Data gathering processes can have many names based on how data is collected and used. Sampling, logging and monitoring all involve data collection, but there are very important differences between the three. This article provides an overview of each concept and when it is applied.
In addition to having differences with each other, data gathering processes can also change based on the application. In weather monitoring, for example, the wind can change much faster than temperature. As a result, wind monitoring tends to be a more demanding application than temperature monitoring: For example, ambient temperature will not jump from 15°C to 25°C within a minute, however the wind can go from 30 mph to 40 mph within just a few seconds.
This is why a fast sampling rate is so important for wind monitoring, we’ll discuss below why.
Sampling consist on gathering a certain number of observations, which is representative of a larger data set. With modern information systems, data can be generated much faster than anytime before in human history, and analysing every single observation is unfeasible. There are many types of samples, based on how the data points are collected:
- A simple random sample is taken from the complete data set, where each observation has the same probability of being chosen. For example, if you choose maximum wind speed measurements for 40 sites out of 200, without any previous classification, the result is a simple random sample.
- A stratified sample is taken by breaking down the data set based on attributes, and then taking a simple random sample is taken from each subgroup. Following the example above, assume you have maximum wind speed data for 150 urban sites and 50 rural sites. If you collect 30 random values from urban data and 10 values from rural data, you have a stratified sample.
- A systematic sample [used for wind] is based on gathering data at predetermined intervals, with a random starting point. In general, weather monitoring equipment collects a systematic sample, since there is a specified time interval between consecutive measurements.
In wind measurement systems, sampling can be used to describe wind patterns with reasonable accuracy. For example, wind measurement systems that meet the IEC 61400 standard have a minimum sampling rate of 0.5 Hz, equivalent to one measurement every 2 seconds, but the recommendation is 1 Hz – one measurement per second. WINDCRANE is our wind monitoring system, and exceeds the standard requirements, with a sampling rate of 1 Hz. Why? Next time is windy observer the changes in wind speed, the wind gusts and you will see that these happen even faster than 1 second.
A fast wind sampling rate is really important in order to fully understand and interpret wind speed, in particular when the safety of machines in operation is involved.
The concept of data logging is very simple: measurements are collected and stored sequentially, regardless of data attributes.
Logging does not involve analysis, since the purpose is to generate a time record of information. However, it is important to collect data at a fast enough rate to reflect important characteristics of the behavior or process being analyzed.
- Assume you use a generic data logger in weather monitoring, and it only picks up information at 1 minute interval.
- The wind can change much quicker than this, and the data gathered will be incomplete as a result.
- On the other hand, if you use an IEC 61400 compliant device like WINDCRANE, wind pattern with a short duration are recorded.
- Gusts and turbulence are of particular interest in wind monitoring, due to the risks they bring. These wind behaviours can change in an instant.
The other factor that limits the logging capacity of a weather monitoring system is the storage space available. For stand-alone devices, once the memory is full, the only way to register new data is by overwriting older data. However, cloud-based devices can upload data to a remote server through GSM connectivity, which allows data to be logged indefinitely – WINDCRANE uploads its data to a cloud server at 10-minute intervals with the statistical data of wind gust/max, average wind speed and standard deviation of the wind (important for looking at turbulence intensity)
In addition to expanding the logging capacity, a specialised wind cloud database provides a key advantage in applications where a logging system is exposed to harsh conditions: If the unit is lost or damaged, the data is preserved as its data is backed up in the cloud in real time.
Monitoring consists on providing live data from a measurement system, normally with the purpose of having 24/7 visibility of site conditions. For example, construction companies can monitor their project sites to suspend activities when weather conditions are unsuitable for work.
Since monitoring often involves long distances, a fast communication system is vital. Knowing yesterday’s wind speed is useful for weather analysis purposes, but not in sensitive applications like construction, where a sudden strong wind can lead to an accident. In construction and other heavy industries, you need to know what is happening NOW across all sites, and this applies for the weather as well. For example, if you are working with tower cranes, a system that tells you the wind was dangerous one hour ago is useless.
To be viable, a monitoring system must present data in a form that is useful for quick decisions. Raw data provides little insight, but it can be aggregated and color-coded to provide a clear snapshot of weather conditions across multiple sites. It is also important to consider that different devices may be used to access the weather monitoring system – a format that is suitable for a computer screen may not be practical in a smartphone, and vice-versa.
Logic Energy, the company behind WINDCRANE, bring a successful track record of more than one decade in heavy industries like construction and energy, and also work with users from non-technical fields in residential and commercial applications. As a result, Logic Energy has accumulated real expertise in data processing, and know how to extract valuable information from large volumes of raw data.
After all, if you do not have wind data backup for your operations, you may as well lose a blank cheque in a windy day.