What is "thousands of people and thousands of faces"?
In the past, the engine design used to cover most application scenarios and working conditions (traction, flat panel, special vehicle and dump truck, high-speed, urban and suburban working conditions, etc.) through a set of designs. In different application scenarios and working conditions, some performance of the engine may have limitations, while the customer hopes that the engine can adapt to different application scenarios and actual application conditions to optimize the performance as much as possible.
Based on the big data algorithm, Cummins has developed a set of "running portraits" to guide the customized design of engines and "private customization" for each customer. By understanding the key information of the engine, big data analysis and algorithm evaluation of engine operating conditions, OTA technology is used to dynamically adjust software calibration to adapt to different operating conditions, optimize performance and fuel consumption, and improve attendance.
Next, let's watch how to realize customized development!
In the era of Internet of vehicles and digitalization, more insights can be gained through big data online, and users can have a deeper understanding of how to use engine products, thus providing unlimited possibilities for customized product development and optimization. In order to achieve this goal, the segmentation of vehicle use scenarios is particularly important. The key to the segmentation of vehicle use scenarios is to list meaningful feature labels, and to divide the type of feature labels into different dimensions according to the business meaning. Based on these labels, customization is realized in engine calibration design, hardware selection and test verification:
Customized calibration design
1. big data model analysis
Based on big data, engineers sorted out market segments and analyzed the actual operating conditions of different market segments. Analyze the characteristics of tens of thousands of parameters in the engine calibration library, and develop a recommended system method for engine calibration, so as to better guide the engineering team to customize and distribute the calibration.
The method of the recommendation system is to take the calibration database in the operation of the existing market as the training set by comparing the classification regression tree and the principal component analysis, convert the customer demand into various parameter characteristics, and predict the value of each parameter through the establishment of the machine learning model, so as to customize the optimal calibration for customers in different market segments. This method not only improves the quality of Basic Calibration and development efficiency, but also allows different customers to experience differentiated engine performance.
2. calibration, customization and distribution
After completing the calibration customization, through the OTA function of the Internet of vehicles, the targeted calibration will be issued to vehicles in different market segments, so that each market segment has its own calibration to achieve better engine performance.
Customized hardware selection optimization
1. big data model analysis
After establishing the market segment classification, the big data model is used to sort out the operating conditions of the engine, and these actual operating conditions are used as the input in the engine development, so as to guide engineers to better select and optimize the key components of the engine, such as air compressor, aftertreatment system and supercharger system.
2. customized hardware selection optimization
After the analysis of the Internet of vehicles and big data, it provides design boundary conditions for engine development, better guides engineers to select and optimize engine hardware, and provides customers with engines with better performance, more reliability and lower carbonization.
Customized test validation optimization
1. big data model analysis
According to the market segment classification, the actual working conditions and operation data analyzed by big data are combined with engine design verification standard (ESW) and failure mode analysis (FMEA), so as to better guide engineers to carry out engine design verification planning and optimize engine road test.
For example, select more areas with high engine failure or typical operating conditions to test, rather than average all operating conditions. In addition, some bench tests and actual road verification in engine development are assisted or replaced by combining vehicle networking data and big data analysis and software simulation.
2. customized road test optimization
The Internet of vehicles and big data analysis are adopted to provide actual operating boundary conditions for engine development and verification, and better guide engine design verification and road test. By identifying the actual operating conditions of the engine, effective experiments can improve product reliability and quality, and ensure the reliable application of customers under various operating conditions.

Source: Cummins China