Another look at measurement
The sensitivity of formulated materials to thermomechanical conditions is a key factor in the diversity of properties and applications of processing industries. It is also the inevitable cause of many of the technical difficulties encountered during the stages of research, development, production and use.
On these issues, measurement has long been at the heart of industrial strategies, but sometimes in a strange way, which sometimes raises more questions than it solves problems.
Measurement as evidence
The measure seems self-evident to many. It must be said that it is generally taught as a simple tool for extracting values. At the risk of stating the obvious, it seems natural to say that “the measuring instrument measures what it measures” : the thermometer the temperature, the pressure sensor the pressure, the viscometer the viscosity. When a thermometer struggles to detect a noticeable difference in temperature, we conclude that it lacks precision or that it is malfunctioning.
For many instrumentations, the question of measurement raises other questions. When a measurement produces identical values for obviously different products, when does it cease to be imprecise to simply become inadequate ? Inadequate with regard to what purpose, moreover? These questions, far from being trivial, are the source of many industrial difficulties in the use of measurement.
The example of the viscosity cup
To approach them, I propose to rely on an old and rudimentary device mentioned in a previous article: the viscosity cup - or consistometric cup. The advantage of this device is that its operation is particularly simple. It consists of two elements:
- the cup, a kind of funnel or cup pierced at its bottom, with standardized dimensions, in which the fluid to be tested is made to flow;
- an external stopwatch to measure the time necessary for the passage of all the fluid through the device (It is often criticized for lacking precision - in particular due to the variability of the human operator).
The principle of the measurement consists in filling the cup then calculating the flow time of all the fluid through the orifice. This flow time makes it possible to compare the flow trends of different products. Under certain conditions to which we will return, a kinematic viscosity value can be associated with the measured flow times.
Ford type viscosity cup (photo credit Kubek-forda – from Wikipedia)
The viscosity cup and its variants have been, and still are, commonly used for rather liquid products, in particular in the fields of dyes, inks, paints, oils or petroleums. ASTM, the American agency for standards, states this in standard D4212 (for immersion viscosity cups): " Viscosity cups are designed for testing of Newtonian and near-Newtonian liquids » [1].
Why is the viscosity cup considered suitable for Newtonian (or near-Newtonian) fluids and not for others?
When the measurement works more
If we fill the cup with a jelly product – for example ketchup: the product does not flow – like when we turn the bottle upside down with the cap open. Filling the cup with a ball or a stone would give an identical result.
Therefore, as a sensor would saturate, the viscosity cup does not works not for measuring the flow behavior of ketchup or jelly products.
Can we conclude that the gelled product cannot flow, just like the stone? Obviously not. By shaking a little, the ketchup will flow out. We are dealing here with a characteristic of a particular non-Newtonian product: the yield point fluid, on which I will not dwell here. We will come back to this in other articles.
Normative ambiguity
If we now replace the gelled product with a non-Newtonian fluid with no flow threshold - a suspension with low additive content for example - it flows in the device, again allowing the measurement of a flow time.
The switchboard tells us what happens next. If our product is “near newtonian”, we are still within its scope of validity. But what is it to be “almost-Newtonian”? In the absence of a scientific definition, another standard guides us: Near-newtonian liquid, a liquid in which the variation of viscosity with shear rate is small and the effect on viscosity of mechanical disturbances such as stirring is negligible. » [2]. The concrete experience of mixed is therefore enough to tell us if a fluid is almost-Newtonian.
The standard, on the other hand, tells us that “ […] If the test material is non-Newtonian, for example, shear-thinning or thixotropic, another method, such as Test Methods D 2196, should be used”. However, “Under controlled conditions, comparisons of the viscosity of non-Newtonian materials may be helpful, but viscosity determination methods using controlled shear rate or shear stress are preferred”.
It's all pretty vague. Why does the standard seem to have so much difficulty in clearly defining its scope of validity?
The double reference of the standard
The viscosity cup standard indicates that the device is designed for Newtonian and near-Newtonian liquids.
Although implicit, this mention seems to be based at least partially on a scientific reference. What is scientifically established in the context of the viscosity cup [3] is thatit is possible to translate the measured flow time into a kinematic viscosity value by applying the laws of capillary flow. Gold these laws are strictly valid only in the case where one can consider an "absolute" viscosity, that is to say the cases where the viscosity does not depend on shearing: “Newtonian” fluids.
In cases where the equation between flow time and viscosity is no longer strictly valid, the viscosity cut does not become in practice unusable for comparing, even qualitatively, flow trends. It is this empirical scope which is confusedly reflected behind the absolutely non-scientific notion of “almost-Newtonian”.
For non-Newtonian products, the notions of absolute, kinematic or dynamic viscosity are no longer intrinsic constants fluid, but functions shear conditions imposed on the product (we then speak of apparent viscosity). It is as if the viscosity cup suddenly changed its operating principle: from a neutral device in which the fluid flows without any particular disturbance, it becomes a device which, by its geometry, imposes shear stresses. In the standard dimensions of the viscosity cups and for products of usual density, the shear experienced by the product is maximum at the level of the flow orifice and varies over the course of the flow (between 0 and ~400s-1) [4].
The viscosity cut therefore quantifies a flow time under undefined and changing shears. No simple relationship can therefore be derived between flow time and apparent viscosity., although in practice variations in non-Newtonian fluid flow times are likely to reflect differences in actual behavior.
Thus, the viscosity cup device is scientifically valid for measuring kinematic viscosity values for Newtonian fluids but only empirically acceptable - under conditions which must be defined on a case-by-case basis - for non-Newtonian fluids. Theoretical and empirical validity therefore do not coincide and the confusion of the standard translates the ambiguity of such an implicit double reference.
It therefore seems useful to go beyond the normative vagueness relating to the viscosity cut by posing the existence of two regimes of validity of the measurement, which I will define in a framework going beyond that of the viscosity cut.
Regime " Measure-Sensor »
I propose to define the first regime as the area within which a device measures a quantity that has otherwise been scientifically established as being traceable to a property - via models most often invisible to the operator. For common sense, it is the mode of operation of any measure, of which the sensor (temperature, density, pressure, conductivity, pH, etc.) is the best illustration. More complex fully automated techniques, such as particle size measurement, also fall under this regime.
The use of the viscosity cup also falls under this regime, but only for the measurement of Newtonian fluids. In this case, the flow time is translated via tables resulting from the application of flow models - strictly valid for Newtonian fluids only - into kinematic viscosity.
Thus, certain devices operate by construction according to this regime, others only in certain configurations of use. The limits of validity of this mode of operation are fixed by, on the one hand the limits of validity of the underlying models, on the other hand the material limits of the device (sensitivity, saturation, etc.).
The validity of the measurement is established by the robustness of the relationship between the measured parameter and a scientific property. The realities to which property relates are most often intuitive (the thermometer measures the temperature). When they are not, they are simply not addressed, or in very vague terms. Thus, in the case of the viscosity cup, the relationship between the viscosity measurement and the concrete application is rather vague: “Viscosity data are useful in the determination of the ease of stirring, pumping, dip coating, or other flow-related properties of paints and related fluids”.
“Measurement-Experience” scheme
In the second regime, the challenge is rather toestablish the link between the measured parameter and the real. That the measured parameter relates to a scientific property is not a necessary condition. The validity of the measurement in this regime is exclusively determined by the context of the experiment.
Many application tests, called "empirical", proceed from this principle (typed density, texturometry, etc.). The use of the viscosity cut for non-Newtonian fluids falls within this framework. It does not matter that the flow times cannot be formally translated into viscosity values, the device can be used with good reason as long as it makes it possible to anticipate concrete behaviors.
Far from being solely empirical, this regime is also characteristic of experimental science in the phase of building, testing and adjusting its models. The development contexts of the different generations of viscometer and rheometer that I presented in the previous article are an illustration of this.
Finally, this regime is also a regime for the use of configurable instrumental techniques, to which we will return at length.
The crux of the matter
By introducing the concepts of measurement validity regime, the aim was to clarify the ambiguity of scientific and empirical references. Although at the draft stage, these concepts seem to me to present the interest of highlight this double reference of the measured parameter to the property and/or to the phenomenon.
That a system as rudimentary as the viscosity cut poses difficulties for normative formalism may surprise at first sight, but I believe illustrates quite well the consequences of the sensitivity of certain products to the conditions of their implementation on the sequence of relations " Parameter measured – Property – Phenomenon”.
Indeed, the relation of the property to the phenomenon, which presents itself as intuitive in many cases (as for the viscosity of Newtonian fluids), becomes impossible to apprehend in the same way for many states of matter extremely common in the industry (notably non-Newtonian fluids, but powders are also involved).
The “Measurement-Sensor” regime, which underlies most of the normative measures, leaves the relationship between property and phenomenon out of consideration, as if resolved in advance. The reality on the ground of the observed difficulty of normative methods in allowing effective discrimination of products, interpretation or anticipation of their real behavior could constitute clues, if not proof of the negative effects of such a presupposition.
I will have the opportunity to show later that this presupposition -deeply rooted in the history of the development of industrial methods-, which is also found in all quality approaches, continuous improvement methods or the speeches of the Industry 4.0 is the crux of the problem in a large majority of the technical issues of the processing industries.
References :
[1] Standard Test Method for Viscosity by Dip-Type Viscosity Cups – ASTM D 4212 – 99 (reapproved 2005)
[2] Standard Test Method for Viscosity by Ford Viscosity Cup – ASTM D1200 – 94 (reapproved 2005)
[3] Standardization of the Saybolt universal viscometer – Bureau of standards – 1918
[4] Mezger, Thomas G. The Rheology Handbook: For Users of Rotational and Oscillatory Rheometers. Hanover: Vincentz Network, 2006
[5] Standard Test Methods for Rheological Properties of Non-Newtonian Materials by Rotational (Brookfield type) Viscometer – ASTM D2196-05
Last Updated on September 15, 2022 by Vincent Billot