Keimei Fujita, TOYOTA MOTOR CORPORATION
Toyota must prepare many variations of products and units to meet a wide variety of customer needs in different countries and regions. This requires a huge number of tests during the development process, which means more time spent analyzing data. Furthermore, the team must spend additional time collecting accurate information to reutilize data for other tests and model development.
DIAdem was introduced as a standard tool to create an analysis environment that could be shared for use by each department. The departments could quickly customize and develop necessary functions. By introducing DataFinder Server Edition, the team created a data management system that anyone from any department can use.
Keimei Fujita - TOYOTA MOTOR CORPORATION
Shinji Hattori - TOYOTA MOTOR CORPORATION
In 2016, Toyota Motors integrated everything from development to production and switched over to a new system to implement a directive known as Creating a Better Car. The most prominent characteristic of this new directive was to implement a system focused on the products. Toyota started Creating a Better Car at seven companies, including the powertrain company. The role of a powertrain is to create power for cars to run and to manage the flow of power to the tires. As the engine/motor generates power, a transmission toggles that power and a differential distributes it among other main components.
In accordance with this new company system, the team set up the development, production technology, and manufacturing departments within the powertrain company. They conducted the process from planning to production consistently; however, they still needed to perform various types of required tests, including a functionality assessment, compliance, and reliability assessment.
During the functionality assessment, they measure various physical quantities for each engine and decide whether improvements are needed. As for compliance, they conduct tests to bring out the performance required from each engine and derive optimal solutions regarding the details of the control method. In the reliability assessment, they conduct tests to confirm the suitability of the design and compliance.
One of the main factors in deciding the types and number of tests needed is the regulations in countries and regions around the world. For example, only four countries had regulations concerning fuel costs in 2010. That number had increased to 13 in 2015, and is expected to increase to 23 by 2020. More countries impose new regulations every year, and the regulation values are becoming stricter. This results in more complicated and advanced products, and more advanced control is required as the number of hardware parts increases. Toyota conducts more than 1,500 tests to develop a single engine.
Having expanded to 53 manufacturing entities in 28 countries/regions, in 2015 Toyota sold over 5X more vehicles overseas compared to the number sold in Japan. To continue expanding their global business, Toyota must meet the needs of different countries and regions, including basic form, emissions, performance, and drive systems. This means the total number of tests had to increase by the number of tests needed for one engine multiplied by the number of variations.
After the number of required tests skyrocketed, Toyota discovered one primary (and serious) issue. They needed more time to analyze the data obtained, which took up an extremely high proportion of work time. This increase in analysis time also meant less time to focus on new tasks. Therefore, reducing and optimizing analysis time became extremely important.
The analysis team performed conventional analysis tasks. They temporarily sent test data obtained from each test room to a file server and stored it. Next, the team downloaded the data from the file server onto computers, then sorted and graphed the data. Lastly, the team carried out the management of this data individually, with various people using different tools to conduct the analyses.
This does not mean the team made no effort to reduce analysis time. Departments would prepare their own tools to sort and graph the data, but this method did not develop laterally. That meant that any other department needing to perform analysis had to do it manually, which resulted in a high volume of man-hours.
Furthermore, it wasted time and effort for each department to individually attempt to improve efficiency because the biggest factor for not laterally developing was that multiple tools were already in use. These tools included Microsoft Excel, The MathWorks, Inc. MATLAB® software, and internal tools (each different across departments). This led to a reluctance to transition to other tools.
As for the test data itself, Toyota faced another issue. Because they did not manage this data in a manner in which they could reutilize it, they could not conduct functionality assessments in which a team could refer to data like temperature or pressure obtained from a compliance test. Furthermore, when conducting reliability analyses, a team would sometimes want to use data from the compliance tests to determine conditions for future tests. Alternatively, in preparation for next-term developments, a team would sometimes want to use the data from the compliance test to identify the models used in model-based development and ensure accuracy. To achieve this reutilization, the team must be able to store test data in a manner that allows it to be searched.
One condition that must be met to reuse test data is the specification required for each piece of data and each set. In short, Toyota can’t reuse test data unless it includes information such as when, who, and for what purpose a test was conducted. Take cylinder pressure as an example. In its test data, detailed specification information such as the model, the amount of emissions, the number of cylinders, and the fuel must be stored in a set.
Conventionally, Toyota managed test data individually, so if someone from another department wanted to reuse the data, he or she needed to find the information through hearsay. As a result, even if he or she found the intended data, that person would need to make additional inquiries if the specification information was not clear. If that person could not find the intended data, in the worst case scenario, he or she had to conduct another test altogether. The team occasionally came across cases in which it was extremely difficult to reuse data because of wasted time, hard-to-find data, or data with unclear specification information.
Toyota introduced DIAdem and DataFinder Server Edition (now called SystemLink TDM DataFinder Module) to resolve these issues. The DIAdem software tool focuses on searching, analyzing, and displaying reports on data obtained from tests and simulations. The DataFinder Server Edition data management system focuses on the server side. It can search the server for data from the numerous clients who use DIAdem (My DataFinder).
Figure 1 shows a system Toyota built using both products. They built this system for compliance test environments. One key goal to improve analysis efficiency was introducing standard analysis tools so that anyone could perform the tasks in the same environment. They selected DIAdem for that purpose.
Furthermore, they needed two measures for data management—adding specification information and allowing anyone to search for the data. Storing data in the file server was the same as the conventional method, but the new system automatically creates a folder upon completion of a test. Then the system stores test data in that folder. The team also built a mechanism that automatically adds specification information to each piece of data in the file server. On top of that, they created an index using DataFinder Server Edition for specification information data, so they can search data from clients who have installed DIAdem. The people in charge of analyses who searched for and obtained the necessary data could then conduct analyses immediately using DIAdem. Selecting NI products gave Toyota an opportunity to get in contact with the Association for Standardization of Automation and Measuring Systems Open Data Service (ASAM ODS) in consideration of future expandability.
According to Keimei Fujita, “As the number and types of tests increases, the need for analyses also rises, which means we need highly customizable analysis tools. In that regard, we found DIAdem particularly suitable. It is hard to obtain a degree of freedom equivalent to that of DIAdem using other products. We could also use the existing file server to manage the data, which left few obstacles to introducing DataFinder Server Edition.”
As mentioned, the team could use NI products to create mechanisms to resolve conventional issues, but that was not enough. The Toyota system introduction department had been preparing various types of systems for some time and sharing them with the development department. However, the development sites were not using these new systems. Engineers at the work sites did not want to change their conventional methods. They felt burdened by learning new tools, so the systems were not disseminated.
With this understanding, Toyota put new initiatives in place to promote the dissemination of the new systems. They called the first initiative Penetration. The goal of this initiative was to identify the needs of the development sites. It put into practice the cycle of grasping the current situation, discovering issues, and applying measures by establishing close contact with a special group from the development department.
The subsequent needs identified from this initiative were reflected in the development of the analysis tools. They needed a function that allows one to find data instantly, organize it quickly, and then output it in a typical format. However, they did not use any methods that prepared functions dedicated to specific analysis tasks. Too many types of analyses used this method, which made it difficult to support. Instead, they selected a method called Integrating the Function Into Components.
They can use this method to prepare many component-based functions in advance and combine them to provide the necessary analysis functions. They can also use this method to quickly create the analysis functions that each person in charge of analysis wants to implement. In actuality, they prepared stratified function groups. Stratified functions include common functions for use across multiple domains, specific functions for use in specific domains (for engines, and more), and individual functions to use for specific tasks. Each function was prepared in a Toyota-specific menu created on DIAdem, and then the functions could be combined freely.
Toyota also conducted internal training as an initiative for dissemination. They held a group training once a month. It promoted the dissemination of the new system by teaching basic DIAdem usage methods as well as methods that followed actual practices, and also by introducing case studies.
Taking all factors into consideration, the team successfully shortened problematic analysis times using the initiatives described above. “One department, for example, reduced the man-hours needed for analysis by about 50 percent. Furthermore, being able to reuse test data meant the team could eliminate unnecessary tests and improve the accuracy of model-based development,” says Fujita. In addition, the team could conduct analyses from various perspectives in a short time, so they also benefitted from having more opportunities to hold meetings while looking at analysis results, which made discussions more dynamic.
As mentioned previously, Toyota is conducting a multitude of tests. They do not have countermeasures for all those tests; therefore, they would like to expand the application range of the new system in the future. For example, they plan to introduce the CompactDAQ hardware platform for data collection and LabVIEW system design software for function assessment tests. Thus, teams at the test sites can perform various measurements and collect data using CompactDAQ. They can organize and analyze typical data, and automatically add specification information using programs developed with LabVIEW. Additionally, teams can store processed data on the file server, and combine DataFinder Server Edition and DIAdem to deliver functions similar to compliant examples.
Because NI offers a variety of open, software-centric platform products, Toyota feels confident in this plan moving forward. They believe they can create solutions that cover a wider domain by using NI products, which have a high degree of connectivity and interoperability with other companies' products and technologies.
Keimei Fujita
TOYOTA MOTOR CORPORATION
Japan