to achieve simulation.
The Solution Analysis module can be subdivided into static linear subsystems, dynamic analysis subsystems, and many other branches according to the type of problem being handled.
Product flow
From the perspective of core workflow, the CAE software simulation process includes pre-processing, solving, post-processing, optimization, and reporting.
Pre-processing: The pre-processing module is mainly used to model the project or product, complete the input of analysis data, and establish a reasonable finite element model. Functions include geometry processing, model import, boundary definition, material definition, mesh definition, etc., and the meshing of excellent pre-processing software can avoid topological errors of complex models and improve the quality of model elements and simulation effects;
Solver: The solution analysis is mainly to analyze the element characteristics, assemble the finite element element, solve the finite element system and generate the finite element results of the finite element model; the sub-simulation forms of the model, such as stiffness, strength, heat, etc., are analyzed and calculated;
Post-processing: The post-processing module mainly performs data smoothing and other processing on the model based on the results of the solution analysis, and displays the visual effects of the force, displacement, stress, deformation, etc. of the model.
Simulation optimization: If the simulation results do not meet the design requirements, the optimization module will modify the parameters for the simulation result design and return to the pre-processing for the simulation process.
The core module of CAE simulation software is the solver
The solver is the core tool for reducing simulation: the solver refers to the objective restoration of the physical laws and mathematical principles in the scene in a program-coded manner for specific scenarios (such as liquid flow, temperature propagation and structural deformation), that is, the realization of the so-called "simulation". The core competitiveness of simulation software is the matching degree between the simulation results and the real scene, that is, the matching ability of the solver results.
Underlying mathematical solutions and physical logic: The mathematical physics model of the underlying simulation usually has two solution methods, the analytical solution and the numerical solution. The analytical solution is precise and efficient, but in reality many problems cannot be solved analytically (i.e., precise answers). CAE software mainly uses numerical solutions, that is, numerical calculation methods based on structural discretization. At present, the most mainstream on the market is the use of the finite element method (FEM) for calculations.
Technical point splitting
Simulation software development is divided into 4 layers from top to bottom:
Layer 1 is the upper application layer, which belongs to the computer front-end processing;
Layer 2 is the pillar or core content and framework of the upper-layer application of industrial software;
Layer 3 is the core technology of the overall industrial software;
Layer 4 belongs to the underlying basic theory.
Even if the development of simulation software revolves around multiple technical points, with the support of funds and manpower, it is not difficult to complete the development of an industrial software product in the short term (1-2 years). What really needs to be paid attention to is the data accumulation, scene restoration, degree of restoration of engineering problems, and iteration of underlying basic research during the long-term use of the product.
Technical pain points and potential solutions
The technical pain points of CAE software are mainly simulation errors, and the reasons for this error are divided into the underlying calculation error and the low solution efficiency.
Underlying calculation error:
Physical model error + mathematical calculation error;
Human error: Because CAE software requires the user to have a lot of experience in the professional fields such as model establishment, simulation parameter setting, process experience and so on to successfully output, the error caused by human operation on the one hand increases the difficulty of software rollout, on the other hand, it also makes the output more unstable and reliable.
The solution is too inefficient:
Pre-processing errors: When CAD and CAE do not connect and cause complex models to change between software, operators need to spend a lot of time to build models, deal with topology errors, re-mesh and other pre-processing work;
The solution time due to insufficient hashing power is too long.
There are four main solutions.
Increase the authenticity of physical modeling and the accuracy of results:
3D rendering engine;
A breakthrough in the mathematical theory of underlying physics.
Reduce the factor of manual debugging:
Add AI to I/O (input and output) for optimization.
Integrated CAD/CAE:
BY actively developing dedicated interfaces for each CAD software, CAE software imports high-quality models, reduces the difficulty of geometry cleaning, and enhances the software's pre-processing capabilities.
Cloud-based software:
Increase business synergy;
Save local hash rate and solve time.
02
Industry markets
Market capacity
Market size: Credence research data, the global CAE market size is expected to reach $8.1 billion in 2020, and it is expected to reach $12.8 billion in 2025, with an average annual compound growth rate of 9.6%. According to CCID Consulting's 2016-2021 China industrial software market size data, the scale of China's CAE market will reach 1.8 billion yuan in 2020, and it is expected to reach 2.1 billion yuan in 2021, with an average annual compound growth rate of 16.2%, which is significantly higher than the global growth rate.
Market CR3: The top three suppliers in the global CAE market in 2020 are Siemens, ANSYS and Dassault Shrews, with a market share of 47%. China's local simulation solution providers include Zhongwang Software, Hollevo and so on. In the Chinese market, the market share of the global CAE big three is more than 95.7%, and the localization rate is less than 5%.
Market drivers
Policy encouragement (see APPENDIX1): Since the 21st century, the government has successively issued a series of supportive policies to promote the development of the CAE industry. Such policies are conducive to promoting the development of CAE technology, improving the industry standard system, promoting the application of CAE software industry, and accelerating the establishment of an industrial ecosystem, which is of great significance for helping downstream industrial enterprises to intelligentize and informatize the process.
Demand upgrade: The ostensible reason for weak domestic demand is the prevalence of pirated software. With the improvement of domestic copyright awareness and industry competition, domestic CAE will gradually enter the market by relying on perfect functions and high cost performance. The development of domestic high-end manufacturing will provide CAE with enough local simulation software market to fundamentally drive the development of the domestic CAE industry.
Basic technical support: CAE/CAD is the foundation for Industry 4.0, digital twins and digital delivery. With the rise of cloud computing, digital twins, and the industrial Internet, CAE provides a more powerful and close segmentation scenario. Compared with the software of the traditional C/S architecture, the software of the B/S architecture can complete the calculation task more quickly in a more favorable subdivision scenario
Market development
The development of the industrial software industry can be divided into three stages:
The stage of software's own development;
Collaborative application stage of software, business process realization collusion and optimization stage;
From providing customers with a single tool to provide customers with a "software + services" overall solution.
Development stage of domestic manufacturers: China's CAE industry is in the first stage and the second stage, and the transformation to collaborative application has become the general trend. At present, a large number of core industrial software such as design and manufacturing in China are occupied by foreign brands. Under the background of the transformation and upgrading of China's manufacturing industry, industrial enterprises have begun to accelerate the pace of integration of the two modernizations (the integration of industrialization and informatization) and gradually change the development model.
The development stage of overseas manufacturers: The development of the foreign CAE industry is already in the third stage. Leading enterprises such as Dassault Siemens and other leading enterprises have realized the technical accumulation of software itself, and realized the application synergy of software in the practice of national industrialization, and are transforming and upgrading to the overall solution of "software + service".
Head player ecology Mapping
History of foreign head players
Overseas industrial software manufacturers were mostly established in the 1960s, and began to develop rapidly in the 1990s. The overseas giants that have developed so far have product applications in all walks of life, and have a strong first-mover advantage in the versatility, reliability and stability of CAE products.
Industry giants use frequent acquisitions of technology companies in niches to complement their own ecological closed loops to achieve a more complete tool chain. The acquisition not only realizes the shortcomings of different CAE algorithm functions, but also reflects the trend of gradual integration of CAD/CAE/CAM. Throughout the first quarter of 2020 to 2021, there were 15 major acquisitions in the CAE market. Among them, Altair completed the acquisition of 7 companies, which is the largest number of acquisitions; followed by Hexagon (4 companies) and Ansys (2 companies).
03
Track financing
Major players and situations in China
Comparison of domestic and foreign leaders: Zhongwang VS Dassault
Business model: Compared with Dassault's long-term transformation to a subscription system and a modular sales business model, Zhongwang Software is still relying on the revenue of buying out Licenses, and it can be expected that due to the monetization cycle restrictions of the current business model, Zhongwang will also transform to a subscription system in the later stage.
Overall scale: Compared with Dassault, it can be seen that the overall scale of Zhongwang Software is still small, but the revenue growth rate is higher, with the growth of the domestic high-end manufacturing market and the tide of industrial software, it is foreseeable that Zhongwang (Chinese manufacturers) still has a high market ceiling.
Domestic player development trend
Changes in the market situation: the industry is accelerating
Since the beginning of the 21st century, domestic CAE software has gradually begun to develop, but it is limited by two aspects, and mature overseas products have completed commercial landing in China, and the intention of user replacement is low; the internal application is limited by the demand scenario and quantity. Since entering 2006, in order to meet the needs of "Made in China", domestic CAE software companies have entered the industry vision from the secondary development of foreign software business.
Product development: CAX integration
After the industrial software is applied from a single item to achieve full coverage and in-depth penetration of various business links such as R&D, production, operation, logistics and other business links of manufacturing enterprises, it gradually develops in the direction of comprehensive integration to achieve business processes and production and operation mode changes. Creating a systematic software solution that runs through the front and back end of industrial production, so as to improve the management efficiency and accuracy of the whole process has become a new direction for the development of industrial software. After the CAD/CAE integration, CAE simulates and analyzes the CAD model, and repeatedly corrects the original design or model through the feedback data to achieve the best results.
Historical changes in financing: industrial funds enter the market
1998-2009: Domestic CAE software has just started, most of the companies established during this period have experienced a long research and development cycle and basically after 2010 to gradually open up financing opportunities to the equity market, most of the companies have become the leader of the domestic industry, but there are only a few listed companies.
2009-2019: With the encouragement of national policies and the upgrading of industrial resources, more high-tech talents choose to start businesses in the CAE track. However, it can be seen that many enterprises use simulation computing as the entrance and target a larger market than industrial software, such as cloud computing, industrial Internet, digital twins, etc. The backgrounds of the investors of these emerging companies are also more diverse, and in addition to government funds, CVC is not far behind. Manufacturing companies such as BYD and Internet companies such as Huawei are constantly laying out in the simulation software track.
Investment characteristics
M&A-oriented combined policy-driven investment strategies are likely to become mainstream.
CAE track belongs to a segment with a high degree of product standardization, and the products are highly reproducible and versatile. However, the software functions and the underlying algorithm need to be iterated for a long time, and the payment cycle is lengthened for capital;
CAE belongs to the typical track of "high barriers, long cycle, small market, large investment, and unclear exit channels", and is not particularly suitable for the segmentation of VC institutions to enter;
However, since CAE simulation is one of the most critical links in the entire industrial field, most institutions need to combine industrial status and policy support to find potential development opportunities and growth points and pay special attention to the corresponding segments.
04
Automotive scenarios are relevant
Application scenario: Automobile manufacturing
Automobiles are currently the most important application scenarios of CAE software. According to marketintellica, 36% of the revenue in the global CAE market in 2019 comes from the application of the automotive industry, followed by electrical and electronics, aerospace and defense, accounting for 22% and 21% respectively.
Automotive Development Scenarios:
System dynamics analysis: mainly analyze the drivability and maneuverability of the car, and often use multi-body (multi-rigid body, multi-flexible) system dynamic analysis methods;
Fatigue life analysis: Automobile fatigue life analysis mainly studies the dynamic and static fatigue life of the whole vehicle and its components;
Collision analysis: Collision analysis methods mainly include finite element method, multi-rigid body system dynamics method river mechanical vibration method. Automobile crash analysis mainly conducts the study of collision resistance of body structure, collision biomechanics and occupant restraint system and safety interior parts;
NVH analysis: The NVH analysis methods commonly used in engineering are: multi-rigid body system dynamics method, finite element method, boundary element method, statistical energy analysis method;
Aerodynamic analysis: Mainly analyzes the aerodynamic noise of automobiles at high speed, and analyzes the influence of airflow field on the control stability of the airflow field of high-speed driving of automobiles.
Key CAEs in automotive development:
Pre- and post-processing software: Heperworks, Patran
Crash analysis: LS-Dyna, Pam-Crash, Radioss
Mechanical Dynamics Simulation: Adams
Nonlinear analysis: Abaqus, Marc, Ansys
Fatigue Analysis: Fatigue
Fluid Analysis: Fluent, Star-cd, AVL-Fire
Forging Process Analysis: SuperForge
Predictive analysis of noise in cars: Akusmod
Multidisciplinary Intelligent Optimization: iSight
Automotive Automation Modeling: SOFY
Engine Thermodynamic Analysis: GT-Power
Vehicle performance analysis: GT-Drive
Application scenario: Autonomous driving
Since another trend in the development of in-vehicle hardware is getting smaller and smaller, simulation can be combined with hardware, similar to today's FPGA programming. Autonomous driving is a typical example, autonomous driving is actually a real-time simulation system, cameras, sensors to collect data in real time. Based on the collected data, the processor calculates and predicts the driving route and adopts the corresponding strategy, relying on the transmission of data and the simulation of the data in this process. If the collision simulation software is solidified in the car hardware, the car can simulate the results of collisions with other cars in real time during the driving process, so that automatic driving can take more favorable prevention and avoidance measures when encountering danger, and its significance is self-evident.
Currently, simulation functions are used in test and validation decision modules and path planning modules in the development of autonomous driving algorithms. However, the capabilities of the current simulation environment remain to be developed. For example, Carla and LGLVS, which are based on game engines, try to use computer graphics models, rendering algorithms, and physical models to try to produce a high-fidelity environment. However, this is only based on the rendering capabilities of the underlying engine to visually achieve "high fidelity", from the physical feedback, running logic, structural perception and other aspects of the current virtual simulation to really fit the real environment there is still a gap.
Appendix: National industrial software related policies