AI Under the Microscope: Reliability in Safety-Critical Sectors
.png?width=200&height=200&name=rasyiqah%20photo%20(3).png)
Enhancing Thermal Solutions with Smart Sensors using Physics-Informed Neural Networks
Battery State Estimation in Electric Vehicles: A Data-Driven Machine Learning Approach
The advent of machine learning (ML) has led to an exponential rise in the exploration of new methods to estimate the states of lithium-ion batteries (LIBs) in electric vehicle (EV) applications. Data-driven methods involving ML are increasingly engaged to estimate the state-of-charge and state-of-health due to greater availability of battery datasets in the public domain, alongside improvements in computing system efficiency. At present, the battery management system of EVs tend to face challenges in attaining highly accurate state estimation results using traditional methods, since the LIBs possess strong time-varying and non-linear traits, whilst remaining susceptible to the influence of external factors. In light of this issue, this presentation reviews the existing ML methods for state estimation of EV application-based LIBs and provides a proof-of-concept workflow for these methods. The insights gained from this presentation will contribute towards the future development of advanced, LIB state estimation algorithms.
Reduced Order Modeling with AI
In this talk, learn about different ROM techniques and methods, covering AI-based approaches, linear-parameter varying (LPV) modeling, and strategies for bringing large-scale sparse state-space matrices from FEA tools into MATLAB® and Simulink® for applications such as flexible body modeling and control. The focus of the talk, however, will be on AI-based ROM. See how you can perform a thorough design of experiments and use the resulting data to train AI models using LSTM, neural ODE, and nonlinear ARX algorithms. Learn how to integrate these AI models into your Simulink simulations, whether for hardware-in-the-loop testing or deployment to embedded systems for virtual sensor applications. Learn about the pros and cons of different ROM approaches to help you choose the best one for your next project.
Developing Energy-Efficient Solutions for Cooling Data Centers
Are you looking for solutions to cool buildings in a cost and energy-efficient manner? Are you struggling to simulate various components of a building to test out various energy consumption scenarios? In this talk, we will use the example of a data center to show how to develop cooling systems and how we can map computational tools to different stages of a technology development cycle. You will see how to create synthesized data from a data center cooling simulation over a one-year period that can be used to optimize technical vs economic factors. Our demo will also show an optimization layer to a data center cooling simulation that can optimize key thermal system parameters, and an optimized energy management system on the microgrid system that will update operational setpoints based on day-ahead forecasting.
Demystifying Optimization for Industry in Battery Management, Grid Control, and PLC Systems using MATLAB
In sectors crucial to modern infrastructure, such as battery management and grid control, reluctance towards optimization remains a significant hurdle. Mathematical and meta-heuristic optimization methods are non-deterministic, often yielding multiple outputs for the same scenario. Moreover, the addition of AI introduces further uncertainty to the optimization process. This presentation delves into MATLAB's strategies for overcoming this industry hesitation, focusing on battery simulation, grid control, and energy storage systems (ESS). In the realm of research and industrial engineering, MATLAB is a transformative force, revolutionizing traditional approaches. It elucidates how MATLAB emerges as a game-changer, providing unparalleled capabilities, especially in industrial scenarios where PLC serves as an essential workbench. By leveraging MATLAB's extensive toolkit, both researchers and engineers can efficiently navigate complex grid dynamics, optimize energy distribution, and enhance grid stability with unprecedented precision. Supported by illuminating case studies, attendees witness how MATLAB empowers researchers and industry engineers to break barriers, pioneer new frontiers, and shape a sustainable future through optimization.
Hybrid Digital Twin Architecture for Power System Cyber Security Analysis
This talk explores into an innovative Hybrid Digital Twin (HDT) architecture, emphasizing its crucial role in enhancing the cyber security of power system. By leveraging on MATLAB-SIMULINK digital model in conjunction with single-board computers, the HDT provides a highly configurable, scalable, and cost-effective platform for rigorous cyber security analysis on power system communication protocols such as IEC 61850. The talk will outline the methodology for establishing the HDT, demonstrating how it can simulate real-world vulnerabilities and test security measures without compromising critical grid operations. Beyond power systems, the HDT's versatile framework can be adapted for other industries, such as manufacturing, to ensure robust and secure operational environments across various sectors.
Smart Motor Control: Streamline Algorithm Implementation and Tuning
Developing motor control systems can be complex, but using a robust simulation environment can significantly simplify this process. Such an environment allows for the thorough validation of control algorithms and supports the generation of compact, optimized code directly from the models.
Learn how to develop motor control algorithms quickly with reference examples. Implement an automated workflow for Field-Oriented Control that updates based on motor parameter inputs, eliminating the need for redeployment. This enables you to deploy production-ready embedded software on target hardware and verify the performance of the algorithms across different motor ratings.
The workflow is further enhanced with specialized control blocks optimized for code generation, along with tools for decoding sensor signals and implementing observers. Additionally, features like motor parameter estimation and automatic controller tuning make this approach an essential resource for engineers aiming to efficiently develop and fine-tune motor control systems.
.png?width=200&height=200&name=romeo%20(3).png)
Hardware-Ready: Taking MATLAB/Simulink Algorithms to FPGA and SoC
Want to deploy your MATLAB and Simulink algorithms to FPGA or SoC? This session will walk you through the key steps and considerations.
Developing algorithms in an integrated environment allows you to effectively explore, refine, test, verify, and deploy your ideas. Learn how to transition your algorithm from simulation to FPGA/SoC hardware, tackling challenges such as fixed-point math and performance vs area tradeoffs.
Key topics include leveraging MATLAB as a reference for HDL (Hardware Description Language) implementation, understanding hardware architectures, working with fixed-point implementations, generating and optimizing HDL code, and verifying the HDL code on FPGA/SoC.
By the end of the session, you will understand the workflow from simulation to deployment, addressing common challenges and optimizing your algorithms for FPGA and SoC hardware.
Engineering in the Cloud 'Unlocking New Possibilities'
Polyspace for C/C++ Code Verification in the Cloud
Unleashing Power of Heterogeneous Compute with MATLAB
Explore the capabilities of AMD Versal adaptive SoCs, which integrate processing cores, programmable logic, and AI Engines into a versatile, software-programmable platform. This session will delve into how these heterogeneous compute platforms can be effectively used for a variety of high-performance applications. By combining programmable logic and AI Engines, these devices enable the development of cutting-edge solutions across different industries.
Maximizing Data Processing with Versal AI Engines
Discover how Versal AI Engines can exponentially enhance data processing performance for DSP and AI applications. This session will highlight practical use cases where AI Engines boost performance by dozens of times compared to traditional CPU-based or FPGA PL-based solutions. Learn how these engines can simultaneously process multiple AI models, leading to more efficient and powerful data processing.
Understand the advanced capabilities of Versal SoCs, which include rich programmable logic resources, high-speed crypto IPs, and 800G Ethernet MACs. This session will showcase how these features can be leveraged to provide fast, secure, end-to-end data processing within a single chip. Additionally, explore how the hardened PCIE/QDMA subsystem enables a high-speed data transfer link between x86 servers and Versal chips, offloading and accelerating CPU workloads.
Learn about the fully automated and reusable design flow utilizing MATLAB, Vitis, Vivado and Platforms. This session will demonstrate how these tools and reference designs can significantly reduce the design cycle for Versal-based solutions from months to days. Attendees will gain insights into streamlining their development processes for faster time-to-market.
Real-World Solutions Enabled by Versal SoCs
This presentation will provide concrete examples of complex solutions developed using Versal SoCs. From Digital Pre-distortion (DPD) and 4G/5G signal processing to AI implementations like ResNet18, attendees will see how Versal technology is driving innovation in various fields. Learn how to leverage these solutions to meet the demands of modern applications and improve performance across different sectors.
.png?width=200&height=200&name=Xue%20Chunhua%20(1).png)