cpacsType Complex Type |
CPACS root element
Namespace: Empty
Schema: Empty
V3.5
2023-12-19
The Common Parametric Aircraft Configuration Scheme (CPACS) is an XML-based data format for describing aircraft configurations and their corresponding data.
This XML-Schema document (XSD) serves two purposes: (1) it defines the CPACS data structure used in the XML file (e.g., aircraft.xml) and (2) it provides the corresponding documentation (see picture below). An XML processor (e.g., TiXI or XML tools in Eclipse) parses the XSD and XML files and validates whether the data set defined by the user (or tool) conforms to the given structure defined by the schema.
This documentation explains the elements defined in CPACS and its corresponding data types. Data types can either be simple types (string, double, boolean, etc.) or complex types (definition of attributes and sub-elements to build a hierarchical structure). In addition, the sequence of the elements and their occurrence is documented.
To link the XML file to the XSD file, the header of the XML file should specify the path of the schema file. An example could look like this:
<cpacs xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="pathToSchemaFile/cpacs_schema.xsd">
CPACS is an open source project published by the German Aerospace Center (DLR e.V.). For further information please visit www.cpacs.de.
CPACS data is modeled in a hierarchical structure whose underlying concept follows a top-down description of a system-of-systems which decomposes a generic concept (e.g., an aircraft or rotorcraft) into a more detailed description of its components. This originates from the conceptual and preliminary design of aircraft, where the level of detail is initially low and continues to increase as the design process progresses.
For some concepts within CPACS, however, a bottom-up approach is applied where the components are first defined in detail (sometimes referred to as library) and then linked within an instantiated higher-level concept. This is advantageous when used multiple times within complex systems, such as engines, which only have to be defined once in order to be referenced several times on the aircraft. The combination of these two methodologies is known as middle-out approach and enables the goal to fully parametrize aeronautical systems.
3.1. CPACS coordinate system
Coordinate systems are a regular cause for ambiguous interpretation of data. In CPACS, the reference coordinate system is the CPACS-coordinate system. This coordinate system is used for most of the data. A single exception is made in order to keep aerodynamic data in an aerodynamic coordinate system. The following paragraphs outline the determination to known coordinate systems.
The CPACS coordinate system is the coordinate system identified by TiGL, CPACS's geometry library. It is a right-handed coordinate system. If an aircraft is defined in the CPACS coordinate system it will usually follow the directions listed in the table below.
Therefore, the CPACS coordinate system can be confused with the body-fixed coordinate system. While often the CPACS coordinate system and the body-fixed coordinate system overlap, this must not always be true. Several definitions for body-fixed coordinate systems exist (x-axis through nose and tail, x-axis perpendicular to nose plane). For non-symmetric aircraft, body-fixed coordinate systems become even more complicated. Hence, analysis tools should stick to the CPACS coordinate system. It remains to the designer to model the geometry accordingly.
The CPACS coordinate system does not rotate with flow. Hence, aerodynamic calculations do rotate their flow relative to the CPACS coordinate system. If not stated explicitly different, e.g. for target lift-coefficients, results are returned in the CPACS coordinate system, i.e. the cfx-coefficient is parallel to the CPACS x-coordinate, regardless of the way the geometry is defined.
The following table gives a "best-practice" advice on how to locate a geometry within CPACS. Different approaches are, of course, valid as well.
Axis | Direction | Description |
x | tailwards | from nose to tail |
y | spanwise | from symmetry plane to the right wingtip |
z | upwards | from landing gear to tip of vertical tailplane |
The following figures show an example of a geometry that is aligned with the CPACS coordinate system, i.e. the body-fixed coordinate system corresponds to the CPACS coordinate system.
The aerodynamic analysis is relative to the CPACS coordinate system. That is, the angle of attack is represented by the dashed orange line. Results of the aerodynamic calculation are given in the CPACS coordinate system.
The following figures give an example of a geometry that is not defined in alignment with the CPACS coordinate system. It is a valid CPACS file, but only used in this example for demonstrative purposes.
The body axes and the CPACS coordinate system do not align. That is, the origin of the geometry is not at CPACS (0,0,0) but at a point in positive x- and z-direction.
Again, the aerodynamic analysis is relative to the CPACS coordinate system. That is, the angle of attack is represented by the dashed orange line. Results of the aerodynamic calculation are given in the CPACS coordinate system.
3.2. Local coordinate systems via parentUID and transformation
Some elements in CPACS, in particular the geometric components, are described in local coordinates. The hierarchical data structure allows to define a local coordinate system either with respect to the coordinate system of the parent element or with respect to the global CPACS coordinate system. This is achieved by combining the two elements <parentUID> and <transformation>:
parentUID: An individual data hierarchy can be set up using the optional <parentUID> element. Here it is important that exactly one element does not contain the <parentUID> in order to identify the top element of this user-specific hierarchy. As soon as the parentUID (which refers to the uID of the parent element) is set, a local coordinate system of the corresponding node is instantiated.
transformation: This allows the coordinate system to be transformed via <translation>, <rotation> and <scaling>. As soon as the <parentUID> is set, this transformation refers to the local coordinate system (in the current CPACS version this only affects <translation>). An attribute refType is used to either make this explicit (refType="absLocal") or to override this and reference the global CPACS coordinate system instead (refType="absGlobal").
The following table summarizes the possible combinations of <parentUID> and <transformation> and the resulting coordinate system (local or global):
<parentUID> not set | <parentUID> set | |
<transformation> without refType | global | local |
<transformation> with refType="absLocal" | global | local |
<transformation> with refType="absGlobal" | global | global |
Note: The combination of <transformation> with refType="absLocal" and no <parentUID> is global, because the local coordinate system to which the transformation is referring to via refType equals the global coordinate system (see fuselage in the following example).
An exemplary use case further illustrates the concept of the coordinate system hierarchy. The CPACS schema shall not specify in advance that a wing is always be part of the fuselage and engines must always be part of the wing. In other cases the engine could be attached to the fuselage, which would not be possible via a predefined XML tree. The following figure shows how components of the aircraft are related to each other via the <parentUID>. The fairing is a child of the wing and is therefore automatically translated when the wing is translated. Likewise, the horizontal tailplane is a part of the vertical tailplane and is therefore affected by translation of the latter:
There are no explicit attributes describing units in CPACS. The general convention is that all values must be given in the following SI-units:
[m] | Position, Distance |
[m2] | Area |
[m3] | Volume |
[kg] | Mass |
[s] | Time |
[K] | Temperature |
or in derived units, e.g.:
[N] | Force |
[Nm] | Moment |
[W] | Power |
[J] | Energy |
The only non SI unit used throughout CPACS is the angle in degrees [°]. For the sake of an intuitive use the angles are given in degrees rather than in radian [rad].
[°] | Angle |
To provide a better overview, it is possible to split up a CPACS dataset into several files. This can be done by inserting an <externaldata> node at an arbitrary position into the dataset. This node contains a <path> node with a URI to the external file(s), followed by one or more <filename> nodes, containing each a name of a file to be included at that position. Below, an example of such external data is given:
<?xml version="1.0" encoding="utf-8"?> <cpacs xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="pathToSchemaFile/cpacs_schema.xsd"> <vehicles> <profiles> <wingAirfoils> <externaldata> <path>file:://airfoils</path> <filename>NACA0010.xml</filename> <filename>NACA2412.xml</filename> </externaldata> <airfoil uID="NACA0012"> <name>NACA 0012 Airfoil</name> <pointList>...</pointList> </airfoil> </wingAirfoils> </profiles> </vehicles> <cpacs>
Such an external file would look like:
<?xml version="1.0" encoding="utf-8"?> <airfoil uID="NACA0010"> <name>NACA 0010 Airfoil</name> <pointList>...</pointList> </airfoil>
The file would be included completely, except for its title line <?xml version="1.0" encoding="utf-8"?>. This concept can also be used recursively (external files of external files), but it is important to prevent circle connections (file "A" loading file "B" loading file "C" loading again file "A" ...).
For path URI addresses, the trailing file separator "/" may be omitted. Below, some examples for path URIs are given:
With the help of the TiXI XML Interface TiXI, a CPACS dataset that is split into multiple files can be reassembled into a single tree structure for subsequent validation against the CPACS schema. The following commands are used to link external data sets:
The merged data tree for the example above would look like:
<?xml version="1.0" encoding="utf-8"?> <cpacs xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="pathToSchemaFile/cpacs_schema.xsd"> <vehicles> <profiles> <wingAirfoils> <airfoil uID="NACA0010" externalFileName="NACA0010.xml" externalDataDirectory="file://airfoils" externalDataNodePath="/cpacs/vehicles/profiles/wingAirfoils"> <name>NACA 0010 Airfoil</name> <pointList>...</pointList> </airfoil> ... <airfoil uID="NACA0012"> <name>NACA 0012 Airfoil</name> <pointList>...</pointList> </airfoil> </wingAirfoils> </profiles> </vehicles> <cpacs>
The CPACS-dataset often uses references between nodes. Typically, these references define connections between elements which are located somewhere else in the hierarchical dataset (e.g. a <wing> is connected to a <fuselage>; a specific <engine> is connected to a <pylon>; etc.). These connections are defined by unique identifiers (uID) which are specified as attributes. Thus, there are elements which can be referenced via a uID attribute, e.g. a fuselage:
<fuselage uID="ATTAS_fuselage">...
as well as elements which refer to the former, e.g. a wing pointing to its geometrical parent:
<wing uID="e382bf5j"> <name>ATTAS main wing</name> <parentUID isLink="True">ATTAS_fuselage</parentUID> ...
In previous CPACS versions, referencing elements were identified via the isLink="True" attribute. Since this is superfluous due to the explicit definition of the element properties via the CPACS schema, this attribute no longer needs to be listed. It is nevertheless a valid optional attribute to ensure compatibility with older datasets, but might be removed in future versions.
Since uIDs are only used to link nodes within the XML file, no naming convention is required. The characters only have to conform to the conventions of the xsd:ID type standardized by the W3C. UIDs, however, must be unique! Although a common practice for naming uIDs is their position in the data hierarchy (e.g. uID="mainWingSection3"), uIDs as shown in the above example are absolutely valid as well. It is therefore recommended to use the name element to convey human-readable meanings.
CPACS is designed to serve as a central data exchange format in fully automated process chains. A key requirement is therefore that tools can automatically read and process an incoming CPACS file. A second requirement is that users can interpret the data set. To address both requirements, the following usage of the <name> and <description> elements in combination with the uID attribute is proposed:
name: A specification of the <name> element is usually mandatory for sequences of elements (e.g., if max occurrence is unbounded [1..*]). Typical examples are wings/wing, aeroPerformance/aeroMap or missions/mission. Such elements must be able to be listed by tools, especially for visualization and reporting purposes, where the <name> element serves as a concise and human-readable indicator of the actual meaning of the corresponding element in the list (e.g., which wing, which aeroMap, which mission). This is usually a single word or a small number of words.
description: The <description> element is usually optional and is used to add comprehensive and human-readable explanations. This is usually at least one explanatory sentence.
uID: As described in more detail in Section 6, the uID attribute is mainly used for internal referencing of CPACS elements. Further processing software, e.g. TiXI and TiGL, also use the uIDs to improve the robustness of the data query. Consequently, the uID attribute serves as a machine-readable indicator and does not claim to be interpretable by human users. In some practical use cases, the same string is chosen for uID and <name>. However, restrictions on the choice of characters for the uID attribute must be considered, for example that no spaces may be used and the uID must be unique.
<wing uID="main_wing"> <name>Main wing</name> <description>This is the main wing which was designed by my awesome wing sizing design tool. Your tool should not try to read and interpret what I'm writing here as typos are not recognized by XML processors.</description> </wing>
8.1. Specification of symmetric elements
Sometimes it might be useful to specify a part of the aircraft as symmetric instead of holding all the data twice in nearly identical form in the dataset (e.g. left and right wing are usually identical, except for the sign of the y-coordinate). Hence, some parts offer the option to set a symmetry attribute:
<wing symmetry="x-z-plane">
There are six possible attribute values:
One example of how to apply the symmetry attribute is shown in Sec. 3.2. Another simplified example shown below illustrates the combination of different symmetry properties of 4 wings:
Note: The corresponding transformations are not shown here.
8.2. Referencing symmetric elements
All nodes (e.g., <parentUID>) in CPACS that refer to a component holding the symmetry attribute (e.g., <wing>) might also have a symmetry attribute to specify how symmetry is propagated through the resulting element hierarchy.
The symmetry attribute of a referencing element may take three values: symm, def, full:
For example, to refer to the "other" side of a mirrored wing the following the following syntax might be used:
<enginePylons> <enginePylon uID="pylon"> <parentUID symmetry="symm">wing</parentUID>
Note: This feature is not implemented in TiGL. The upper figure is manually processed to illustrate the principle. In addition, there is an ongoing debate whether the approach is suitable for CPACS due to rapidly increasing complexity and unresolved implicit assumptions as to whether it is one or two components after mirroring. Therefore, it is advised to avoid using the symmetry attribute if possible.
For large data sets (e.g. increments of aerodynamic coefficients due to control surface deflections) it is advantageous to map them via vectors and arrays instead of using a sequence of nodes for each data value. Therefore vectors and arrays are defined as semicolon-separated lists in CPACS. Via the documentation (derived from the XSD) of the corresponding nodes it has to be checked whether it is a vector or an array.
The vector is meant as a one-dimensional-array. In such a node, the values are given in a semicolon separated list:
<angleOfAttack>0.;1.5;3.;4.5;6;7.5;9.</angleOfAttack>
As for vectors, multi-dimensional arrays provide values in a semicolon separated list. An array is always preceded by a sequence of vectors, containing the dimensions and index values. Which vectors of an array are dimensioning is specified in the respective documentation of the array.
<altitude>1000.;2000.;3000.</altitude> <!-- vector element --> <incrementMaps> <incrementMap uID="incMap_b3ac2"> <controlSurfaceUID>InnerWingFlap</controlSurfaceUID> <controlParameters>-1;-0.5;0;1</controlParameters> <!-- vector element --> <!-- array of dimension length(altitude) x length(controlParameters): --> <dcl>11.;12.;13.;14.;21.;22.;23.;24.;31.;32.;33.;34.</dcl>
Control parameter = -1 | Control parameter = -0.5 | Control parameter = 0 | Control parameter = 1 | |
Altitude = 1000m | 11. | 12. | 13. | 14. |
Altitude = 2000m | 21. | 22. | 23. | 24. |
Altitude = 3000m | 31. | 32. | 33. | 34. |
Control parameters are abstract parameters, linking a generic value (i.e., the control parameter) to a configurational state of a control device (e.g., control surface, landing gear, engine settings, ...). The basic idea is that this control parameter can be used in different CPACS nodes (e.g., aeroMaps), while the relationship between the abstract control parameter and the configurative state of a controllable component is defined in the latter. Controllable compents can have multiple control functions (referred to as control devices), e.g., extraction and rotation state as well as the braking state of a landing gear.
Control parameters are predominantly used for control surfaces, which is why they are discussed in more detail below as an example. However, the approach also applies to other components, such as landing gears. In future the engines and other components will also be controlable via control parameters.
For control surfaces, the translation from the abstract control parameter to its physical state (i.e., deflection = rotation + translation) is defined in a so-called <path>, which is componsed of a list of <step> elements.
The control parameter values for each step are arbitrary floating point values. However, it is strongly recommended to use values between -1 and +1, or between 0 and +1 (depending on the type of control surface). The smallest and the largest value implicitly define the maximum deflection limits. It is mandatory, that the value “0” is within the specified range, as this value is treated as undeflected and used to specify a “clean” aircraft configuration (e.g. used in the clean aero performance map). Furthermore, it it is mandatory for the <step> elements to be sorted in ascending order of the control parameters. It is recommended, but not mandatory to specify a <step> with a <controlParameter> of 0. Consequently, no <controlParameter> must be used twice within a single <path> definition. Deflection values between two specified steps are handled by linear interpolation.
The following example shows the usage of control parameters within a control surface deflection path definition:
<controlSurfaces> <trailingEdgeDevices> <trailingEdgeDevice uID="InnerWingFlap"> ... <path> ... <steps> <step> <controlParameter>-1</controlParameter> <hingeLineRotation>-20.</hingeLineRotation> </step> <step> <controlParameter>-0.5</controlParameter> <hingeLineRotation>-10.</hingeLineRotation> </step> <step> <controlParameter>0</controlParameter> <hingeLineRotation>0.</hingeLineRotation> </step> <step> <controlParameter>1</controlParameter> <hingeLineRotation>5.</hingeLineRotation> </step> </steps> ...
There is a possibility that more than one deflection command is applied to a control surface at the same time (e.g. coming from a <configurationDefinition> and from an explicit deflection). Furthermore, a control surface could be deflected by one (or even multiple) control distributor and control device command(s) in parallel. In all these cases, the deflection commands have to be superposed in the following way:
At some places in CPACS, an atmosphere has to be selected (e.g. for connecting an altitude with a certain pressure or density). Currently, CPACS does only support a single atmospheric model: The ICAO Standard Atmosphere (ISA) from 1993 (see ICAO Doc 7488/3 'MANUAL OF THE ICAO STANDARD ATMOSPHERE', third edition, 1993). It covers temperature, pressure, density, speed of sound, dynamic viscosity and kinematic viscosity with respect to altitude. In CPACS, <altitude> means what is called 'geopotential altitude' (H) in the ISA reference document and is given in [m]. For details, see ISA manual, section 2.3, page E-viii f. ISA covers a range from -5000 m to 80000 m.
Temperature offsets are introduced on top of the definitions in the ISA manual (which does not cover such variations). The offset model is based upon the idea that the pressure at a fixed geopotential altitude is independent from temperature offset (pressure altitude). The temperature offset changes only the density (following rho = p / Gas Constant / T) (and viscosity, of course)
Release in November 2023
Release in April 2022
Release in June 2021
Release in February 2020
Release in August 2019
Release in Jul 2018
Release in Jul 2016
Release in Nov 2015
CPACS 2.3 is the fourth public release of CPACS. Major changes include:
Release in Feb 2015
Release in May 2014
Release in May 2013
Release in Nov 2012
Release in Mar 2012
Release in Jul 2011
Release in Feb 2011
Release in Nov 2010
Release in Aug 2010
Release in May 2010
Release in Feb 2010