BNL, Upton, Long Island, New York, USA
The repository of data for the Relativistic Heavy Ion Collider and associated pre-injector accelerators consists of well over half a petabyte of uncompressed data. By todays standard, this is not a large amount of data. However, a large fraction of that data has never been analyzed and likely contains useful information. We will describe in this paper our efforts to use machine learning techniques to pull out new information from existing data. Our focus has been to look at simple problems, such as associating basic statistics on certain data sets and doing predictive analysis on single array data. The tools we have tested include unsupervised learning using Tensorflow, multimode neural networks, hierarchical temporal memory techniques using NuPic, as well as deep learning techniques using Theano and Keras.
BNL, Upton, Long Island, New York, USA
The development of HTTP service interfaces* to the BNL Collider-Accelerator Department (C-AD) controls system opens up the ability to more quickly and easily adapt existing codes developed for other systems for use at RHIC. A simple particle accelerator online model built for commissioning the NSLS II** was adapted for use with the Low Energy RHIC electron Cooling project (LEReC)*** and the Coherent Electron Cooling (CeC)**** proof of principle experiment. For this project, a set of python modules and a python application were adapted for use in RHIC by replacing NSLS II control system interfaces with python modules that interface to the C-AD controls HTTP services. This paper will discuss the new interfaces and the status of commissioning them for operations.
* T. D'Ottavio, et al., these proceedings
** S. Seletskiy et al., TUPMA054, IPAC15, 2015.
*** A. Fedotov et al., WEA4CO05, NAPAC16, 2016.
**** V.N. Litvinenko et al., THPS009, IPAC11, 2011
BNL, Upton, Long Island, New York, USA
In keeping with a long tradition in the BNL Collider-Accelerator Department (C-AD) controls environment, we try to provide general and simple to use interfaces to the users of the controls. In the past we have built command line tools, Java tools, and C++ tools that allow users to easily access live and historical controls data. With more demand for access through other interfaces, we recently built a set of python and MATLAB modules to simplify access to control system data. This is possible, and made relatively easy, with the development of HTTP service interfaces to the controls*. While this paper focuses on the python and MATLAB tools built on top of the HTTP services, this work demonstrates clearly how the HTTP service paradigm frees the developer from having to work from any particular operating system or develop using any particular development tool.
* T. D'Ottavio, et al., these proceedings
Stony Brook University, Stony Brook, New York, USA
BNL, Upton, Long Island, New York, USA
Stony Brook University, Computer Science Department, Stony Brook, New York, USA
Large distributed control systems typically can be modeled by a hierarchical structure with two physical layers: Console Level Computers (CLCs) and Front End Computers (FECs). The controls system of the Relativistic Heavy Ion Collider (RHIC) consists of more than 500 FECs, each acting as a server providing services to a potentially unlimited number of clients. This can lead to a bottleneck in the system. Heavy traffic can slow down or even crash a system, making it momentarily unresponsive. One mechanism to circumvent this is to transfer the heavy communications traffic to more robust higher performance servers, keeping the load on the FEC low. In this work, we study this client-server problem from a different perspective. We introduce a novel game theory model for the problem, and formulate it into an integer programming problem. We point out its difficulty and propose a heuristic algorithms to solve it. Simulation results show that our proposed schemes efficiently manage the client-server activities, and result in a high server throughput and a low crash probability.
Stony Brook University, Stony Brook, New York, USA
BNL, Upton, Long Island, New York, USA
The Relativistic Heavy Ion Collider (RHIC) complex of accelerators at Brookhaven National Laboratory (BNL) operates using a large distributed controls system, consisting of approximately 1.5 million control points, over 430 VME based control modules, and thousands of server processes. We have developed a new testing platform that can be used to improve code reliability and help streamline the code development process by adding more automated testing. The testing platform simulates the control system using the actual controls system code base but by redirecting the I/O to simulated interfaces. In this report, we will describe the design of the system and the current status of its development.
BNL, Upton, Long Island, New York, USA
This paper describes the development and use of an HTTP services architecture for building controls applications within the BNL Collider-Accelerator department. Instead of binding application services (access to live, database, and archived data, etc.) into monolithic applications using libraries written in C++ or Java, this new method moves those services onto networked processes that communicate with the core applications using the HTTP protocol and a RESTful interface. This allows applications to be built for a variety of different environments, including web browsers and mobile devices, without the need to rewrite existing library code that has been built and tested over many years. Making these HTTP services available via a reverse proxy server (NGINX) adds additional flexibility and security. This paper presents implementation details, pros and cons to this approach, and expected future directions.
BNL, Upton, Long Island, New York, USA
Schottky monitors are used to determine important beam parameters in a non-destructive way. In this paper we present improved processing of the transverse and longitudinal Schottky signals from a hi-Q resonant 2.07 GHz cavity and transverse signals from a low-Q 245 MHz cavity with the main focus on providing the real-time measurement of beam tune, chromaticity and emittance during injection and ramp when the beam condition is changing rapidly. The analysis and control is done in python using recently developed interfaces to Accelerator Device Objects.
FRXPL01
BNL, Upton, Long Island, New York, USA