Grinding mills fulfill an essential role in many manufacturing plants. They are found at different stages of the process, from the initial step of pre-processing the raw feedstock up to the final milling of the product into the required grain size. In all these cases this installation represents an essential element of the full process, and thus one wants it to be in top shape and continuously available. One way to achieve this, is to “run to failure”. At the other end of the spectrum one can choose to conduct an extensive preventive maintenance program based on lubricating and replacing components well ahead of time. Although having their merits, both approaches have one major drawback in common. The issue is not connected to the core financial and organizational aspects of “preventive” or “failure”, but to the fact that both approaches consider maintenance as an isolated operation. As such both ignore significant gains that can be made by looking at the situation from a bit more distance. The present piece tries to shed some light on an alternative view, and how to implement it.
The Background
The process of “grinding” a material is known for far more than 2000 years. It is thus no surprise that in the mean time various principles have been developed and put into action to obtain the same goal: reduce particle size. As a result, today one can encounter different grinder types when stepping into a production site. None can be considered as being the “best”, and most often an individual site selects the most optimal type based on the process it conducts, the specific nature of its material supply, the synergy with the overall process as well as aspects related to maintenance and efficiency. The decision on what specific type of grinder to use is thus not purely based on maintenance aspects, an important clue.
In a grinding process, particles are reduced in size either by applying high pressure, or through impact. This is achieved through the application of a subset of the following components: balls, hammers, cylinders, rotating tables, and/or rotating milling stones. Consequently, the name of the installation type is often derived from the type of element used in the milling process: a ball mill, hammer mill, cone crusher, ring roller mill, vertical roller mill... We will not get into detail on how each type of installation functions, but in general a material supply is fed into the central installation in a batch or continuous mode. The final product will leave the installation either through gravitational forces, or through an air flow taking away the lighter, thus smaller particles. Some operate in a batch mode with consecutive periods of filling and emptying the installation, others process a continuous stream of materials.
It is necessary to step away from a “one sensor → one issue” relationship
Grinding as a process is not specific to an individual industry either. A reduction in particle size is relevant in multiple processes, such as crushing the grains before they enter the kettle in a brewery when making beer, and in the process of producing cement, for reducing the clinker in size and mixing it with gypsum to create the desired end product. This last example indicates some of the variations that can be encountered as well, as in this specific case the installation is used for grinding and mixing at the same time. Other applications can be found in metal production when transforming the ores into the optimal granulometry for further processing, in metal recycling for breaking and reducing size of recovered materials prior to entering in the smelter process and in primary steel making when preparing the coal in the process of making cokes. Even more examples can be found in the production of building materials such as lime and plaster or in the production of more course granulates used in concrete for example.
Possible gains?
As stated in the introduction: historically the main activities conducted on these type of machines have remained very much isolated. Depending on the company philosophy maintenance was done on a very regular basis (preventive) or whenever a machine broke down or started making a lot of strange noises. A different department implemented improvements or optimizations based on long-term insights, theoretical studies or best practices picked up form others. This approach has its merits, but today we know that it results in a lot of value being lost for the company. Capital efficiency is however key in the present industrial world: create as much value as possible for a given amount of capital injected. For installations such as grinders this can only be reached through combining multiple domains that used to be managed by different entities: process efficiency, product quality, maintenance and production. In an ideal case also the department responsible for investment and upgrades is involved. A choice or decision made in any of these domains will impact the other aspects, knowingly or unknowingly, these impacts not necessarily being negative. The only way to quantify these influences and make this process tangible is by implementing an integrated, multi-aspect monitoring approach, including all relevant parameters related to the process as well as the behavior of the installation. As one should strive to optimize beyond the scope of pure maintenance, this means that the approach also has to include the parts of the installation involved in feeding the raw materials as well as the components involved in extraction and/or separation.
And how to obtain these?
The optimal way of running and managing the installation cannot be obtained by implementing some vibration monitoring on the motors on the one hand and some regular offline product quality monitoring on the other hand. What is required, is an approach where relevant parameters are measured at the appropriate locations and combined with process and product data already available on-site. Using such an approach where all the data is gathered and continuously automatically analyzed in a central location allows one to optimize a broad range of aspects. Some of these are relatively straightforward, such as the health of motors, gearboxes, bearings and transmissions. It is however important to not just focus on detecting degradation, but rather on identifying conditions that would lead to degradation, in order to prevent it. Other aspects are more process related, such as ensuring a smooth startup process that is also short, such that energy consumption is reduced and efficiency as well as product quality can be increased. The same tool will also help to achieve a smooth startup and a smoothly running process, leading to reduced wear as well as a much more stable product quality. This in turn ties closely to optimizing the process speed to still reach optimal product quality, but avoiding inducing too much wear and without impacting product quality. By tracking the evacuation or extraction stage as well, associated issues that would lead to halted operations can be identified and effects such as clogging of tubing, leading to higher energy consumption and productivity losses, can be tracked too, allowing to take timely action. Last but not least this approach also allows to determine optimal process conditions in case different product types are processed on the same installation or when additives are involved, where optimal dosing impacts product quality, cost and/or throughput of the installation.
All of this is based on the knowledge that it is necessary to step away from a “one sensor → one issue” relationship. In reality the data collected using a single sensor can and has to be used to serve multiple purposes. An illustration: electrical currents and potentials, when measured with sufficiently high resolution and at sufficiently high frequency, can be used to track motor health, energy consumption, and process efficiency, but also as a part of a feedback loop to maximize throughput. On the other hand it is also often necessary to combine data from multiple sensors in order to implement a decent follow-up of a specific phenomenon. In order to track clogging of the ducts for example it is necessary to combine data about the RPM of the fan, the temperature as well as the pressure drop in multiple positions along the route. When correlating all these parameters, indeed clogging or sub-optimal performance can be identified and reacted upon.
Combining all relevant aspects
Ultimately the economical efficiency of an installation is determined by how much can be delivered (money in) for a certain amount of input (money out) for a given unit of time. Money comes in for each quantity of material produced within the required specifications. Costs are related to initial investment (CAPEX) and all process inputs: base materials, maintenance and repair, energy and labor (OPEX). When implementing a fully integrated smart monitoring, the overall balance will become more beneficial. Using the same platform one can have an impact on 4 aspects at the same time:
- Ensure a maximal availability and throughput: by avoiding sudden stops and making sure the installation can work at full regime
- Increase energy efficiency: a more efficient process means a lower energy consumption, and tracking the relationship between energy consumption and product quality as well as production rate allows the user to operate according to optimal setpoints.
- Improve process efficiency and product quality: the smoother the process runs, the more consistent the product quality will be, also leading to lower amounts of losses. Examples are products that have to be treated twice, prolonged or frequent startup or shutdown cycles…
- Reduce running costs: grinders inherently have “wear parts”. When monitoring the wear and process efficiency together, the optimal moment for replacement can be identified. This in turn also impacts product quality and throughput.
Conclusion
It is thus safe to conclude that, even on a relatively well known setups such as a grinder installation, still a lot of gains can be obtained. As a result of increased costs of energy and spare parts, this optimization is to be done all through the operational life and as a multitude of aspects plays a role, it cannot be obtained purely based on physical or straightforward relationships. Measuring the right things, combining data streams, and injecting them into a smart analysis process enables one to continuously detect and implement operational improvements and avoid unexpected stops and high intervention costs through early detection and proper understanding of the background of deviations