Next-Generation Battery Materials Development Enabled by Data Science

As countries around the world commit to achieving carbon neutrality by 2050, the adoption of electric vehicles (EVs)—a cornerstone of that transition—is accelerating.

The electrification of automobiles depends fundamentally on the development of next-generation batteries. Batteries that surpass today’s mainstream lithium-ion technology in both performance and cost will enable longer driving ranges and lower vehicle prices, driving the shift away from gasoline-powered and other internal combustion engine vehicles toward electric mobility.

Against this backdrop, competition to develop next-generation batteries is intensifying, and DENSO has been actively advancing its own R&D efforts. The question then becomes: what kind of approach is required to accelerate the development of battery materials?

Keisuke Kobayashi of the R&D Division points out that conventional materials development processes face inherent limitations.

“In frontline materials development, it has long been common to rely on the experience of individual engineers, iterating experiments one by one to identify suitable conditions and meet technical requirements. However, such craftsmanship-driven approaches are time-consuming and often lack sufficient theoretical grounding. If critical parameters are overlooked, development must be repeated—driving up both time and cost. To avoid this, we must fundamentally rethink and redesign the development framework itself.”
ー Kobayashi

In next-generation battery materials development, an enormous number of parameters—temperature, humidity, pressure, mixing ratios, and more—directly influence performance. As a result, the technical requirements for improving performance and quality become increasingly complex.

In response, DENSO undertook a comprehensive transformation of its development process by integrating data science. The initiative brought together the R&I Division and the Advanced Testing & Evaluation Division, both part of the R&D Center.

The Advanced Testing & Evaluation Division participates from the earliest planning stages, incorporating a manufacturing perspective to translate engineers’ concepts into practical solutions. Through this role, it has established agile development processes and effectively served as a partner that turns ideas into reality.

To further accelerate materials development for next-generation batteries, DENSO has now made a decisive shift toward a development approach centered on data science.

To accelerate the development of materials, it is crucial to systematically collect, accumulate, and analyze data that can be meaningfully interpreted. To achieve this, we introduced AP+DN7 (Analysis Platform + Digital Native Quality Control 7 Tools).

AP+DN7 is a data analysis platform developed by DENSO and released free of charge to support the effective use of manufacturing process big data (https://sites.google.com/view/analysisplatform-dn7). It enables data ingestion from multiple sources, preprocessing, integration, and analysis.
The platform also incorporates DN7, a DX-based evolution of the traditional Quality Control Seven Tools*, offering a wide range of analytical functions tailored to process data. This enables the rapid application of big data insights to process improvement.

*DX-Based Quality Control Seven Tools
A reconfiguration of the traditional Quality Control Seven Tools—seven foundational tools for quality control—from the perspective of digital technology and DX (digital transformation). This next-generation toolset actively incorporates AI and IT to support data utilization, visualization, and frontline problem-solving.

The adoption of data science, however, was not without challenges. One of the most significant hurdles was designing new workflows in collaboration with frontline R&D teams. Shogo Hayakawa of the Materials Laboratory in the Advanced Testing & Evaluation Division explains.

“Because ‘data accumulation’ was not part of conventional workflows, it initially placed an additional burden on frontline operations, making implementation difficult. We therefore designed a system in which data is captured naturally within existing workflows and can be used directly for analysis.

This approach reduced rework and helped boost frontline motivation. New development methods do not take root unless engineers can clearly experience their benefits. By improving efficiency and reducing workload, we were also able to steadily build trust with frontline teams.”
ー Hayakawa

As a result of these efforts, cumbersome tasks involving handwritten records, spreadsheets, and server-based management were significantly streamlined through dedicated applications and systems. According to Kaito Yamaguchi of the Materials Laboratory, Advanced Testing & Evaluation Division, this also led to reductions in prototyping effort and material costs.

“Previously, low reproducibility often resulted in repeated trials, increasing both cost and labor. By incorporating data analysis, we can now precisely identify the processes where issues occur and the specific areas requiring corrective action. This enables targeted interventions, reducing rework and contributing directly to cost savings.”
ー Yamaguchi

To ensure that data-driven materials development is firmly embedded in frontline operations, human resource development is equally critical. In the Advanced Prototyping Section, five engineers are currently certified at the Shodan level under DENSO's internal Manufacturing DX Personnel* qualification system. This fiscal year, three additional members will focus on data-driven initiatives.

“By helping even non-certified members understand how data science transforms our work—and allowing them to experience its benefits firsthand—we aim to build an organization in which everyone can adopt this way of working.”
ー Hayakawa

*Personnel capable of utilizing and analyzing manufacturing process data. Classified into four skill levels: Literacy → Introductory → Shodan → Master

The materials development methodologies enabled by data science are not limited to next-generation batteries. They are expected to find applications across a wide range of fields. One such example is the next-generation Solid Oxide Electrolysis Cell (SOEC) system that DENSO has been developing.

SOEC is a next-generation water electrolysis technology designed to efficiently produce hydrogen using renewable energy. Compared with conventional low-temperature electrolysis, SOEC offers fundamentally higher efficiency in environments where low-carbon electricity is consistently available and waste heat can be utilized.

• Related article: DENSO is challenging itself to develop a solid oxide electrolysis cell system looking ahead to a future of hydrogen society

Like batteries used in electric vehicles, SOEC operates on the fundamental principle of efficiently transporting ions. As a result, data science techniques can be applied across research and development of materials and manufacturing methods aimed at improving both performance and reliability.

Addressing the technical challenges of SOEC is essential to realizing the carbon-neutral, circular society DENSO envisions. Keisuke Kobayashi of the R&I Division’s Research Planning Department outlines this outlook.

“There is no single path to achieving carbon neutrality; multiple technologies must be developed in parallel. By leveraging data science to advance development more efficiently and within shorter timeframes, we aim to contribute to the realization of a carbon-neutral society.”
ー Kobayashi

In advancing data-driven materials development, DENSO draws on its extensive manufacturing knowledge accumulated over decades. Hayakawa elaborates on this advantage.

"At DENSO, we do more than convert material properties into data—we develop with a clear view of how those materials ultimately function as part of a system. No matter how advanced one’s data science expertise may be, it has little value unless it translates into real manufacturing outcomes. Because we understand the realities of the frontline, we know which parameters truly matter and what data must be captured. This perspective is a defining strength of DENSO.”
ー Hayakawa

In practice, critical insights often reside beyond numerical data on the frontline: the equipment in use, the sequence of operations, and the tacit knowledge and atmosphere of the workplace—factors not captured by metrics such as temperature or humidity. Incorporating these realities into data analysis leads to more effective and actionable improvements.

“It is essential to work on two axes: what can be explained by data, and what reflects the realities of the frontline that data alone cannot capture. Our goal is not digitalization for its own sake, but data utilization grounded in a deep understanding of manufacturing.”
ー Yamaguchi

Looking ahead, DENSO will continue to strengthen data science–driven materials development and accelerate the creation of technologies essential to a carbon-neutral society, including next-generation batteries and SOEC systems. By balancing data-driven methodologies with the inheritance of manufacturing craftsmanship cultivated over decades, DENSO remains committed to redefining the future of manufacturing.