组织人：余亮教授 孟岭超副教授 张超教授（民航学院副院长）
报告题目：Bayesian Deep Learning or Deep learning with Bayesian principles
摘要：In this seminar,I discuss about Neural Nets(NN),Deep Learning(DL),Bayesian inference,Bayesian DL and Deep Learning with Bayesian principles. The main content of this talk,follows:
1，Introduction to NNs. Whats are good for?
2，Learning by training, testing, evaluating and using NNs
3，Main established tasks of NN and Deep Learning
4，Bayesian inference and computational methods
5，Deep Learning and/or Bayesian Learning?
6，Deep Learning with Bayesian principles
7，Available tools, challenges
阿里 嘉法理（Ali Djafari）
报告人简介：阿里 嘉法理（Ali Djafari）博士曾是法国国家科学研究中心研究主任，法国巴黎第十一大学终身教授（全球大学排名第14、法国第1位，拥有4位诺贝尔奖、5位菲尔兹奖），智能探测领域资深专家，为贝叶斯推断方法在工业智能诊断应用起到了极为关键的作用。他提出的高斯-波特图像分割算法、快速贝叶斯变分法、超参数贝叶斯推断方法等，受到国际学术和工业界高度认可。在无损探测、机械故障诊断、医学图像识别和工业大数据分析领域等广泛应用，相关技术被欧洲空中客车、法国泰勒斯、达索飞机、法国核电等直接采用，支持项目超过1千万欧元。发表论文300余篇、2部专著、8部合著、12本教材、培养博士生21位、硕士生31位。现为浙江上风高科专风实业股份有限公司（SZ000967）战略科学家、总工程师，首位浙江省引进的产业类国际顶尖人才、受到浙江省主要领导接见、浙江省西湖友谊奖候选人。
报告题目：Modal composition integrated fast-rotating sound source localization under Bayesian framework for axial fans
摘要：The modal composition beamforming (MCB) method can quickly achieve the localization of rotating blade sound sources, however, its relatively low resolution and unclear application range lead to certain limitations in its practical application. Therefore, this study aims to investigate in depth the MCB-based high-resolution localization method for rotating multiple sound sources and apply it to the field of axial-flow fan noise reduction. The study employs both variational Bayesian approximation (VBA) and subspace variational Bayesian (SVB) methods to solve the MCBbased rotating sound source power propagation (RSP) model in a Bayesian framework. In this talk, the RSP model is experimentally validated for the first time. In addition, a pioneering parameter selection scheme is proposed by investigating the applicable boundary of the MCB method, and its effectiveness is verified by simulation. Also, for the sound source localization results, three evaluation metrics are proposed, and the processing times of the related simulations are given to assist in verifying the above scheme. The proposed method of highresolution localization of rotating sound sources is fast in calculation, effective and high in resolution, and can realize the leaf tip noise localization of multi-blade high speed axial fans. And the proposed MCB parameter selection scheme can effectively guide the acoustic measurements and the practical application of the method.
报告题目：Multiphase Modelling quantifies Trade-Offs in Porous Media Drying: Kinetics, Energy Consumption, Quality and Environmental impacts.
The current study demonstrates the contribution of numerical simulations to gaining further insights and improving the life cycle assessment (LCA) of a convective drying process. The LCA findings indicate that energy consumption accounts for approximately 80% of the environmental impacts associated with drying. The research recommendations include targeting hotspots and reducing overall or specific emissions by reducing electricity consumption or transitioning to renewable energy sources. While minor environmental indicators are linked to agricultural production and transportation, the study highlights the drying process as an energy-intensive process and, thus, a primary contributor to the climatic footprint of food.
Understanding the interplay of thermophysical properties, heat and mass transfer, and mechanical coupling in porous media allows for optimising energy consumption, enhanced efficiency, and improved material quality. Furthermore, the presented methodology can be extended to similar sectors, exploring the wide-ranging applications of porous media beyond agricultural products and drying processes. Porous media, consisting of materials with interconnected void spaces, have relevance in various sectors such as geology, environmental engineering, chemical engineering, filtration processes, groundwater management, and biomedical applications. By integrating quality attribute prediction and expanding databases, stakeholders can make well-informed decisions to foster the implementation of eco-friendly industrial processes.
1.Overview of porous media drying processes and technologies.
2.The current state of energy-efficient drying technologies.
3.Physics-Based Digital Twins and various modelling approaches.
4.Decarbonisation of production systems and environmental assessments.
5.Challenges and future outlook.
报告人简介:Pursuing a PhD in Refrigeration and Cryogenics at Zhejiang University, with a Master's degree in Thermal and Renewable Energies Engineering from “École Nationale Supérieure d'Arts et Métiers” at Moulay Ismail University in Morocco. Solid background in advanced mathematics, thermodynamics, mechanical design and manufacturing, and heating and cooling systems. Fluent in English, bilingual in French and Arabic, with moderate proficiency in Chinese, and basic knowledge of SpanishMultiphase Modelling quantifies Trade-Offs in Porous Media Drying: Kinetics, Energy Consumption, Quality and Environmental impacts.Demonstrated leadership and teamwork abilities as a Team Leader and Co-founder of the Energy Club at Moulay Ismail University in Morocco, where mentored new members, organized workshops, and collaborated on innovative energy-harnessing concepts. Served as an English tutor at Zhejiang University English Centre, assisting students with academic writing and speaking skills. Gained practical experience in industrial fluid flow behaviour and heat transfer through work at the R&D Direction of the OCP Group and other prestigious companies and industries in Morocco. Passionate about sustainable engineering and dedicated to developing environmentally friendly solutions. Strong communicator and team player, excelling in interdisciplinary projects. Eager to
contribute knowledge and make meaningful impacts.
Simulation of aerodynamic noise source, sound transmission and sound impedance of blowers
风机的主要噪声一般为气动噪声，要真实地预测风机噪声水平，需从噪声源的产生和噪声的传播两方面进行计算。莱特希尔声类比方法（Lighthill’s Acoustic Analogy）是目前应用最广泛的气动噪声计算方法，先通过计算流体力学（Computational Fluid Dynamic, CFD）对流场进行求解后，通过声类比法将流场数据转化为声源，从而得到噪声源结果。该方法还可计算声场辐射，得到远场噪声结果。在上述方法的基础上，引入专业的声学计算方法，被称为混合计算方法（Hybrid Method）。通过声类比方法得出的噪声源结果转化为等效声源，并通过快速傅里叶变换（Fast Fourier Transform, FFT）将时域数据转化为频域数据。有限元或边界元的方法将声源在结构中的反射、散射和吸收等效应进行计算与后处理，即声场计算。最后，对风机系统中的降噪组件进行参数化后，考虑声阻抗作为一部分声学边界条件，对噪声的传播进行计算，以更真实地预测噪声水平和噪声控制效果。
The main source of the blowers’noise is aerodynamic noise. In order to predict the noise level of the blower, it is necessary to calculate both the generation of noise source and the propagation of noise. The most wildely used metohd for calculating aerodynamic noise is Lighthill's Acoustic Analogy. Firstly, solve the flow field by Computational Fluid Dynamic (CFD). Then, the flow filed data is converted into sound source by analogy method, yielding the result of noise source in time space. This method can also calculate the sound field radiation to get the far-field noise result. Based on the above method, a professional acoustic calculation method is introduced, which is called the Hybrid Method. The noise source results obtained by analogy method are transformed into equivalent sound source, and the time domain data is transformed into frequency domain data by Fast Fourier Transform (FFT). The finite element method or boundary element method is used to calculate the acoustic field, considering the effect of reflection, scattering and absorption of the structure to the sound. Noise control measures are usually included in blower systems, so the sound impedance should be parameterized and introduced into the acoustic boundary conditions in a acoustic field simulation. Combined those methods, the noise level and the noise control effect can be realistically predicted.