Data Intelligence for Better Banknote Fitness Sorting
Banknotes get creased, folded, soiled, torn. So they need to regularly undergo a simple test: Fit or unfit? In making the decision, central banks must continually balance quality against cost. The BPS® Eco-Comparator software tool makes it possible to simulate their own threshold values for fitness, and then to select the optimal settings for banknote sorting.
Replacing banknotes is expensive, which is why finding the ideal balance is essential when determining banknote fitness. Machines from G+D are able to evaluate various parameters to assist in this respect, with graffiti, soiling, stains, tears, dog-eared corners, and holes among the most important fitness criteria. Threshold values may also vary by denomination, as notes in larger denominations need to be cleaner than small ones.
Each banknote series has around 150 quality criteria, which makes fitness inspection a complex matter. When central banks are selecting the ideal threshold values for banknote condition, cost efficiency is the most important factor.
The successful search for the perfect parameters
- BPS® Eco-Comparator visualizes and analyzes the distribution of banknote property values. Sorting threshold adjustments are simulated, and their impact on the unfit rate is immediately displayed.
- It supports the comparison and analysis of sorting results from various BPS® systems, for increased consistency and reproducibility.
- BPS® Eco-Comparator allows automated or manual scaling of graphs and histograms, as well as automated or manual tuning of sorting thresholds and sensitivity analyses.
- Automated threshold setting and a macro record/replay function make it fast and efficient to work with the tool, whether directly on the BPS® or offline in the office.
Setting your own thresholds
The settings must be perfect; the machine needs to be readjusted if the shredder rate is too high for the customer, or the fitness threshold for notes is too low. However, central banks increasingly want to be able to set threshold values themselves, in order to save time and money – the BPS® Eco-Comparator tool makes this level of independence available to central banks. The software is able to determine the best sorting thresholds from a limited amount of data, can provide simulations and make automated optimizations. The software learns which notes should be classified as fit and which as unfit in advance, based on manually sorted banknotes – resulting in more plausible fitness sorting.
In order to train a machine such as the BPS® M7, one stack of fit banknotes and one stack of unfit banknotes – each consisting of at least 1,000 notes – are processed on the machine, and then the BPS® Eco-Comparator software evaluates the data obtained from this exercise. Rather than analyzing banknote by banknote, it works on the principle of assessing data from a certain volume of notes. The result is more-plausible fitness sorting, with a sharper distinction between fit and unfit, and without excessive shredding rates. The software is able to calculate the ideal combination of threshold values to enable optimum banknote fitness sorting.
Explore different scenarios
Another option is to simulate fit and unfit. To do this the central bank gathers data from at least 10,000 circulated banknotes, and the tool uses this data to calculate the unfit rate based on current sorting thresholds. The customer is then able to explore various threshold scenarios for which the BPS® Eco-Comparator simulates the new unfit rate: “How many banknotes would be destroyed if less contamination on the front were acceptable?” This enables the central bank to model how many banknotes it wants to replace.
Where adjusting sorting thresholds used to be a matter of gut instinct, the BPS® Eco-Comparator is a decisive step towards fact-based decision-making. And it’s a methodology that pays off quickly too: If the cost of producing a banknote is €0.10, and the processing volume of a BPS® is around 500,000 banknotes per day, 1% less shredding results in savings of around €100,000 per machine per year.