Internal combustion engine: combustion mechanism analysis to 1D model construction
Complex system optimization is required to meet the various performance requirements of internal combustion engines (thermal efficiency, output, emissions, NV, etc.). Combustion development, where physical phenomena and chemical reactions are intricately intertwined, is said to be the most difficult area to predict by numerical calculation.
In particular, phenomena such as misfire and knocking have a large impact on system performance, and the ability to analyze their causes quickly and accurately enables efficient system development with minimal rework. We can provide a wide range of support, from mechanism analysis to performance factor analysis, analysis tool development and 1D simulation model construction.
Optimizing robot motion
We support the introduction of model-based control that utilizes model information of the controlled object, and the associated parameter tuning of feedback controllers.
We also provide motion command generation technology to realize highly accurate and efficient robot operation. Specifically, we optimize trajectories and velocity profiles to suppress residual vibration after reaching the target position while maintaining stability during operation, and to reach the specified position in the shortest possible time. We optimize control according to the characteristics of robots and actuators, contributing to improved reliability and performance at manufacturing sites and in precision equipment.
Combustion chemical reaction analysis to lightweight model development
“Ignition delay”, “laminar burning velocity” and “chemical equilibrium” are said to be important parameters for estimating intermediate indicators of combustion performance such as knocking, misfire and emissions. These have nonlinear characteristics with respect to physical quantities and require combustion chemical reaction calculations.
General combustion phenomena are intricately intertwined with flows such as turbulence, but compared with fluid calculations, chemical reaction calculations take an enormous amount of time, which tends to make numerical calculation difficult.
We not only calculate chemical reactions and examine mechanisms, but also build predictive models from the calculation results, contributing to the reduction of calculation cost.
Building control laws through reinforcement learning
We apply the theory of reinforcement learning to build flexible controllers for nonlinear systems, making it possible to handle targets that were difficult with conventional control theory. What is called a “policy” in reinforcement learning is positioned as a “control law” in control engineering, and it is acquired through learning.
We also have experience building controllers with inverse reinforcement learning, which uses expert control data to aim for better performance while keeping behavior close to that of the expert.
Chemical Reaction Simulation Development Service
We provide high-precision numerical analysis tools for a wide range of fields involving chemical reactions, including combustion research, engine development, catalytic reactions and materials processing.
We design and implement systems flexibly according to your requirements — GUI applications, cloud systems, parallel-processing platforms and batch analysis workflows. Cantera serves as the computational core, supporting advanced analysis needs such as reaction mechanism handling, reaction path analysis, pressure and temperature dependency evaluation and large-scale parameter sweeps.
We develop a broad set of capabilities such as ignition delay time, laminar burning velocity, adiabatic flame temperature, reciprocating engine cycle analysis, YAML mechanism file generation for Cantera and model reduction. Building on these technologies, we can extend features or develop new applications tailored to your product development and research themes.
AI-Powered Automated Document Generation
By combining AI and OCR technologies, we provide software that automatically extracts the necessary information from handwritten notes, scanned documents and other sources, and populates it into specified formats such as application forms, reports and business templates.
It significantly reduces manual entry, transcription and verification work, achieving both operational efficiency and quality.
Because it adapts flexibly to diverse document formats, it contributes greatly to automating back-office operations and optimizing internal workflows.