Database - Overview

Components

A component is an instance of a component type. Component types describe the meta information of the components. An example of a component type is PT100. The type has a name and a description. Furthermore, the component type can have parameters that are measured by the component. For example, the sensor PT100 records the temperature. This approach makes sure that every instance of the sensor always has the same name and parameters. Components (instances of a type) can have a serial number and a status. With the status, it can be tracked when the component was temporarily defective or being repaired. To classify the components, every type has a component class. In the example above, the PT100 has the class Sensor.

Assembled components

To create an assembled component, you have to relate several components. Therefore you have to create the components first and set them in relation afterwards. The table component_relation stores all relationships between the components and describes all installations. An entry of the table describes that the component is installed in the parent component. Both components are a part of the main component. The attributes installation timestamp and removal timestamp store the period of the installation. If the field removal timestamp is null, it means that the component is currently installed.

Calibrations

During a deployment, the efficiency of the instruments decreases, which also influences the data quality. Thus a pre-calibration and a post-calibration will be performed to analyze this effect. These calibrations will be used to create calibration functions that will be applied to improve the data quality. These calibration functions have to be stored in the database to retrace the process chain. A calibration function has a function type and several calibration parameters that are required for calculations. Moreover, a function could have additional function parameters if needed.

Deployments

Projects and campaigns

A project has a name, a long name, a period, and a description. There are third-party funding projects and research infrastructures. Third-party funding projects run over 3-4 years, and research infrastructures run for over 20 years. During a deployment several deployments will be performed. Every deployment must be assigned to a project and is limited in time. Campaigns are a part of the project and are limited in time as well. They also have a name and a short description. Each campaign has to be assigned to a project and describe a subset of the deployments. These deployments always need to be attached to the project and never directly to the related campaign. Only the campaigns refer to the associated deployments and not vice versa.

Deployments

Components can be installed into an aircraft or at a fixed measurement location. For each component that is installed, an entry will be created in the database. Installed components are usually assembled. For example, the unit NOD will be installed into the component Package 2b and the Package 2b into an aircraft. In that case, there is only an entry for the Package 2b. This approach reduces the memory space and management. During the deployment, several series are measured. These series always refer to the deployment.

Aircraft

Every airline contains several aircraft and every aircraft has to be assigned to an airline. Furthermore, every aircraft needs to be assigned to an aircraft type. An aircraft can perform several flights. Every flight has a departure airport and arrival airport. Every airport needs to be assigned to a country. Note that every aircraft has an availability period that saves when the aircraft was available for operations.

Measurement locations

As already mentioned, a component can also be installed at a fixed measurement location (e.g., a station). These locations have a fixed position and an availability period. The period saves in which period the location was available for measurements. An example of a measurement location is the SAPHIR chamber.

Dataseries

Data series

Every deployment has a number of data series that were recorded by the components. Series can only relate to components that measure data. Every series has a number of data points and a process level (e.g. Raw data). For each timestamp, only one data point exists. A data point can have additional information. For example, if the data point is the sum of values, an extension can be created, which refers to the data point and contains the values as a JSON-text. Moreover, every data point has a flag that marks the validity of the data point.