MADDOX learns to recognize faults based on complex machine data and links them with empirical knowledge. This knowledge is proactively presented to the user in the event of a fault. An innovative search algorithm based on machine learning is a core component of MADDOX and makes large databases, photos, and videos easy to use. Users can confirm, correct, reject, or add to these suggestions. In this way, MADDOX – comparable to a colleague – learns with every disturbance.
The particularly high learning rate due to special feature reduction and extraction based on process knowledge is outstanding. In addition, the integration of psychological expertise helps to take into account human drivers such as motivation, helpfulness, appreciation, and fear of loss.
Increase production efficiency by reducing the number and duration of unplanned shutdowns
Motivation & Appreciation
Use of experiential knowledge and continuous training according to the “training on the job” principle
Documentation and planning
Reliability in production planning and simplified documentation (e.g. automatic recording of micro-stops)
Optional extension to a data platform and combination of other technologies for new digital business models
MADDOX in Detail:
Documentation processes, software applications used, and responsibilities often differ greatly between production companies. For this reason, we attach great importance to maximum flexibility of our software right from the start. In this way, MADDOX adapts to our customers – and not the other way around.
MADDOX runs on Linux and uses Docker technology. This means that it can be operated as a standalone installation on a PC at the plant, as well as on a virtual machine on the company’s server. Thus, it is possible to regulate access rights and protection of knowledge data individually, according to customer requirements.
For the recognition of (fault) conditions, data from the machine and its process are recorded, stored, and analyzed with machine learning algorithms. In standard cases, the analysis of internal machine data takes place without external additional sensors. However, at the customer’s request, environmental parameters, external measuring devices, other data storage devices, or higher-level systems can also be integrated.
A distinction is made between low-frequency (<500 Hz) and high-frequency (>500 Hz) data acquisition. Low-frequency data acquisition is usually performed via standard interfaces such as OPC UA/DA and is the basis for status description. For high-frequency acquisition, MADDOX is integrated into the machine’s bus as a slave via a special interface card. This also allows anomalies to be detected, for example in sensor switching signals.
The interaction with the user takes place via mobile devices such as tablets. Here, we rely on platform-independent web technologies in combination with a wrapper app. This means that MADDOX can be operated on Android, iOS, and Windows devices, and we can fully orient ourselves to the customer’s requirements. The same applies to the type of device. Here, both low-cost consumer devices with splinter protection foil and industrial tablets are used. In this context, outdoor devices have proven to be an interesting combination of protection, performance, and cost in individual cases. In terms of device size, we offer 7″ to 10″ devices, and are also guided by user preferences.
MADDOX as an assistance system for machine operators is used in industrial production environments. The diverse boundary conditions require a wide range of solutions, for example, for machine interfaces. On the other hand, the software should be able to be integrated as flexibly as possible into the customer’s existing operational processes and system landscape. Our assistance system is therefore not just one product that is simply rebuilt after each change and made available to all customers for download. At Peerox, we have recognized that special solutions are needed in a system’s architecture to meet these special requirements.
MADDOX can be integrated into both existing and new plants. Therefore, we cooperate with plant manufacturers and operators.
Since the beginning of the development, we have pursued a distinctive cooperation strategy. We work on our product together with professional UX designers, engineering psychologists, and control manufacturers. We are convinced that by bundling competencies, a strong team achieves the high speed and professionalism needed to implement such an ambitious vision. Become a part of this team!
MADDOX offers numerous possibilities for implementing your own digital services and innovative platform concepts. We would be happy to present our ideas to you personally and consider together how MADDOX can be combined with your digital products.
MADDOX is a manufacturer-independent, plant-wide system. Due to the uniqueness of each production plant as well as the different operational and technical requirements, a simple plug&play solution is rarely possible. Together with you, we check how MADDOX can be used as efficiently as possible and in the right places, how current documentation processes can be integrated, and which existing systems (e.g. MES, BDE) can be used. To this end, potential users are closely involved from the very beginning. The goal is to implement a system that increases your production efficiency, but also establishes itself as a popular, central documentation tool among employees.
We want nothing less than to revolutionize your operational knowledge management of production disruption. This is not done through the simple installation of a software. Rather, we see ourselves as a total solution provider, and accompany you from initial workshops (in which we work with you to discover success scenarios and – also in our own interest – critically examine them) to the process of internal scaling within your company. To ensure efficient achievement of our goals and that we also draw in and inspire your employees, we have developed a detailed process for product introduction.
Pilot customers and projects
In pilot projects, we develop MADDOX into a marketable product together with our customers and partners during ongoing production operations. The following three examples of our ongoing pilot projects illustrate the variety of industries and technical constraints to which we are aligning MADDOX.
At Bayer AG, a tablet packaging line at the Leverkusen site is being equipped with MADDOX. The line includes the production of blister packs using the thermoforming principle and filling with tablets as well as the subsequent cartoning of the blisters, production of bundles, cartoning of bundles and palletizing. The companies involved in the project include Seidenader Maschinenbau GmbH and pester pac automation GmbH. Special challenges exist in the variety of data interfaces as well as very high GMP and IT security requirements.
At Heinrichsthaler Milchwerke GmbH, MADDOX is implemented on a line from GEA Food Solution GmbH. In this production line, which is exemplary for the food sector, blocks of cheese are cut, sliced, packaged and fed to a secondary packaging system. MADDOX is installed as an add-on solution on a stand-alone hardware server and accesses an edge device of the plant manufacturer via OPC UA.
Essel Deutschland GmbH prints and manufactures high-quality, open-sided tubes made of plastic film for subsequent filling at the customer’s premises. MADDOX is installed on a virtual machine in the company’s internal cloud and is connected to a proven existing system. Here, we are gaining experience in implementing MADDOX as a retrofit solution on controllers without modern data interfaces, such as OPC UA.