Around 1900, the horse was still the better choice for most users. Well into the 1910s, animal traction did much of the hauling in industrial cities, even though the automobile already existed (McShane and Tarr, 2007; Greene, 2008). Early cars were costly, mechanically unreliable and dependent on poor roads and a thin fuel supply (Flink, 1988). Judged against the requirements of the day, the horse was the sensible choice, and the metrics were not wrong.
The mistake lay in the comparison itself. The horse had been bred and refined for centuries and sat close to its physical ceiling. The car started far lower, on a curve with a great deal more room above it. What separated the two was the distance each stood from its own limit, and day-one performance said very little about that.
We think fouling control in tangential flow filtration deserves the same kind of reading.
A Mature Curve
TFF is an old and heavily worked field. Crossflow and TMP optimization, backflushing protocols and sensor-based adaptive control have each taken a share of the fouling problem, and each now returns less per unit of effort than it did. Foster (1986) describes this as the upper end of an S-curve: performance climbs slowly, then fast, then saturates, and near the plateau every further gain costs more than the last.
These methods share an assumption. Fouling is treated as a disturbance, something to be observed while the batch runs and corrected as it develops. Improvement therefore means better sensing, better models, faster reaction.
Paradigm and Architecture
Dosi (1982) uses the term technological paradigm for the accepted model of how a class of problems gets solved: which problems count, which principles apply, which tools are considered legitimate. The paradigm also fixes the trajectory along which the technology improves. Kuhn (1962) made the same observation about science. Inside a paradigm one optimizes, and changing the paradigm changes the question being asked.
AFRM® starts from a different question. Membrane regeneration is a planned event with a schedule computed beforehand from the rheological properties of the fluid, and not a reaction to measured drift. The difficulty moves from real-time estimation to design, and with it the skills required, the failure modes, and the way the result has to be validated. The reasoning against adaptive and data-driven feedback is developed in Deterministic Invariance vs. Empirical Extrapolation.
Henderson and Clark (1990) separate improvements to components from architectural innovation, in which the components stay the same and the way they are linked changes. It is hard to recognize from the inside, because established organizations encode their knowledge in the existing architecture and tend to read anything new as a variant of the old. AFRM® is an architecture of this kind. It is a module that integrates into an existing TFF line without replacing the membrane, the pump or the validated process around them, so a customer evaluates an addition to a line that stays as the reference. In the vocabulary of process intensification (Stankiewicz and Moulijn, 2000), the aim is to get more from equipment that is already installed.
What Early Data Can and Cannot Say
Abernathy and Utterback (1978) observed that first versions of a new architecture are usually poorly optimized. The early phase serves to locate the real constraints and settle a reference design; incremental improvement follows once the design has stabilized.
A first-generation dataset therefore marks a starting point and says little about the ceiling. The more informative question is whether the architecture removes a constraint that the previous paradigm could not remove. If it does, the space left for optimization is larger than before.
A first version shows that the mechanism operates. How much it is worth is settled by quantitative validation on real fluids, with process partners.
Limits of This Reading
Innovation theory explains how new technologies tend to be judged. It does not replace experiment, and nothing on this page should be read as demonstrated AFRM® performance.
Nor is AFRM® disruptive in Christensen's (1997) sense. It does not enter at the low end of a market and climb; it addresses the same customers as the incumbents and strengthens their existing technology. Architectural innovation with a sustaining effect is the closer category.
A patent application is pending (Patent). Technical details covered by confidentiality are shared only with partners under NDA.
References
- Abernathy, W.J., Utterback, J.M. (1978). Patterns of industrial innovation. Technology Review, 80(7), 40–47.
- Christensen, C.M. (1997). The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press.
- Dosi, G. (1982). Technological paradigms and technological trajectories: a suggested interpretation of the determinants and directions of technical change. Research Policy, 11(3), 147–162.
- Flink, J.J. (1988). The Automobile Age. MIT Press.
- Foster, R.N. (1986). Innovation: The Attacker's Advantage. Summit Books.
- Greene, A.N. (2008). Horses at Work: Harnessing Power in Industrial America. Harvard University Press.
- Henderson, R.M., Clark, K.B. (1990). Architectural innovation: the reconfiguration of existing product technologies and the failure of established firms. Administrative Science Quarterly, 35(1), 9–30.
- Kuhn, T.S. (1962). The Structure of Scientific Revolutions. University of Chicago Press.
- McShane, C., Tarr, J.A. (2007). The Horse in the City: Living Machines in the Nineteenth Century. Johns Hopkins University Press.
- Stankiewicz, A.I., Moulijn, J.A. (2000). Process intensification: transforming chemical engineering. Chemical Engineering Progress, 96(1), 22–34.