Real-time Artificial Intelligence for Accelerator Control: A Study at the Fermilab Booster
Jason St. John, Christian Herwig, Diana Kafkes, Jovan Mitrevski, William A. Pellico, Gabriel N. Perdue, Andres Quintero-Parra, Brian A. Schupbach, Kiyomi Seiya,… 展开作者
Jason St. John, Christian Herwig, Diana Kafkes, Jovan Mitrevski, William A. Pellico, Gabriel N. Perdue, Andres Quintero-Parra, Brian A. Schupbach, Kiyomi Seiya, Nhan Tran, Malachi Schram, Javier M. Duarte, Yunzhi Huang, Rachael Keller
备注：
16 pages, 10 figures. Phys. Rev. Accel. Beams vol 24, issue 10. Published 18 October 2021. For associated dataset and data sheet see http://doi.org/10.5281/zenodo.4088982
We describe a method for precisely regulating the gradient magnet power supply at the Fermilab Booster accelerator complex using a neural network trained via reinforcement learning. We demonstrate preliminary results by training a surrogate machine-learning model on real accelerator data to emulate the Booster environment, and using this surrogate model in turn to train the neural network for its regulation task. We additionally show how the neural networks to be deployed for control purposes may be compiled to execute on field-programmable gate arrays. This capability is important for operational stability in complicated environments such as an accelerator facility.
The extraction bucket position in the Fermilab Booster is controlled with a cogging process that involves the comparison of the Booster RF count and the Recycler Ring revolution marker. A one RF bucket jitter in the extraction bucket position results from the variability of the process that phase matches the Booster to the Recycler. However, the new slow phase lock process used to lock the frequency and phase of the Booster RF to the Recycler RF has been made digital and programmable and has been modified to correct the extraction notch position. The beam loss at the Recycler injection has bee…
Integrable optics is an innovation in particle accelerator design that potentially enables a greater betatron tune spread and damps collective instabilities. An integrable rapid-cycling synchrotron (RCS) would be an effective replacement for the Fermilab Booster, as part of a plan to reach multi-MW beam power at 120 GeV for the Fermilab high-energy neutrino program. We provide an example integrable lattice with features of a modern RCS - dispersion-free drifts, low momentum compaction factor, superperiodicity, chromaticity correction, bounded beta functions, and separate-function magnets.
At the Fermilab Booster, and many other proton facili-ties, an intense proton beam is accumulated by multi-turn injection of an H- beam through a stripping foil. The circu-lating beam scatters off the injection foil and large-angle Coulomb scattering leads to uncontrolled losses concen-trated in the first betatron period. We measure the foil scat-tering loss rate in the Fermilab Booster as a function of LINAC current, number of injection turns, and time on the injection foil. We find that current Booster operation has ~1% foil scattering loss and we make projections for the Proton Improvement …
We propose a model-independent quantity, $L/G$, to characterize non-Gaussian tails in beam profiles observed with the Fermilab Booster Ion Profile Monitor. This quantity can be considered a measure of beam halo in the Booster. We use beam dynamics and detector simulations to demonstrate that $L/G$ is superior to kurtosis as an experimental measurement of beam halo when realistic beam shapes, detector effects and uncertainties are taken into account. We include the rationale and method of calculation for $L/G$ in addition to results of the experimental studies in the Booster where we show that …
A perpendicularly biased tuneable 2nd harmonic cavity is being constructed for use in the Fermilab Booster. The cavity's tuner uses National Magnetics AL800 garnet as the tuning media. For quality control, the magnetic properties of the material and the uniformity of the properties within the tuner must be assessed. We describe two tests which are performed on the rings and on their corresponding witness samples.
It is important to have experimental methods to estimate the maximum beam intensity for the Fermilab Booster as objective input into long term program commitments. An important existing limit is set by the available rf power. This limit is difficult to set a priori because the real longitudinal impedance is not well known. The synchrotron phase at transition crossing was measured using both the mountain range plot and the direct phase measurement of the RF accelerating voltage relative to the beam, and results were consistent. They were applied to predict 6E12 maximum Booster beam intensity wi…
A perpendicularly biased 2nd harmonic cavity is being designed and built for the Fermilab Booster. Its purpose is to flatten the bucket at injection and thus change the longitudinal beam distribution to decrease space charge effects. It can also help at extraction. The cavity frequency range is 76 - 106 MHz. The power amplifier will be built using the Y567B tetrode, which is also used for the fundamental mode cavities in the Fermilab Booster. We discuss recent progress on the cavity, the biasing solenoid design and plans for testing the tuner's garnet material.
The Fermilab Booster is a fast ramping (15Hz) synchrotron which accelerates protons from 400MeV to 8GeV. During commissioning of a transverse digital damper system, it was shown that the damper could provide a measurement of the machine tune throughout the cycle by exciting just 1 of the 84 bunches with minimal impact on the machine operation. The algorithms used to make the measurement have been incorporated into the damper FPGA firmware allowing for real-time tune monitoring of all Booster cycles.
Increasing proton beam power on neutrino production targets is one of the major goals of the Fermilab long term accelerator programs. In this effort, the Fermilab 8 GeV Booster synchrotron plays a critical role for at least the next two decades. Therefore, understanding the Booster in great detail is important as we continue to improve its performance. For example, it is important to know accurately the available RF power in the Booster by carrying out beam-based measurements in order to specify the needed upgrades to the Booster RF system. Since the Booster magnetic field is changing continuo…
To date, the 120 GeV Fermilab Main Injector accelerator has accelerated a single batch of protons from the 8 GeV rapid-cycling Booster synchrotron for production of antiprotons for Run II. In the future, the Main Injector must accelerate 6 or more Booster batches simultaneously; the first will be extracted to the antiproton source, while the remaining are extracted for the NuMI/MINOS (Neutrinos at the Main Injector / Main Injector Neutrino Oscillation Search) neutrino experiment. Performing this multi-batch operation while avoiding unacceptable radioactivation of the beamlines requires a previ…
In order to meet the needs of Fermilabs planned post-collider experimental program, the total proton throughput of the 8 GeV Booster accelerator must be nearly doubled within the next two years. A system of 48 ramped corrector magnets has recently been installed in the Booster to help improve efficiency and allow for higher beam intensity without exceeding safe radiation levels. We present the preliminary results of beta function measurements made using these corrector magnets. Our goal is to use the correctors to reduce irregularities in the beta function, and ultimately to introduce localize…
The Fermilab Booster - built more than 40 years ago - operates well above the design proton beam intensity of 4x10**12 ppp. Still, the Fermilab neutrino experiments call for even higher intensity of 5.5x10**12 ppp. A multitude of intensity related effects must be overcome in order to meet this goal including suppression of coherent dipole instabilities of transverse oscillations which manifest themselves as a sudden drop in the beam current. In this report we present the results of observation of these instabilities at different tune, coupling and chromaticity settings and discuss possible cur…
The difference between the rf voltage seen by the beam and the accelerating voltage required to match the rate of change of the Booster magnetic field is used to estimate the energy loss per beam turn. Because the rf voltage (RFSUM) and the synchronous phase can be experimentally measured, they can be used to calculate the effective accelerating voltage. Also an RFSUM reduction technique has been applied to measure experimentally the RFSUM limit at which the beam loss starts. With information on beam energy loss, the running conditions, especially for the high intensity beam, can be optimized …
A new beam injection scheme is proposed for the Fermilab Booster to increase beam brightness. The beam is injected on the deceleration part of the sinusoidal magnetic ramp and capture is started immediately after the injection. During the entire capture process we impose Pdot=0 in a changing B field. Beam dynamics simulations clearly show that this method is very efficient with no longitudinal beam emittance dilution and no beam loss. As a consequence of preserved emittance, the required RF power on a typical Booster cycle can be reduced by ~30% as compared with the scheme in current operation…
Detrimental beam dynamics effects limit performance of high intensity rapid cycling synchrotrons (RCS) such as the 8 GeV proton Fermilab Booster. Here we report the results of comprehensive experimental studies of various beam intensity dependent effects in the Booster. In the first part, we report the dependencies of the Booster beam intensity losses on the total number of protons per pulse and on key operational parameters such as the machine tunes and chromaticities. Then we cross-check two methods of the beam emittance measurements (the multi-wires proportional chambers and the ionization …
Fermilab Booster synchrotron requires an intensity upgrade from 4.5x1012 to 6.5x1012 protons per pulse as a part of Fermilabs Proton Improvement Plan-II (PIP-II). One of the factors which may limit the high-intensity performance is the fast transverse instabilities caused by electron cloud effects. According to the experience in the Recycler, the electron cloud gradually builds up over multiple turns in the combined function magnets and can reach final intensities orders of magnitude greater than in a pure dipole. Since the Booster synchrotron also incorporates combined function magnets, it is…
Over the past decade, Fermilab has focused efforts on the intensity frontier physics and is committed to increase the average beam power delivered to the neutrino and muon programs substantially. Many upgrades to the existing injector accelerators, namely, the current 400 MeV LINAC and the Booster, are in progress under the Proton Improvement Plan (PIP). Proton Improvement Plan-II (PIP-II) proposes to replace the existing 400 MeV LINAC by a new 800 MeV LINAC, as an injector to the Booster which will increase Booster output power by nearly a factor of two from the PIP design value by the end of…
The Fermilab Booster is being upgraded under the Proton Improvement Plan (PIP) to be capable of providing a proton flux of $2.25^{17}$ protons per hour. The intensity per cycle will remain at the present operational $4.3^{12}$ protons per pulse, however the Booster beam cycle rate is going to be increased from 7.5 Hz to 15 Hz. One of the biggest challenges is to maintain the present beam loss power while the doubling the beam flux. Under PIP, there has been a large effort in beam studies and simulations to better understand the mechanisms of the beam loss. The goal is to reduce it by half by c…
The Fermilab booster has an intensity upgrade plan called the Proton Improvement plan (PIP). The flux throughput goal is 2E17 protons/hour, which is almost double the current operation at 1.1E17 protons/hour. The beam loss in the machine is going to be the source of issues. The booster accelerates beam from 400 MeV to 8 GeV and extracts to the Main Injector. Several percent of the beam is lost within 3 msec after the injection. The aperture at injection energy was measured and compared with the survey data. The magnets are going to be realigned in March 2012 in order to increase the aperture. …
Fermilab is committed to upgrade its accelerator complex to support HEP experiments at the intensity frontier. The ongoing Proton Improvement Plan (PIP) enables us to reach 700 kW beam power on the NuMI neutrino targets. By the end of the next decade, the current 400 MeV normal conducting LINAC will be replaced by an 800 MeV superconducting LINAC (PIP-II) with an increased beam power >50% of the PIP design goal. Both in PIP and PIP-II era, the existing Booster is going to play a very significant role, at least for next two decades. In the meanwhile, we have recently developed an innovative bea…
The development of magnetic cogging is part of the Fermilab Booster upgrade within the Proton Improvement Plan (PIP). The Booster is going to send 2.25E17 protons/hour which is almost double the present flux, 1.4E17 protons/hour to the Main Injector (MI) and Recycler (RR). The extraction kicker gap has to synchronize to the MI and RR injection bucket in order to avoid a beam loss at the rising edge of the extraction and injection kickers. Magnetic cogging is able to control the revolution frequency and the position of the gap using the magnetic field from dipole correctors while radial positio…
We have developed a computer program interfaced with the ACNET environment of Fermilab accelerators to measure energy spread of the proton beam from the LINAC at an injection into the Booster. It uses a digitizing oscilloscope and provides users an ability to configure the scope settings for optimal data acquisition from a resistive wall current monitor. When the program is launched, a) a ${\it one\ shot\ timeline}$ is generated to initiate beam injection into the Booster, b) a gap of about 40 ns is produced in the injected beam using a set of fast kickers, c) collects line charge distribution…
In synchrotron machines, the beam extraction is accomplished by a combination of septa and kicker magnets which deflect the beam from an accelerator into another. Ideally the kicker field must rise/fall in between the beam bunches. However, in reality, an intentional beam-free time region (aka "notch") is created on the beam pulse to assure that the beam can be extracted with minimal losses. In the case of the Fermilab Booster, the notch is created in the ring near injection energy by the use of fast kickers which deposit the beam in a shielded collimation region within the accelerator tunnel.…
We have measured the total energy spread (99 persent energy spread) of the Booster beam at its injection energy of 400 MeV by three different methods - 1) creating a notch of about 40 nsec wide in the beam immediately after multiple turn injection and measuring the slippage time required for high and low momentum particles for a grazing touch in line-charge distribution, 2) injecting partial turn beam and letting it to debunch, and 3) comparing the beam profile monitor data with predictions from MAD simulations for the 400 MeV injection beam line. The measurements are repeated under varieties …
The Fermilab booster has an intensity upgrade plan called the Proton Improvement plan (PIP). The flux throughput goal is 2E17 protons/hour which is almost double the current operation at 1.1E17 protons/hour. The beam loss in the machine is going to be an issue. The booster accelerates beam from 400 MeV to 8GeV and extracts to The Main Injector (MI). Cogging is the process that synchronizes the extraction kicker gap to the MI by changing radial position of the beam during the cycle. The gap creation occurs at about 700MeV which is 6msec into the cycle. The variation of the revolution frequency …
Currently a fast vertical 1.08-m long kicker (notcher) located in the Fermilab Booster Long-5 straight section is used to remove 3 out of 84 circulating bunches after injection to generate an abort gap. With magnetic field of 72.5 Gauss it removes only 87% of the 3-bunch intensity at 400 MeV, with 75% loss on pole tips of the focusing Booster magnets, 11% on the Long-6 collimators, and 1% in the rest of the ring. We propose to improve the notching efficiency and reduce beam loss in the Booster by using two horizontal kickers in the Long-12 section. The STRUCT calculations show that using such …
Fermilab is committed to upgrading its accelerator complex towards the intensity frontier to pursue HEP research in the neutrino sector and beyond. The upgrade has two steps: 1) the Proton Improvement Plan (PIP), which is underway, has its primary goal to start providing 700 kW beam power on NOvA target by the end of 2017 and 2) the foreseen PIP-II will replace the existing LINAC, a 400 MeV injector to the Booster, by an 800 MeV superconducting LINAC by the middle of next decade, with output beam intensity from the Booster increased significantly and the beam power on the NOvA target increased…
Currently, Fermilab Booster accelerates ~4.5E12 protons per pulse (ppp) in 81 bunches from 400 MeV to 8 GeV at 15 Hz to provide beam to multiple HEP experiments and is being upgraded to handle higher beam intensity >6.7E12 ppp at a repetition rate of 20 Hz. In the current mode of operation, we control longitudinal coupled bunch instabilities (LCBI) using combination of passive and active dampers. The issues with LCBI are expected to worsen at higher beam intensities. Here, we would like to investigate at what time in the beam cycle a particular mode is going to originate and how much it contri…
The Fermilab Booster uses multi-turn beam injection with all its cavities phased such that beam sees a net zero RF voltage even when each station is at the same maxi-mum voltage. During beam capture the RF voltage is increased slowly by using its paraphase system. At the end of the capture the feedback is turned on for beam acceleration. It is vital for present operations as well as during the PIP-II era that both the HLRF and LLRF systems provide the proper intended phase and RF voltage to preserve the longitudinal emittance from injection to extraction. In this paper, we describe the origina…
The Fermilab Proton Source machines, constituted by Pre-Injector, conventional Linac and Booster synchrotron, at Fermi National Accelerator Laboratory (Fermilab) had have a long history of successful beam operations. Built in late '60s, the Fermilab Proton Source began operations early in the '70s and since then it has successful provided protons to support the laboratory physics experiments. During the past decade, Booster performance reached unprecedented proton flux delivery of the order of 1.0-1.1E17 protons per hour, corresponding to 40 kW of beam power while maintained an allowed upper l…
We describe the offline machine learning (ML) development for an effort to precisely regulate the Gradient Magnet Power Supply (GMPS) at the Fermilab Booster accelerator complex via a Field-Programmable Gate Array (FPGA). As part of this effort, we created a digital twin of the Booster-GMPS control system by training a Long Short-Term Memory (LSTM) to capture its full dynamics. We outline the path we took to carefully validate our digital twin before deploying it as a reinforcement learning (RL) environment. Additionally, we demonstrate the use of a Deep Q-Network (DQN) policy model with the c…
A second harmonic RF cavity which uses perpendicularly biased garnet for frequency tuning is currently being constructed for use in the Fermilab Booster. The cavity will operate at twice the fundamental RF frequency, from ~76 - 106 MHz, and will be turned on only during injection, and transition or extraction. Its main purpose is to reduce beam loss as required by Fermilab's Proton Improvement Plan (PIP). After three years of optimization and study, the cavity design has been finalized and all constituent parts have been received. We discuss the design aspects of the cavity and its associated …
The 40-year-old Fermilab Proton Source machines, constituted by the Pre-Injector, Linac and the synchrotron Booster, have been the workhorse of the Fermi National Accelerator Laboratory (Fermilab). During this time, the High Energy Physics Program has demanded an increase in proton throughput, especially during the past decade with the beginning of the neutrino program at Fermilab. In order to achieve a successful program, major upgrades and changes were made in Booster. Once again, the Proton Source has been charged to double their beam throughput, while maintain the present residual activati…
We present the current and planned beam physics research program and accelerator education program at Advanced Superconducting Test Accelerator (ASTA) at Fermilab.
This white paper presents opportunities afforded by the Fermilab Booster Replacement and its various options. Its goal is to inform the design process of the Booster Replacement about the accelerator needs of the various options, allowing the design to be versatile and enable, or leave the door open to, as many options as possible. The physics themes covered by the paper include searches for dark sectors and new opportunities with muons.
The completion of the PIP-II project and its superconducting linear accelerator will provide up to 1.2 MW of beam power to the LBNF/DUNE facility for neutrino physics. It will also be able to produce high-power beams directly from the linac that can be used for lower-energy particle physics experiments as well, such as directing beam toward the Muon Campus at Fermilab for example. Any further significant upgrade of the beam power to DUNE, however, will be impeded by the limitations of the present Booster synchrotron at the facility. To increase the power to DUNE by a factor of two would requir…
An integrated program is described, starting with muon experiments in the Booster era, continuing with a 2 MW target station, a 4 GeV Neutrino Factory and a 3 TeV Muon Collider, all driven by Project X. This idea provides an integrated approach to the Intensity and Energy Frontiers at Fermilab.
Fermilab is currently constructing a superconducting electron linac that will eventually serve as the backbone of a user-driven facility for accelerator science. This contribution describes the accelerator and summarizes the enabled research thrusts. A detailed description of the facility can be found at [\url{http://apc.fnal.gov/programs2/ASTA_TEMP/index.shtml}].
In fast ramping synchrotrons like the Fermilab Booster the conventional methods of betatron tune evaluation from the turn-by-turn data may not work due to rapid changes of the tunes (sometimes in a course of a few dozens of turns) and a high level of noise. We propose a technique based on phasing of signals from a large number of BPMs which significantly increases the signal to noise ratio. Implementation of the method in the Fermilab Booster control system is described and some measurement results are presented.
After a 14 month shutdown accelerator modifications and upgrades are in place to allow us doubling of the Main Injector beam power. We will discuss the past MI high power operation and the current progress towards doubling the power.
Fermilab's Integrable Optics Test Accelerator is an electron storage ring designed for testing advanced accelerator physics concepts, including implementation of nonlinear integrable beam optics and experiments on optical stochastic cooling. The machine is currently under construction at the Advanced Superconducting Test Accelerator facility. In this report we present the goals and the current status of the project, and describe the details of machine design. In particular, we concentrate on numerical simulations setting the requirements on the design and supporting the choice of machine param…
The use of existing Fermilab facilities to provide beams for two muon experiments -- the Muon to Electron Conversion Experiment (Mu2e) and the New g-2 Experiment -- is under consideration. Plans are being pursued to perform these experiments following the completion of the Tevatron Collider Run II, utilizing the beam lines and storage rings used today for antiproton accumulation without considerable reconfiguration. Operating scenarios being investigated and anticipated accelerator improvements or reconfigurations will be presented.
The Fermilab Proton Improvement Plan II, or PIP-II, would enable the world's most intense high-energy neutrino beam and would help scientists search for rare particle physics processes. The PIP-II goal is to deliver 1.2 MW of proton beam power from the Fermilab Main Injector, over the energy range 60 - 120 GeV, at the start of operation of the LBNF/DUNE program. PIP-II provides a variety of upgrade paths to higher beam power from the Main Injector, as demanded by the neutrino science program and as recommended by the 2014 P5 report. Delivering more than 2 MW to the LBNF target in the future wi…
We report on the status of the Fermilab accelerator complex, including recent performance, upgrades in progress, and plans for the future. Beam delivery to the neutrino experiments surpassed our goals for the past year. The Proton Improvement Plan is well underway with successful 15 Hz beam operation. Beam power of 700 kW to the NOvA experiment was demonstrated and will be routine in the next year. We are also preparing the Muon Campus to commission beam to the g-2 experiment.
New accelerator magnet technology based on Nb3Sn superconductor is being developed at Fermilab since late 90's. Six short dipole models, seven short quadrupole models and numerous individual dipole and quadrupole coils have been built and tested, demonstrating magnet performance parameters and their reproducibility. The technology scale up program has built and tested several dipole and quadrupole coils up to 4-m long. The results of this work are summarized in the paper.
A superconducting RF accelerator test facility is currently under construction at Fermilab. The accelerator will consist of an electron gun, 40 MeV injector, beam acceleration section consisting of 3 TTF-type or ILC-type cryomodules, and multiple downstream beam lines for testing diagnostics and performing beam experiments. With 3 cryomodules installed this facility will initially be capable of generating an 810 MeV electron beam with ILC beam intensity. The facility can accommodate up to 6 cryomodules for a total beam energy of 1.5 GeV. This facility will be used to test SRF cryomodules under…
