26 March 2023
Flight test has always been fueled by data analytics. Flight test takes a system into the unknown to explore and gather data. That data needs to be confirmed for quality, visualized, analyzed for patterns, and interpreted to inform decisions. So, while data analytics is nothing new in flight test, the last 10 to 15 years have seen major advancements in algorithms, programming frameworks, storage, parallel processing, and the use of graphics processing units (GPUs), resulting in what has been called “Big Data Analytics” in the financial and social media industries.
Changes in test instrumentation have pushed flight test teams closer and closer to the realm of “Big Data.” Flight data from a fully instrumented Airbus 380 or F-35 approaches the scale of banking or financial systems, whether "streaming" (real-time data monitored and analyzed on-the-fly either on-board or via telemetry) and "batch" (recorded, then made available later).
There are several different types of data analytics, including:
1. Exploratory: Viewing the data from multiple perspectives provides insight into data quality, outliers, anomalies, unexpected patterns, and suggestions for follow-on analysis.
2. Descriptive: Summarizing and describing data in order to understand what has happened within the dataset.
3. Diagnostic: Investigating the root cause of a problem or issue in order to understand why it occurred.
4. Predictive: Creating statistical models and machine learning algorithms to predict future outcomes or trends based on available data.
5. Prescriptive: Using data and analytics to recommend specific actions or decisions.
Data analytics allows professional flight testers to make data-informed decisions, rather than only relying on heuristics or prescriptive processes.
Python is the preferred tool for data analytics, with a number of libraries and frameworks that are specifically designed for data analysis. These libraries, including NumPy, SciPy, and pandas, provide a range of tools for tasks such as importing and cleaning data, manipulating data sets, and building and evaluating statistical models. Python is widely used in the field of machine learning, which is a subset of data analytics that involves using algorithms to automatically learn patterns in data and make predictions. There are many machine learning libraries available in Python, such as TensorFlow and scikit-learn, which make it a powerful choice for this type of work.
Python has several advantages over MATLAB for data analytics and visualization:
1. Cost: Python is an open-source language, which means it is freely available to use and modify. In contrast, MATLAB is a proprietary language and requires a license to use, which can be expensive.
2. Community and resources: Python has a large and active community of users and developers, which means there is a wealth of online resources, documentation, and support available. MATLAB also has a large user base, but it is not as widespread as Python.
3. Extensibility: Python is a general-purpose programming language, which means it can be used for a wide range of tasks beyond data analytics and visualization. This makes it a more versatile tool than MATLAB, which is primarily focused on technical computing.
4. Libraries and frameworks: Python has a large number of libraries and frameworks that are specifically designed for data analysis and visualization, such as NumPy, pandas, and Matplotlib. These libraries provide a range of tools for tasks such as importing and cleaning data, manipulating data sets, and creating charts and plots. MATLAB also has a number of built-in functions and toolboxes for data analysis and visualization, but the range of options is not as extensive as in Python.
Large datasets demand a versatile tool, and Python can be that tool for your test team! Make it a valuable addition to your flight test toolkit.
Determination of specific excess power (Ps) via the level acceleration flight test technique. Includes flight test demonstration, explanation of governing equations, and Python coding tutorial to create calculations and plots.