Dynamic Thermal Management Research Articles

Investigation of Engine Exhaust Heat Recovery Systems Utilizing Thermal Battery Technology

Over 50% of an engine’s energy dissipates via the exhaust and cooling systems, leading to considerable energy loss. Effectively harnessing the waste heat generated by the engine is a critical avenue for enhancing energy efficiency. Traditional exhaust heat recovery systems are limited to real-time recovery of exhaust heat primarily for engine warm-up and fail to fully optimize exhaust heat utilization. This paper introduces a novel exhaust heat recovery system leveraging thermal battery technology, which utilizes phase change materials for both heat storage and reutilization. This innovation significantly minimizes the engine’s cold start duration and provides necessary heating for the cabin during start-up. Dynamic models and thermal management system models were constructed. Parameter optimization and calculations for essential components were conducted, and the fidelity of the simulation model was confirmed through experiments conducted under idle warm-up conditions. Four distinct operational modes for engine warm-up are proposed, and strategies for transitioning between these heating modes are established. A simulation analysis was performed across four varying operational scenarios: WLTC, NEDC, 40 km/h, and 80 km/h. The results indicated that the thermal battery-based exhaust heat recovery system notably reduces warm-up time and fuel consumption. In comparison to the cold start mode, the constant speed condition at 40 km/h showcased the most significant reduction in warm-up time, achieving an impressive 22.52% saving; the highest cumulative fuel consumption reduction was observed at a constant speed of 80 km/h, totaling 24.7%. This study offers theoretical foundations for further exploration of thermal management systems in new energy vehicles that incorporate heat storage and reutilization strategies utilizing thermal batteries.

Stretchable Metamaterials with Tunable Infrared Emissivity for Dynamic Thermal Management.

Manipulation of infrared emissivity, which is closely related to surface structure and optical parameters of materials, is a crucial approach for realizing dynamic thermal management. In this study, we design a metamaterial consisting of an array of aluminum disks embedded on a surface of a stretchable elastomeric substrate. Mechanical stretching-induced deformation allows dynamic modification of the surface structure and equivalent optical parameters, thus enabling dynamic control of the emissivity. By utilizing the elastomer polydimethylsiloxane (PDMS) as the substrate, the microstructure interdisk gap can be altered by stretching the PDMS. Through theoretical calculations, the plausibility of this approach is explained by the excitation of plasmon resonance and the variation in the exposed area of highly absorbent PDMS, and the optimal structures for tuning the infrared emissivity are revealed to be 6 μm in diameter and 100 nm in height. Based on this design, we prepare samples with periods of 7 and 7.9 μm and experimentally demonstrate that a change in the period can cause a change in the emissivity and thus tunability in thermal control performance. The temperature difference between the two samples reaches 44.1 °C at a heating power of 0.28 W/cm2 for both samples. Furthermore, we construct a stretching platform that enables in situ mechanical stretching to realize dynamic changes in emissivity. The integral infrared emissivity of the sample increases from 0.32 to 0.5 at a biaxial tensile strain of 13%, achieving a 56% modulation rate of the integral infrared emissivity. The material is expected to enable dynamic thermal management.

NeuroTAP: Thermal and Memory Access Pattern-aware Data Mapping on 3D DRAM for Maximizing DNN Performance

Deep neural networks (DNNs) have been widely adopted, owing to break-through performance and high accuracy. DNNs exhibit varying memory behavior involving specific and recognizable memory access patterns and access intensity, depending on the selected data reuse in different layers. Such applications have high memory bandwidth demands due to aggressive computations, performing several billion-floating-point-operations-per-second (BFLOPs). 3D DRAMs, providing very high memory access bandwidth, are extensively employed to break the memory wall , bridging the gap between compute and memory while running DNNs. However, the vertical integration in 3D DRAM introduces serious thermal issues, resulting from high power density and close proximity of memory cells, and requires dynamic thermal management (DTM). To unleash the true potential of 3D DRAM and exploit the enormous bandwidth under thermal constraints, there is a need to intelligently map the DNN application’s data across memory channels, pseudo-channels, and banks, minimizing the effective memory latency and reducing the thermal-induced application slowdown. The specific memory access patterns exhibited by a DNN layer execution are crucial to determine a favourable data mapping method for 3D DRAM dies that potentially causes minimal thermal impact, and also maximize DRAM bandwidth utilization. In this work, we propose an application-aware and thermal-sensitive data mapping that intelligently assigns portions of the 3D DRAM to DNN layers, leveraging the knowledge about layer’s memory access patterns and minimizing DTM-induced performance overheads. Additionally, we also deploy a DRAM low-power states based DTM mechanism to keep the 3D DRAM within safe thermal limits. Using our proposal, we observe a performance improvement of 1% to 61%, and memory energy savings of 1% to 55% for popular deep neural networks over state-of-the-art DTM strategies while running DNN inference.

A Janus textile mimics leaf stoma for dynamic thermal regulation through a shape memory coating

The development of smart textiles with thermal/moisture regulation is crucial for outdoor activities and healthcare applications. However, incorporating dynamic adaptive thermal management into conventional fabrics to provide consistent comfort remains a significant challenge. Inspired by the ingenious stomata of plant leaves, a novel Janus-type cotton fabric is fabricated by covalently linking shape memory polymer (SMP) on the surfaces of cotton fibers via the combination of the “grafting from” strategy and the “mist polymerization” methodology. This unique coating induces yarns to dynamically twist in response to temperature, switching the pore apertures in the fabric and allowing for adaptive adjustment of both heat dissipation capability and moisture permeability. The resulting fabrics provide effective thermal insulation at low temperatures, while demonstrating an enhanced cooling effect at high temperatures. Meanwhile, the asymmetrical distribution of SMP across the fabric enables unidirectional water transport, improving sweat removal and cooling. Additionally, the excellent biocompatibility and fabric flexibility allow for comfortable garment fabrication. This work provides novel insights into the field of self-adaptive thermal regulation, demonstrating promising potential for smart textile applications.

Autonomous Thermal Modulator Based on Gold Film‐Coated Liquid Crystal Elastsomer

AbstractRadiative cooling has been recently intensively explored for thermal management and enhancing energy efficiency. Yet, traditional materials with singular emissivity fall short in dynamic thermal management, highlighting the need for materials that can adjust their thermal radiation in real time. Active modulation methods, requiring external stimuli such as mechanical stretch, electric potential, or humidity change, offer adaptability but can increase energy use and complexity. Passive approaches, using materials' inherent thermal‐responsive properties, face manufacturing and scalability challenges. Here, a scalable yet effective passive approach is introduced for adaptive thermal modulation based on gold (Au) and liquid crystal elastomer (LCE) with a reversible response to environmental temperature changes. This modulator enables a “low thermal resistance” state through actuation‐induced microcracks that expose a high‐emissivity polymer substrate, and a “high thermal resistance” state by closing these microcracks and forming a high thermal resistance air gap between the modulator and the target object. The flexible design and fixed external dimensions of the Au‐LCE thermal modulator make it adaptable to various surface geometries. Furthermore, by adjusting the LCE's chemical composition, the modulator's transition temperature can be tailored, broadening its applications from enhancing building energy efficiency to improving clothing thermal comfort.

The thermal 'Buddha Board'-application of microstructured polyolefin films for variable thermal infrared transparency materials.

In this study, we explore the innovative application of biological principles of scattering foams and structural colouration of white materials to manipulate the transmission properties of thermal infrared (IR) radiation, particularly within the 8-14 μm wavelength range in polyolefin materials. Inspired by the complex skin of organisms such as chameleons, which can dynamically change colour through structural alterations, as well as more mundane technologies such as Buddha Boards and magic water colouring books, we are developing methods to control thermal IR transmission using common thermoplastic materials that are semi-transparent to thermal IR radiation. Polyethylene and polypropylene, known for their versatility and cost-effectiveness, can be engineered into microstructured sheets with feature sizes spanning from 5 to 100 μm. By integrating these precisely moulded microstructures with index-matching fluids, specifically IR transparent oils, we achieve a reversible modification of the thermal transmission properties. This novel approach not only mimics the adaptive functionality of natural systems but also offers a practical and scalable solution for dynamic thermal management. Our results indicate a promising pathway for the development of new materials that can adapt their IR properties in real time, paving the way for smarter thermal management solutions via radiative emission/absorption.

Continuously‐Tunable and Ultrawide‐Range Thermal Regulator Based on Superaligned Carbon Nanotube Aerogels for Dynamic Thermal Management of Batteries and Buildings

AbstractEfficient heat transfer control is highly demanded for dynamic thermal management of equipment and buildings especially when the environmental temperature dramatically changes. Thermal switches, the current approach of heat transfer control, suffer from the issues of low switching ratios below eight and sharp state transition between “on” and “off”. Herein, a continuously‐tuned thermal regulator with an ultrahigh thermal‐conductivity change ratio of 43 based on superaligned carbon nanotube aerogel is reported, which works on the regulation of both thermal interfacial resistance and conduction pathways by compressive deformation‐induced microstructure evolution. This thermal regulator can stabilize the device temperature at 25 °C when the environmental temperature varies by 7 and 11 °C under the natural convection and forced convection conditions, respectively. Toward practical application, the thermal regulator with flexibility is wrapped around a cylindrical lithium‐ion battery to control the operation temperature within the optimal range of 20–40 °C for enhanced discharging performance even at an environmental temperature of −20 °C. Besides, by combining the thermal regulator with radiative cooling film, the house model temperature can be lowered by 2.7 °C during daytime and raised by 1.1 °C during nighttime compared with the bare one. This efficient thermal regulation approach offers an effective solution for practical thermal management.

Bioinspired Stretchable Polymers for Dynamic Optical and Thermal Regulation

Elaborated optical and thermal modulatory systems are of great importance to the survival and evolution of organisms in nature. Inspired by these natural intelligent systems, researchers have made great efforts for developing stretchable polymers and exploring their applications in fields of communication, dynamic camouflage, thermal management, and others. Herein, an up‐to‐date account of the advancements in bioinspired stretchable polymers for dynamic optical and thermal regulation is provided. First, stretchable polymers for dynamic structural colors are presented, including cholesteric liquid crystal elastomers, photonic crystal elastomers, and emerging photonic polymers. Then stretchable polymers for dynamic infrared emissivity are introduced, which are achieved by stretch‐induced wrinkled‐flat surface or stretch‐induced cracked surface. Third, stretchable polymers for dynamic thermal management are discussed, focusing on tunable solar transmittance and dynamic radiative cooling. Moreover, the perspectives on the opportunities and challenges for future research directions of bioinspired stretchable polymers are presented at the end.

A Facile and Effective Design for Dynamic Thermal Management Based on Synchronous Solar and Thermal Radiation Regulation.

Passive solar heating and radiative cooling have attracted great interest in global energy consumption reduction due to their unique electricity-free advantage. However, static single radiation cooling or solar heating would lead to overcooling or overheating in cold and hot weather, respectively. To achieve a facile, effective approach for dynamic thermal management, a novel structured polyethylene (PE) film was engineered with a switchable cooling and heating mode obtained through a moisture transfer technique. The 100 μm PE film showed excellent solar modulation from 0.92 (dried state) to 0.32 (wetted state) and thermal modulation from 0.86 (dried state) to 0.05 (wetted state). Outdoor experiments demonstrated effective thermal regulation during both daytime and nighttime. Furthermore, our designed PE film can save 1.3-41.0% of annual energy consumption across the whole country of China. This dual solar and thermal regulation mechanism is very promising for guiding scalable approaches to energy-saving temperature regulation.

Porous vanadium dioxide thin film-based Fabry−Perot cavity system for radiative cooling regulating thermochromic windows: experimental and simulation studies

Abstract Radiative cooling in smart windows using VO2 – a dynamic thermal management material, is of potential interest for enhancing energy savings in buildings due to its both solar and emittance tuneability in response to changing temperatures. However, studies related to the effects of VO2 thin film microstructure in a multilayer system on emissivity regulation are currently lacking. The present study addresses the thermochromic and emissivity performance of VO2/ZnSe/ITO/Glass Fabry−Perot (F–P) cavity thin film system, by manipulating the porosity in VO2 thin film. The device is fabricated by commercially feasible physical vapor deposition methods such as sputtering and thermal evaporation, most suitable for mass production. The optimized sample with porous VO2 delivers an enhanced long-wave infrared (LWIR) emissivity contrast of Δɛ LWIR ≥ 0.4 preserving a high visible transparency T lum(avg) of ∼41 % compared to dense VO2. Then finite difference time domain (FDTD) simulation is performed to further understand the effects of varying VO2 porosity and ZnSe thickness on the F–P cavity properties. The reduced low-temperature ɛ LWIR (0.1–0.2) gives this film better energy saving in regions where warming demand is dominant as simulated by EnergyPlus.

Modulation of near-field radiative heat transfer between nanoparticles supported by a strained hBN film with graphene covered

Dynamic modulation of near-field radiative heat transfer (NFRHT) requires changing the optical responses of the object, which is expected to be realized with the help of the strain engineering. Based on the electric dipole approximation and the reflection Green's function, in this work, we investigate the effect of the hBN film under uniform in-plane strain on the NFRHT between two graphene/SiC core-shell (GSCS) nanoparticles. By changing the magnitude of the strain on the film, the hyperbolic phonon polaritons (HPPs) can be dynamically tuned to match and deviate from the localized surface resonance (LSR), thus significantly modulating the NFRHT between the nanoparticles. When the two GSCS nanoparticles are placed above a strained hBN film, an amplification of the maximum radiative heat flux by about 6 orders of magnitude can be achieved. In addition, after the introduction of graphene to form graphene/hBN heterostructure, the radiative heat flux can be further enhanced with the help of the hybrid polaritons. Especially at a suitable chemical potential and strain, the NFRHT can be enhanced by 4 more orders of magnitude compared to the bare hBN film. This work provides a new insight into efficient and dynamic thermal management among nanoparticles.

Patterned liquid metal embedded in brush-shaped polymers for dynamic thermal management.

Interface thermal resistance has become a crucial barrier to effective thermal management in high-performance electronics and sensors. The growing complexity of operational conditions, such as irregular and dynamic surfaces, demands thermal interface materials (TIMs) to possess high thermal conductivity and soft elasticity. However, developing materials that simultaneously combine soft elasticity and high thermal conductivity remains a challenging task. Herein, we utilize a vertically oriented graphene aerogel (VGA) and rationally design liquid metal (LM) networks to achieve directional and adjustable pathways within the composite. Subsequently, we leverage the advantages of the low elastic modulus and high deformation capabilities of brush-shaped polydimethylsiloxane (BPDMS), together with the bicontinuous thermal conduction path constructed by VGA and LM networks. Ultimately, the designed composite of patterned liquid metal/vertically oriented graphene aerogel/brush-shaped PDMS (LM-VGA/BPDMS) shows a high thermal conductivity (7.11 W m-1 K-1), an ultra-low elastic modulus (10.13 kPa), excellent resilience, and a low interface thermal resistance (14.1 K mm2 W-1). This LM-VGA/BPDMS soft composite showcases a stable heat dissipation capability at dynamically changing interfaces, as well as excellent adaptability to different irregular surfaces. This strategy holds important application prospects in the fields of interface thermal management and thermal sensing in extremely complex environments.

NeuroCool: Dynamic Thermal Management of 3D DRAM for Deep Neural Networks through Customized Prefetching

Deep neural network (DNN) implementations are typically characterized by huge datasets and concurrent computation, resulting in a demand for high memory bandwidth due to intensive data movement between processors and off-chip memory. Performing DNN inference on general-purpose cores/edge is gaining attraction to enhance user experience and reduce latency. The mismatch in the CPU and conventional DRAM speed leads to under-utilization of the compute capabilities, causing increased inference time. 3D DRAM is a promising solution to effectively fulfill the bandwidth requirement of high-throughput DNNs. However, due to high power density in stacked architectures, 3D DRAMs need dynamic thermal management (DTM), resulting in performance overhead due to memory-induced CPU throttling. We study the thermal impact of DNN applications running on a 3D DRAM system, and make a case for a memory temperature-aware customized prefetch mechanism to reduce DTM overheads and significantly improve performance. In our proposed NeuroCool DTM policy, we intelligently place either DRAM ranks or tiers in low power state, using the DNN layer characteristics and access rate. We establish the generalization of our approach through training and test datasets comprising diverse data points from widely used DNN applications. Experimental results on popular DNNs show that NeuroCool results in a average performance gain of 44% (as high as 52%) and memory energy improvement of 43% (as high as 69%) over general-purpose DTM policies.

Dynamic Thermal Management of 3D Memory through Rotating Low Power States and Partial Channel Closure

Modern high-performance and high-bandwidth three-dimensional (3D) memories are characterized by frequent heating. Prior art suggests turning off hot channels and migrating data to the background DDR memory, incurring significant performance and energy overheads. We propose three Dynamic Thermal Management (DTM) approaches for 3D memories, reducing these overheads. The first approach, Rotating-channel Low-power-state-based DTM (RL-DTM) , minimizes the energy overheads by avoiding data migration. RL-DTM places 3D memory channels into low power states instead of turning them off. Since data accesses are disallowed during low power state, RL-DTM balances each channel’s low-power-state duration. The second approach, Masked rotating-channel Low-power-state-based DTM (ML-DTM) , is a fine-grained policy that minimizes the energy-delay product (EDP) and improves the performance of RL-DTM by considering the channel access rate. The third strategy, Partial channel closure and ML-DTM , minimizes performance overheads of existing channel-level turn-off-based policies by closing a channel only partially and integrating ML-DTM, reducing the number of channels being turned off. We evaluate the proposed DTM policies using various mixes of SPEC benchmarks and multi-threaded workloads and observe them to significantly improve performance, energy, and EDP over state-of-the-art approaches for different 3D memory architectures.

Optimization of an Air-cooling Thermal Management System for Lithium-ion Battery Packs via Particle Swarm Algorithm

Recently, lithium-ion batteries have attracted many researchers and their safety issues such as overheating, combustion and explosion continue to further limit battery application scenarios. These issues are mainly caused by unoptimized battery structure parameters or cooling methods. In this paper, an integrated approach has been proposed to design an efficient air-cooling system using the particle swarm algorithm to find an optimal relationship between air flow rate and battery temperature. Firstly, this method can adjust an optimized air flow rate to ensure that the battery temperature is minimized with the lowest energy consumption via the particle swarm algorithm. Additionally, an optimized air flow rate can still be obtained with the change of structure parameters such as the radius in a lithium-ion battery pack via this novel algorithm. Then, we demonstrate the feasibility of this integrated method in simulations. Compared with the previous work, this method can employ the continuous modulation of the particle swarm algorithm, realizing both the best cooling capacity of the battery cooling system and simultaneously the lowest energy consumption for cooling in cell heat regulation systems. Meanwhile, temperature variations of the entire cell pack are also shown in simulations. In contrast to previous approaches, this integrated method may provide more dynamic thermal management inspirations for designing novel battery thermal management systems.

Thermochromic Conductive Fibers with Modifiable Solar Absorption for Personal Thermal Management and Temperature Visualization.

Thermal management textiles provide an energy-efficient strategy for personal thermal comfort by regulating heat flow between the human body and the environment. However, textiles with a single heating or cooling mode cannot realize temperature regulation under dynamic weather. Furthermore, monocolor textiles do not satisfy aesthetic requirements in a garment. Here, we develop a thermochromic (TC) conductive fiber with a coaxial structure composed of a conductive core and thermochromic shell. The TC conductive fiber-woven fabric has the ability of low-energy dynamic thermal management by combining Joule heating and modulation of solar absorption. Compared with commercial white fabrics, TC conductive fabrics exhibit a maximum temperature drop of 2.5 K, while the temperature of colored commercial fabrics is 7.5-16 K higher than that of commercial white fabrics in the hot. In the cold, the combination of Joule heating and the photothermal effect can provide desired thermal comfort for humans. Meanwhile, heat obtained from solar absorption brings the temperature of a fabric to a predetermined level, which saves energy of 625 W/m2 compared to a conductive-fiber-based textile. In addition, TC conductive fabrics with trichromatic evolution provide a sensitive and instant temperature visualization capable of identification of invisible and intense infrared radiation. These results provide another path to expand potential applications of wearable, flexible electronics.

Dynamic thermal environment management technologies for data center: A review

The energy demand of the data center (DC) industry has accounted for 2% of the total global energy consumption, and its operating power consumption has reached 50 times that of the commercial buildings. Therefore, improving the thermal environment performance to reduce the energy cost is major trend to realize the energy savings of DC. It has become a positive measure to integrate smart technologies to improve the operating efficiency of HVAC systems, especially due to the development of informatization and intelligence. This study provides detailed overview of the dynamic thermal environment management technologies of the DCs. The requirements of thermal environment in DCs were reviewed by different levels, which are the basis of the design and operation. Accordingly, various dynamic prediction technologies were presented concerning the theoretical mechanics, the computation efficiency and the applicable circumstances. The performance of the PID, the model predictive control and the reinforcement learning in the dynamic thermal environment control of DC were analyzed. Moreover, the evaluation systems based on cooling capacity, exergy analysis and entransy analysis were introduced respectively to measure the performance of the thermal environment. By investigating the knowledge related to the “requirement, prediction, control, and evaluation” of the DC dynamic thermal environment, this study aims to provide reference and guidance for energy conservation, airflow organization and smart operation of DCs.

Multifunctional wearable electronic textile based on fabric modified by MXene/Ag NWs for pressure sensing, EMI and personal thermal management

Flexible wearable electronics with multi-functions are bound to become the future development trend in the fields of intelligent device application, involving human healthcare, dynamic thermal management and even radiation protection. Herein, we promote a facile method that combines solution immersion and suspension spraying to assemble a poly(dimethylsiloxane)/Ag NWs/MXene/fabric (PAM@fabric) device. In particular, we use poly (diallyl dimethyl ammonium chloride) (PDAC) to pretreat the cotton fabric, which enables MXene to realize electrostatic self-assembly and greatly improve its stability. The multifunctional textile-based device with three-dimensional conductive network structure of MXene and Ag NWs presents durable and sensitive pressure sensing performance for detecting human joint motions and even subtle pulse. Besides, due to the excellent electrical conductivity of MXene and Ag NWs, the electronic also exhibits outstanding electromagnetic interference (EMI) shielding, as well as electrothermal conversion performance with controllability and extended cyclic stability. Accordingly, rational design provides a feasible path to prepare flexible multifunctional wearable devices.

SoC-Modified Core Temperature Estimation of Lithium-Ion Battery Based on Control-Oriented Electro-Thermal Model

Lithium-ion batteries and their control technologies are the key points to electrification and intelligence of transportation. Dynamic thermal management is one of the key technologies for intelligent battery management systems. Real-time monitoring of information about the temperature characteristics inside the battery is important for effective and safe thermal management. This paper firstly constructs a distributed control-oriented electro-thermal coupling model which contains multidimensional internal information about the cell. Based on the proposed model, improved parameter identification methods are used to construct offline database of model parameters. The electrical and thermal parameters are identified by applying recursive least squares with variable forgetting factor (VFFRLS) and particle swarm optimization (PSO) separately. Finally, a SoC-modified core temperature estimation method is proposed, which adopts discrete-time nonlinear observer (DNLO) to modify SoC and adapative Kalman filter (AKF) to estimate core temperature. The method takes into account the sensitivity of the output results to nonlinear, time-varying battery systems. The results show that the root-mean-square error (RMSE) of SoC estimation is 1.75 % and the mean absolute error (MAE) is 0.65 % for the proposed temperature method under wide temperature points (-5, 25, 45 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula> C). The proposed core temperature estimation method possesses better robustness and universality, with RMSE of 0.61 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{\circ }$</tex-math></inline-formula> C and MAE of 0.56 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^ {\circ }$</tex-math></inline-formula> C. Compared with the open-loop prediction method, the accuracy is improved about 0.5 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^ {\circ }$</tex-math></inline-formula> C under extreme loadfiles with uncertainty.

Adaptive Machine Learning-Based Proactive Thermal Management for NoC Systems

Because of the high-complex interconnection in contemporary multicore systems, the network-on-chip (NoC) technology has been proven as an efficient way to solve the communication problem in multicore systems. However, the thermal problem becomes the main design challenge in the current NoC systems due to the high-diverse workload distribution and large power density. Therefore, proactive dynamic thermal management (PDTM) is employed as an efficient way to control the system temperature. Based on the predicted temperature information, the PDTM can control the system temperature in advance to reduce the performance impact during the temperature control period. However, conventional temperature prediction models are usually built based on specific physical parameters, which are usually temperature-sensitive. Consequently, the current temperature prediction models still result in significant temperature prediction errors. To solve this problem, a novel adaptive machine learning (ML)-based PDTM is proposed in this work. The adaptive ML-based PDTM first uses an adaptive single layer perceptron (ASLP), which is composed of a single-neuron operation and a least mean square (LMS) adaptive filter technology, to precisely predict the future temperature. Afterward, the proposed adaptive reinforcement learning (RL) is used to find the proper throttling ratio to control the system temperature. In this way, the proposed adaptive ML-based PDTM can adapt to the hyperplane of the temperature behavior of the NoC system and provide a proper temperature control strategy at runtime. Compared with related works, the proposed approach reduces average temperature prediction error by 0.2%–78.0% and improves the system performance by 2.4%–43.0% with smaller hardware overhead.