1. Introduction to Data Logging
Electronics controls for mobile equipment applications have been growing at a rapid pace over the last decade, to the point where they have essentially become standard on mobile machines in industries such as agriculture, construction, mining, and transportation.
Electronics were originally introduced for the purpose of enhanced automated controls, and the benefits of monitoring quickly became obvious: improved diagnostics were now available to the operator to make better decisions on machine operation.
While live monitoring of machine parameters is beneficial, it is limited to parameters that can be evaluated in real time by operators. If too many machine parameters are shown to an operator, they are ignored, because an operator can only pay attention to a few key items while they are also trying to operate the machine. However, there are typically many more parameters that provide a picture of both the current state of the machine, and a history of how it has been used.
If this information is measured and understood, it provides various benefits to operators, service people, and equipment manufacturers and designers. This article explores potential benefits of data logging, and methods on how data can be captured and analyzed to be most effective in different mobile equipment control applications.
2. Data Logging Design Considerations
Logging data allows for the benefit of monitoring of many parameters together, and also seeing historical parameters for evaluation of trends. There are a number of different design considerations when determining how to most effectively use logging in an application. The key considerations are: the type of data to be logged, type of device used for logging, size of memory available, types of logs to be stored, full logs strategy, and method of log retrieval.
2.1 Type of Data
Typical data that is stored in logs are time and date, machine hours, and then application specific machine parameters. This can be first-order data that is directly retrieved from machine sensors (e.g. engine temperature, engine RPM, angle sensor values, speed sensor values) or it could be second-order data that is derived from the operating conditions (e.g. operating modes, state machine conditions).
Also, parameters that may not be obviously related to machine function can be valuable in analysis for data, such as location (based on GPS), or orientation information (from accelerometer, gyroscope, and/or magnetometer sensors). Tracking this information along with machine function information can often provide insights for issues that are not obvious at first.
2.2 Type of Logging Device
Data logging devices can be either stand-alone modules that have the sole function of logging, or integrated electronic control modules that perform system control functionality in addition to data logging.
The main benefit of a stand-alone unit is that it is not essential to the core functionality of the machine, so it can be added as an option if it is not required for all machines, and could be added after-market as well. This approach may also be beneficial for companies that have many varieties of different machines and do not want to develop custom logging capability in each unit’s control module. The stand-alone logging module typically receives machine data over a communication interface (e.g. CAN, J1939, ISOBUS, NMEA0183, NMEA2000) that contain machine specific data. Internally generated data for location (from GPS module), or orientation sensing can also be provided for data logging modules that have these capabilities built in, such as the JCA Thrasher module.
Incorporating data logging into an electronic control unit that interfaces with sensors and drives actuators in a system can be beneficial in some cases as well. The benefits of this approach is that the overall system cost may be reduced as a separate data logging module is not required, the control unit is often directly measuring many of the sensors, and can be the source of second-order parameters (such as operating modes). So, integrating data logging into a control unit is often the most desirable solution. This of course depends on the electronic control module having memory for data logging capability, which is the case for some controllers such as the JCA Falcon controller, but may not always be the case.
Figure 1: System Block Diagram Showing Data Logging Connection Options
2.3 Size of Memory, Types of Logs and Full Log Strategies
The size of memory available in a data logging device is a key design parameter that must be taken into consideration in designing the types of logs. There are two basic types of data logs, which are event-based logs, and periodic logs.
Event based logs are logs that are only written when particular conditions (or events) occur; an example would be a log taken only when a machine changes operating modes. Periodic logs are logs that are written at fixed time intervals independent of any particular conditions, for example all key operating parameters may be stored every 10 seconds.
The benefit of the event-based logging is that less memory is typically required, but the benefit of periodic data logging is that trends that were not originally foreseen could be determined from analysis of the logs. A hybrid of each of these types are possible, where an event triggers periodic logging to begin. This would be the case for a black box recorder that may start periodic recording when some significant event (such as a crash) was detected. JCA controllers and data logging modules support multiple data log types in the same application, so some event-based logs can be kept, as well as some periodic logs (or hybrid types).
An estimate of the number of logs over a period of time will provide the length of time that can be logged before the memory is full, so tradeoffs between the size of logs (number of parameters), frequency of logs and type, and the total memory available are balanced to ensure logging over the period of time desired.
In the event that log memory does become full, a strategy is required to determine what to do. The two typical options are to stop logging (fixed buffer) or to write over the oldest logs (circular buffer). The fixed buffer approach preserves the oldest logs in the system, but prevents new logs from being written. The circular buffer approach discards the oldest logs in favor of new logs. The method that should be used is dependent on the application and the type of data, so there is no universally correct method.
2.4 Data Log Retrieval
The method for retrieving data logs can have a major impact on how the data logs are used. The logs can be retrieved over a wired connection (such as CAN, USB, or RS-232), or a wireless connection (Bluetooth, Wi-Fi, or cell modem). Logs can also be retrieved locally to a PC/laptop for analysis, emailed for remote analysis, or have logs stored in the cloud for remote data analysis.
The basic option for data log retrieval would require a wired connection into the machine with a laptop to retrieve the logs. This method requires having a laptop locally on the machine to gather the logs and often a diagnostic tool is required for interfacing the laptop to the machine.
The next option up from this would be a wireless connection. This may still require a laptop for data log analysis but the logs can more easily be retrieved. The next step up would be a wireless retrieval of logs using a tablet or smartphone app that packages the logs into a compressed format that can be emailed for analysis on a remote machine. The benefit of this method is that the person who is retrieving the logs does not have to be in the same location as the person analyzing the logs.
A further step up is loading of logs periodically to a cloud-storage database, allow users to access logs remotely in the cloud. The benefit of this is that logs can be periodically uploaded automatically and regularly when an Internet connection is made (either with a cell modem, or through a tethered Wi-Fi/Bluetooth device), and multiple users can have access to the same logs at any time.
Cloud-storage also facilitates building data analytics into the cloud system making it easier to find machine trends using data from multiple machines.
3. Data Log Analysis
Logging of the data is not particularly useful unless the data is analyzed. Analysis of the data can vary from simple review to find unusual operating conditions, to regression analysis to find statistically significant correlations of parameters to particular conditions.
The format of the data can make a difference on the ease of performing analysis. The most common format of logs is in a table format that can be input into a spreadsheet for further analysis using common spreadsheet functions. Logs with GPS locations stored can also be retrieved using location data as a key to plot on a map (using a mapping format such as KML to view on Google maps). Data stored in the cloud can be retrieved and viewed with custom graphs using a web browser or mobile device app, and can have trends from multiple machines show up immediately for quick analysis.
The benefit of cloud storage is that the analysis of the data logs can be done automatically in the cloud and presented in a clear graphical format for users. Many cloud-based data analytic tools are available for performing advanced data analytics rapidly for continuous evaluation of patterns that can point to systemic issues in a design.
This analysis can provide information on what conditions cause machine damage or higher than normal wear, insight on typical operation of machines by users, evidence of machine misuse for warranty considerations, and information useful to designs for continuous improvement of their machine.
Mobile machines are now at the stage where there is wide variety of data available for monitoring and analysis. Data logging can provide a competitive advantage for OEMs by effectively using the available data to reduce warranty claims by identifying machine misuse, improve speed of service by viewing logs to pinpoint issues, and improve designs through better understanding of how their machines are used.
The design and incorporation of robust, reliable, and easy to use data logging system will soon no longer be an option, but will be a key distinguishing factor between OEMs that lead their market, and those falling behind.