Read data in the 'MTConnect'< http://www.mtconnect.org/> standard. You can use the package to read data from historical 'MTConnect logs' along with the 'devices.xml' describing the device. The data is organised into a 'MTConnectDevice' S4 data structure and some convenience methods are also provided for basic read/view operations. The package also includes some functions for analysis of 'MTConnect' data. This includes functions to simulate data (primarily position data, feed rate and velocities) based on the G code and visualisation functions to compare the actual and simulated data.
This research was supported by the National Institute of Standards and Technology through the cooperative research agreement 70NANB15H090, “Standards Based Infrastructure for Enabling Smart Manufacturing Research and Development Test Bed.”
Just follow the regular installation procedure.
# Install devtools if not installed alreadyinstall.packages("mtconnectR", repos ="")
The MTConnectR package provides a way to read data from MTConnect and G Code to the statistical software R (www.r-project.org). You can use the package to read data from historical MTConnect logs along with the devices.xml describing the device. The data is organised into a 'MTConnectDevice' S4 data structure and some convenience methods are also provided for basic read/view operations. Also, the package further provides the user a way to read and parse G Code data into R, for simulating data in the MTConnect format, and to map the simulated data with the actual MTConnect data.
You can use a ruby parser on the MTConnect agent data dump to get data data from a CNC machine as a log file. This file is a continuous stream of data with data for different aspects of the machine (execution, path feed rate, controller mode, conditions, etc) combined into a coninuous object. One issue faced by the those who want to analyse machine data is the fact that raw MTConnect data, although ideal for stream data processing and canonical storage, doesn't lend itself to aggregation and quick analysis.
In addition, we don't have any context towards what data corresponds to which data item in the Device. This data is stored in another file called the Devices XML. However the metadata from the Devices XML and the raw data are disjoint unless passed through an MTConnect agent.
We try to solve this issue with this package. It can read in a data from the device and parse the Devices XML and store the data in a form more appropriate for analysis with data seperated into different entities called data items.
mtconnectR package also provides functions with further analysis capabilities.
The user can also read and parse G Code data into R using this pacjage. The user can also use this parsed G Code data for simulating data in the MTConnect format. In addition, the user has functions which can be used to map the simulated data with the actal MTConnect data.
We've added a few S4 classes to aid in the correct structuring of MTC data. They are:
This is a virtual class that can be used to represent any type of time series object. The only necessary slots are:
Currently, this is inherited only by the MTCDevice Class
This is the class in which all the main details are stored. The slots(in addition to the ones for the MTCCycle Class ) are:
This class represents the data from a single data item from MTConnect. The slots for these class can be accessed
directly by using the
@ accessor function to access the slots in the class but it is adviced that the methods over the MTCDevice Class be used directly. The slots for MTCDataItem class are
create_mtc_device is a wrapper over multiple functions, and can accept different types of data input including:
The output of the
create_mtc_device function is a MTCDevice S4 Class that has all the data organized into easily accessible forms. In additon, the following convenience functions are also provided to facilitate exploratory analysis:
Within the MTCDevice Class, each data
For more info on individual functions, read the vignette on
simulate_data_from_gcode - This function is used to simulate MTConnect-style data
from gcode data. The inputs are parsed gcode values from the parse funciton. Optionally,
start time can be provided to make the first data point start from that point (or numeric value),
and a data resolution can be provided if you need to interpolate data simulating the
stateful representation of sample values.
parse_gcode function can be used to read G Code data into R. The G Codes currently
supported can be got fromt he
gcode_dict.csv file as follows:
gcode_dict = utils::read.csv(system.file("gcode_dict.csv", package = "mtconnectR")) %>% select(-notes)
Currently, 47 different G Codes are in the dictionary.
simulate_data_from_gcode function can be used to simulate MTConnect data from G Code. Data from G Code including timestamp and values in MTConnect style (stateful, tesselated interpolation for samples). Currently, the follwing data items are implemented:
Given data for one instance of the part and the G Code, the
map_gcode_mtc function maps the G code and raw data using Simulated G code values. Currently the mapping is done using
We can expand using Event (program/Tool etc) mapping, but this is not implemented yet
plot_twoway function can be used to plot the Simulated and the Actual values against each other once the mapping has been completed.
We have also bundled some scripts to do common tasks that can be done with the MTConnectR package. For more information on how to use them, read the README in the scripts folder.
For DMTCD Data deal with:
This package is licensed under AGPL-3 (http://www.r-project.org/Licenses/AGPL-3) and comes with no warranty.
COPYRIGHT HOLDER: System Insights Inc.