基于FPGA的腔式BPM数字信号处理算法研究

Free electron laser (FEL), known as the most potential fourth generation light source after the synchrotron radiation, has wide application prospects in scientific, industrial, medical and military fields. One of the key components to ensure the high brightness and coherence of the FEL facilities is to use the cavity beam position monitors (CBPM) with the resolution as small as of the order of nanometer. There are two major categories of the CBPM based on different quality factors: high Q mode and low Q mode. Commercial CBPM signal processor are still not available up to now. Although many electronics have been developed by institutes all around the world, they are only designed to process the signals from their specified mode and the application is quite limited. Nevertheless, the signal processing procedures of the different modes do not share the same algorithm. The ability of the beam diagnostics would be significantly improved when a general-purpose CBPM signal processor can be developed to meet the requirements on sundry occasions. FPGA chip is becoming the core of the signal processing instruments since it has abundant programmable logic resources with high-speed parallel processing ability. This project will study the algorithms of the signal processing procedure with different CBPM configurations and various accelerator operation modes and their FPGA implementation to achieve the ability to switch the algorithm accordingly on-line and to improve the measurement performance. Laboratory experiments as well as FEL field tests will be carried on to verify the system and have further optimization. The accumulated technology will be reserved to develop the general-purpose CBPM processor, also the technology could be spreaded to the measuring of various machine parameters.

自由电子激光在科学、工业、医学、军事等领域具有广阔的应用前景，被誉为继同步辐射光源之后最有希望的第四代光源。分辨率达到nm级的腔式BPM是保证自由电子激光装置高亮度、高相干性的关键设备之一。 腔式BPM按品质因数可分为高Q值和低Q值两类。目前还没有商业腔式BPM信号处理器，各实验室开发的处理电子学也只针对各自的高Q值或低Q值探头设计，使用范围有限。同时，装置在不同运行模式时，信号处理算法也有差异。如果能研制通用腔式BPM信号处理器，满足不同应用需求，将极大提高束流诊断能力。 FPGA有丰富的逻辑资源，可进行高速并行处理，逐渐成为数字信号处理设备的核心。本项目利用FPGA进行不同腔式BPM和不同机器运行模式信号处理算法研究及实现，可根据情况在线切换算法，提高测量性能，通过实验室测试和带束测试进行优化。项目为通用腔式BPM处理器研究做关键技术储备，其中的技术也可推广到其他机器参数的测量上去。

自由电子激光由于其高亮度、高相干性有极大应用价值，近两年国内已建成大连DCLS和上海SXFEL两个自由电子激光装置。其中，高分辨率的束流位置测量设备是保证自由电子激光装置的高亮度、高相干性的关键设备之一。腔式 BPM 系统轨道测量分辨率达到数十nm量级，成为自由电子激光装置束流位置测量的最佳选择。腔式BPM信号经过前端处理后进行数字化采样与数字信号处理获得通道信号幅度，经归一化计算得到数量位置信号。国内已开展了腔式BPM探头设计研究，但在信号处理系统，包括射频前端、数字信号处理算法、和IOC软件方面比较薄弱，没有可用于腔式BPM信号处理的数据采集处理设备可供使用。.本项目研究了针对高Q和低Q探头的数字信号处理算法，其中曲线拟合算法适合离线使用，重点研究了FFT算法和Hilbert算法及其实现，两种算法分别从频域和时域进行信号处理。在以FPGA为核心的自行研制数据采集处理器上，同时实现了FFT和Hilbert算法，实验测试表明，两种算法能成功实现腔式BPM探头的信号处理，但当前测试下两种算法的处理性能差异并不明显，测试分辨率约为0.4μm。此外，Hilbert算法也可用于条带BPM信号处理，实现了处理器的通用化应用。.此外，对与腔式BPM信号处理相关的前端处理技术、标定平台、EPICS IOC上层应用等各关键环节进行了研究，开发出满足实际需求的相应设备并投入在线使用，促进了与腔式BPM工程相关技术的进步。.将来自由电子激光装置可能运行在束团串模式或高重频，对信号处理提出了新的挑战，以上研究成果的成功应用为将来通用处理器的研制打下了较好的基础。