The Units model describes the relations between UnitType, ServerUnitType and AlternativeUnitType.

The UnitType is the base class and defines details that are relevant of all derived types.

Figure 9 - Units model

It is formally defined in Table 38.

Table 38 – UnitType definition

Attribute	Value
BrowseName	UnitType
IsAbstract	True
References	NodeClass	BrowseName	DataType	TypeDefinition	Other
Subtype of the BaseObjectType defined in OPC 10000-5
HasProperty	Variable	Symbol	LocalizedText	PropertyType	M
HasProperty	Variable	Description	LocalizedText	PropertyType	O
HasProperty	Variable	UnitSystem	String	PropertyType	M
HasProperty	Variable	Discipline	String	PropertyType	O
Conformance Units
Data Access Quantities Base

The DisplayName Attribute of each instance shall provide the name of the unit, e.g. “second”, “degree Celsius” or “square metre”. This matches the “description” field of the EUInformation structure (see 5.6.3.3).

The Symbol Property is used for the symbol of the unit (e.g. "h" for hour or "m/s" for meter per second). If no symbol is defined for the unit, the DisplayName Attribute shall be used as symbol. Symbol matches the “displayName” field of the EUInformation structure (see 5.6.3.3).

The Description Property is used to expose a more verbose explanation.

The UnitSystem Property describes the system of units (e.g. ISQ) in which the unit is specified. If any of the well known systems defined in Table 39 is used, the acronym in the column “UnitSystem” shall be used for the value of this Property.

Table 39 – Non-exhaustive list of well-known systems of units

UnitSystem	System of units
ISQ	International System of Quantities (ISO/IEC 80000) Fully covers and thus replaces SI (ISO1000)
USCS	US Customary Unit System
ISU	British imperial system of units
MSU	Myanmar Units of measurement
TROY	Troy units of weight
CGS	centimetre–gram–second system of units
GAUSS	Gaussian System of units
FPS	foot–pound–second system of units

Syntax References: Instances of the UnitType can identify the unit in a specific external reference work using HasDictionaryEntry References – see 6.3.2.

Instances of ServerUnitType define the units utilizes in the Server. They are assigned to the proper Quantities. EngineeringUnit Properties can refer to it. It is formally defined in Table 40.

Table 40 – ServerUnitType definition

Attribute	Value
BrowseName	ServerUnitType
IsAbstract	False
References	NodeClass	BrowseName	DataType	TypeDefinition	Other
Subtype of the UnitType
HasComponent	Object	AlternativeUnits		BaseObjectType	O
HasProperty	Variable	ConversionLimit	ConversionLimitEnum	PropertyType	M
HasComponent	Object	CoherentUnit		UnitType	O
Conformance Units
Data Access Quantities Base

The optional Object AlternativeUnits contains Objects of the AlternativeUnitsType. These explicitly specify the units into which the conversion can be made.

The mandatory Property ConversionLimit indicates whether the ServerUnit may be converted. A distinction is made between NO_CONVERSION, LIMITED and UNLIMITED.

UNLIMITED means the client may perform conversions based on the rules defined by the given UnitSystem.

LIMITED conversion means that a conversion can not be performed by simply applying the rules defined by the given UnitSystem. Either only the conversions mentioned in the AlternativeUnits are to be used or the client requires application specific know-how for a conversion on his own.

NO_CONVERSION means that no conversion is allowed (e.g. for statistical values).

The CoherentUnit of a value is a derived unit that, for a given system of quantities and for a chosen set of base units, is a product of powers of base units, with the proportionality factor being one. Therefoe it shall share the same UnitSystem as the ServerUnit.

The components of the ServerUnitType have additional subcomponents which are defined in Table 41.

Table 41 – ServerUnitType Additional Subcomponents

BrowsePath	References	NodeClass	BrowseName	DataType	TypeDefinition	Other
AlternativeUnits	HasComponent	Object	<AlternativeUnit>		AlternativeUnitType	MP

The AlternativeUnitType describes alternative units to a ServerUnit. It is required to specify a conversion method for a value from the ServerUnit to this AlternativeUnit. It is formally defined in Table 42.

The use of conversions enables that a server and a client can work with different units and even in different systems of units. E.g. a server works with imperial units and a client uses metric units.

As a single server may distribute values to a number of clients with different needs the actual conversion has to be performed at client side.

Table 42 – AlternativeUnitType definition

Attribute	Value
BrowseName	AlternativeUnitType
IsAbstract	False
References	NodeClass	BrowseName	DataType	TypeDefinition	Other
Subtype of the UnitType
HasProperty	Variable	LinearConversion	LinearConversionDataType	PropertyType	O
HasProperty	Variable	MathMLConversion	String	PropertyType	O
HasProperty	Variable	MathMLInverseConversion	String	PropertyType	O
Conformance Units
Data Access Alternative Units

The Server shall provide either a LinearConversion or a MathMLConversion together with a corresponding MathMLInverseConversion. It may provide Linear and MathML conversions in parallel.

The optional Property LinearConversion represents a simple conversion according to the following formula. The values(a, b, c, d) are given in a Structure as defined in the LinearConversionDataType in chapter 6.6.2. The valuex is published by the server in the named server unit andy is the value converted into the named alternative unit.

Figure 10 MathML example linear conversion

This also defines the inverse conversion to be used if a client wants to write a value to the server. The values (a, b, c, d) are given in a Structure as defined in the LinearConversionDataType in chapter 6.6.2. The valuey1 is the value that the client wants to write to the server in the named alternative unit and x1 is the value the client actually has to write to the server instead

Figure 11 MathML example inverse linear conversion

The optional Property MathMLConversion allows the specification of all kinds of conversion methods. The MathML syntax is used for this. Within the MathML expression X always stands for the value at the server side and Y for the value at the client side. An example (formula of the LinearConversion) looks as follows.