Go to Top

Problem statement and motivation

Production systems deteriorate with the usage and age. The normal strategy to keep them in good conditions is to apply preventive maintenance practices, with a supportive workforce “reactive” in the case of clearly detected malfunctions or machine breakdowns. All these have an impact on quality, cost and in general, productivity. Added to this, the uncertainty of machine reliability at any given time, also impacts on product/production delivery times. It is known also that a worn-out or incorrectly assembled mechanism has higher energy consumption.

The use of intelligent predictive technologies and tools could contribute to improve the situation; it has been proven that deployment of these tools and techniques can detect malfunctions or potential breakdowns, and help and guide an anticipated solution. However the obvious of the advantages of a predictive strategy and techniques are not widely used in the production environment. There are many reasons for this, one of the main obstacles is providing a sound cost-benefit analysis for the business. Often sensors and monitors that are required for the production environment are non standard. Most applications that require a solution to particular problems are solved case-by-case basis with sensors have to be retro fitted toexisting machines, which means the implementation is costly and requires machines down time for access and often ongoing monitoring making the cost benefits harder to quantify . Added to this, for some processes, retro fitting of addition monitoring sensors/devices is not possible due to machine configuration, process constraints and I.T. integrity concerns.


Looking at other types of equipment, elevators for example, the solutions that work are those where the predictive technology is embedded in the product and where it does not require a specific fit i.e. imbedded. In this type of equipment to use of health assessment based on power consumption is widely used.

Looking at “general” industry the relative cost of energy in production processes is growing on an annual basis. However the use of Energy Management Systems (EMS) that are currently available that monitor and optimize the energy efficiency are only deployed by those industries with high demands of energy like paper, steel production, chemical and petrochemical. In these cases, energy management is based on real time information obtained from costly process monitoring and control systems. Of course, opportunities for cost reduction are greatest when both electricity consumption and prices vary over time, which is common in process industries, and open electricity market environments. Typically, the overall cost reduction in these environment can higher than 5 per cent of the total energy cost (given we can maintain a stable supply). It is not only in high consuming process that need to be manage, but it is also possible also manage the energy in buildings or communities with the available metering infrastructure.

All these mean that there is a trend to look for low-cost methods to monitor and analyze energy data in any type of process. This need will be accelerated with the penetration of the so called Smart Grids. If we can use this energy monitoring also to predict machine malfunctions through the power consumption analysis, we will have a clear synergy to facilitate the introduction of predictive technologies in the more complex production environments. Of course, this universal solution should also be compatible with the added value information that could come from existing sources/sensors used at the factory, and jointly this will preserve current and future investment in the field.

Currently there are emerging companies providing “remote services” to companies in order to optimize their energy costs which have been widely accepted. So, it should be more easily accepted companies/services that complement this type of services with added value information about the health of the production system. Of course these service companies could also work for the production machine OEM (Original Equipment Manufacturer), giving them the adequate feedback about their machine behavior over time.