EP 2 834 517 B1 relates to a twisted blade root.
Brief outline of the case
The OD held that the claim 1 as granted lacked N over D1=WO 2012/007058, not mentioned on the ISR established by the EPO, but decided maintenance according to AR3.
Both proprietor and opponent appealed.
The board disagreed with the lack of N decided by the OD, but considered claim 1 as granted to lack IS over D1. For the surplus it confirmed maintenance according to AR3.
The case is interesting as it deals with the rounding of numeric values.
At stake is the feature F of “a 3° twist difference in the longitudinal range smaller than 25%L”
The value of twist angle variation defined in claim 1 is representative of dimensions of a wind turbine blade that set clear cut geometrical and structural limitations to that blade in terms of twist distribution that should comprise the lower threshold value of 3°.
A critical question is whether the difference in twist in the first quarter of the blade disclosed in D1 comprises a value of 3°.
The opponent’s point of view
The opponent considered that the same level of accuracy should be applied to the values disclosed in the prior art as in the claim and supported its arguments on case law, by quoting T 0175/97, T 1186/05, T 0871/08 and T 2203/14.
The opponent further argued that to compare the disclosure of document D1 with the claimed subject-matter the 2.5° of twist difference disclosed in D1 has to be rounded to the value of 3° to have the same accuracy as the one claimed.
The OD’s decision
The maximum value of 2.5° of twist variation in the first quarter of the blade which is explicitly disclosed and expressed in D1, is clearly below the threshold value of 3°.
The OD used the rounding convention expressed in the Guidelines G-VI, 7.1, to consider that the requirement expressed in feature F includes an error margin from 2.5° to 3.4°, the disclosure of D1 thus anticipating feature F.
The board’s decision
N over D1 is given
For the board, at best, according to D1, the maximum twist difference lies below 2.5°.
The Board disagreed with the finding of the OD because the present case simply requires a comparison between clearly expressed and disclosed values.
Thus to compare the lower threshold of twist variation or amplitude of 3°, which is an integer value without decimal within the claimed scope no measurement, calculation, conversion of values or any other estimation or approximation is necessary and thus no rounding is required.
The board was however unconvinced that the case law referred to by the opponent is applicable in the present case, especially because they concern decimal value of parameters in the field of chemistry requiring measurements on chemical compounds or layer composition.
T0175/97 concerns accuracy in weight percentage concentration of a chemical compound within an electrolyte including decimals, thus not applicable to structural geometrical dimensions of an object as in the present case.
In T 1186/05 rounding is required because it concerns a third decimal of a density value in the prior art to be compared to a value with only two decimals.
T0871/08 is again in the field of chemistry and requires the conversion of units.
Finally, T 2203/14 the measurement of the thickness of a microscopic layer with a certain accuracy is also involved that is to be compared with a value that requires an approximation.
Thus none of these cases can be compared to the geometry on a macroscopic solid object, which defined dimension values are remote from manufacturing tolerances and are explicitly disclosed.
The board was thus unable to follow the opponent’s further argument.
The effect of this rounding would increase the given value of 2.5 by 20%, well beyond any measurement accuracy limits, and would therefore in the board’s view lead to a misinterpretation and an extension of the disclosure of D1 beyond what a skilled person in the art, familiar with normal workshop practice tolerances, would reasonably understand, and consequently beyond what can be directly and unambiguously derived from D1.
The board therefore could thus not agree with the finding of the OD that D1 discloses feature F of claim 1, and instead concluded that the subject-matter of claim 1 as granted is novel with respect to D1 by this sole feature.
Lack of IS over D1
The board however decided that the skilled person would have considered to increase the twist angle from the value of 2,5° to 3° as a straightforward measure to improve the aerodynamic behaviour in the root portion.
Working on the twist and thus improving the lift of the blade aerodynamic profile would be considered as the most obvious parameter to vary by the person skilled in aerodynamics seeking improvement to the yield of a wind turbine and is therefore not based on hindsight.
The board did not deny that the design of the blade in its root region involves other considerations than aerodynamic ones. However, in the present case the magnitude of an increment of 0.5° required to reach the threshold value of 3° is so small, that it falls within what a skilled person would routinely consider in normal workshop practice, without anticipating major interactions with other design parameters of the blade of D1, and thus lies within the range that the skilled person would naturally consider when working around D1’s design values.
The board also disagreed that such minor change in angle of twist constitutes a deviation from the design of D1, because the modification of the twist angle by 0,5° would not be seen as departing from its teaching comprised within a range of twist angles from 1,5° to 2,5°, that is half the width of the disclosed range.
Comments
When it comes to N in presence of numerical values, it is normal to take into account the measurement accuracy.
The board’s decision is interesting in that it held that, if the effect of a rounding increases the disclosed value by 20%, a rounding should not be applied.
On the other hand varying a value directly and unambiguously disclosed in the prior art by 20% could be considered lacking IS.
The present decision, as interesting it might look, should however not generalised.
Comments
Leave a comment