How do I assess the OEE potential of a modern powder mixing plant with automated cleaning?
The assessment of OEE potential should be based on real batch, time, quality and cleaning data. OEE stands for Overall Equipment Effectiveness and results from availability, performance and quality:
OEE = Availability × Performance × Quality
In powder mixing plants, charging, dosing, mixing, discharging, cleaning, product changeover, sampling and release must all be taken into account. Whether planned cleaning forms part of the planned production time must be defined clearly in advance. Otherwise OEE values between plants, products or sites are not comparable.
Establishing baseline data
The basis is a sufficiently long reference period with a representative product mix, for example several weeks or complete campaigns. Start and end times of charging, dosing, mixing, discharging, cleaning, changeover and product changes should be recorded. In addition come unplanned stoppages, waiting times for material or releases, faults, remixing, scrap, rework and the batch quantity actually achieved.
Availability describes the proportion of the defined production time during which the plant is actually ready for operation. Performance compares the actual batch or cycle time with a product- and recipe-specific target time. Quality assesses the proportion of batches or product quantities that can be released without rework. Not every quality deviation should be counted as an OEE loss as a matter of course; the attribution of remixing, scrap, yield loss and waiting time must be applied consistently over time.
A fixed data-collection period of four to eight weeks is not generally sufficient. Where product changes, seasonal recipes or faults are infrequent, a longer period may be necessary. What matters more than a calendar period is that the data reflect actual operation.
Assessing cleaning
Automated cleaning can improve availability if it reduces dismantling, manual activities and process variation. Cleaning in Place, or CIP, refers to cleaning in the installed state without extensive dismantling. Wet in Place, or WIP, refers to wet cleaning in the installed state. The achievable time saving depends on the plant, the product and the degree of soiling; general figures such as 30 to 90 minutes for CIP or several hours for manual cleaning are therefore not reliable.
The assessment should capture the complete changeover time: discharge, securing the plant, preparation, cleaning, rinsing, drying, inspection, sampling where applicable, and release. Only this gives the production time actually lost. Activities carried out in parallel, such as pre-weighing the next batch, can reduce the net changeover time, provided personnel, material provision and safety procedures allow it.
Cleaning sensors such as conductivity, flow, pressure, temperature or return-flow turbidity can monitor the process. However, they are not automatic proof that all product-contact surfaces are clean. A riboflavin test, for example, assesses the wetting and possible spray shadows of a cleaning process; it does not replace product-specific cleaning validation. The FDA expects written cleaning procedures and a documented assessment of their effectiveness.
Condition-based cleaning can be sensible where soiling and cleaning need can be assessed with suitable, reliable data. In many regulated applications, fixed, validated cleaning programmes remain required. Shortening or omitting cleaning cycles may only be done if it is supported by risk analysis, data and, where applicable, validation.
Improving performance and quality
Performance potential is often found in mixing time, dosing, charging, discharging and material transfer. Target times should not be defined as blanket machine values, but should apply to the product, batch size, recipe and permissible process window. Shortening the mixing time is only sensible if the required homogeneity and product quality continue to be achieved reliably.
Inline analytics can help, with suitable products, to assess mixing progress more quickly, for example through spectroscopic inline moisture measurement. Torque or current profiles provide indications of product behaviour but are not general proof of homogeneity. Their suitability as a mixing-endpoint criterion must be demonstrated for the recipe, sensor position and process objective.
Further performance drivers are short interruptions, dosing problems, bridging, slow discharge, residual quantities, conveying faults and waiting times. A Pareto analysis ranks loss causes by their contribution to lost time and focuses improvement measures on the largest causes. Condition monitoring can help to detect changes in the drive, bearings, seals or discharge elements at an early stage. It requires suitable measured variables and clearly defined responses; sensor technology alone does not prevent a fault.
The quality component covers above all batches that cannot be released, scrap, rework and, where applicable, remixing. Relevant factors are mixing quality, moisture, particle-size distribution, yield, product build-up and the risk of segregation during discharge and conveying. The discharge strategy, drop heights, conveying technology and container geometry should therefore be part of the assessment. Process Analytical Technology, or PAT, refers to the use of suitable measurement and analysis technology to monitor and control manufacturing processes. It can make quality deviations visible earlier but must be qualified or validated for its specific purpose.
Calculating the potential
In practice, the following sequence is recommended: first, loss times and quality losses are recorded in a consistent event structure. The largest sources of loss are then prioritised by Pareto analysis. Only realistically achievable scenarios are then calculated, for example a demonstrably shorter cleaning time, a reduced share of remixing or faster discharge.
The economic benefit does not arise solely from a higher OEE figure but from additional usable production capacity, avoided rework, reduced scrap, lower use of cleaning agents and possible investment or staffing effects. A free machine hour only has an economic value if it is actually used for additional production or removes a capacity bottleneck.
Blanket benchmarks such as 65 to 85 percent OEE have little validity for batch-wise powder processes. Cleaning requirements, campaign lengths, product range, containment, quality releases and the chosen definition of planned production time strongly influence the value. Plants with the same time model and a similar product portfolio are the ones most suitable for comparison.
Assessing and unlocking OEE potential with amixon®
amixon® does not assess OEE potential using general guide values, but on the basis of the specific product, the intended batch size and the complete process chain. In the pilot plant, mixing time, discharge, cleaning effort, product changeover and, where applicable, dosing are examined with the original product. Realistic target times and a robust comparison with the existing plant can be derived from this. Around 35 test units are available for this purpose in Paderborn; supplementary pilot plants are located in the USA and several countries in Asia.
Availability can be supported by the design of the mixing plant. In many amixon® mixers, the mixing tool is mounted exclusively above the mixing chamber. This eliminates a lower shaft passage as an additional product-contact sealing point. Low rotational speeds and a reduced number of wear-relevant components can lower the maintenance effort. In addition, amixon® supports operators with inspections, maintenance, modernisation, spare parts supply and, on request, condition-based maintenance concepts. The actually achievable availability continues to depend on the product, mode of operation, cleaning, maintenance strategy, spare-parts management and operation.
For the performance side of OEE, mixing time, discharge and product changeover are particularly important. The vertical mixer type HM generates superimposed mixing flows and can thereby support intensive yet gentle mixing. For the KoneSlid® mixer type KS, amixon® states an ideal mixing quality after approximately 20 to 30 revolutions for suitable applications, as well as short discharge times. These values should be understood as product- and application-related guidance and must be confirmed for the specific recipe.
ComDisc® technology can convey product residues towards the discharge and thereby improve yield, product changeover and cleaning effort. A high degree of residual discharge reduces the amount of product that remains in the mixer and subsequently has to be cleaned or discarded. The achievable discharge performance depends on product adhesion, moisture, particle size, bulk density and equipment design. It is therefore sensibly assessed as part of the trials with the original product.
For automated wet cleaning, amixon® offers, among other things, the WaterDragon® system. The washing lance moves automatically into the mixing chamber, cleans with water at a comparatively low pressure and can be combined with automated drying. Good residual discharge before cleaning begins can reduce water and cleaning effort. The actual cleaning duration and cleaning effectiveness, however, depend on the product, soiling, geometry, nozzle arrangement, flow rate, temperature, cleaning chemistry and drying. Automated wet cleaning is not automatically validated cleaning; its effectiveness must be demonstrated for the intended process.
With the container mixer type COM, weighing-in, mixing, cleaning, intermediate storage and filling can be organisationally decoupled. This can improve the utilisation of the mixing station with frequent product changes, provided sufficient Mixtainer®, cleaning and transport capacity as well as suitable release procedures are available. The cleaning time therefore does not necessarily limit the overall performance of the system to the same extent as with a single, permanently installed mixing station.
The quality component of OEE can be supported by reproducible recipe management, suitable mixer design, high dischargeability and controlled product changes. Mixing programmes can be stored in a programmable logic controller, or PLC, and executed batch by batch. Barcode-supported material identification as well as project-specific interfaces to Enterprise Resource Planning systems, or ERP systems, or Manufacturing Execution Systems, or MES, can link recipe, batch and process parameters. For such data to be usable for traceability, root-cause analyses and OEE evaluations, the data model, timestamps, user rights, data integrity and downtime categories must be clearly regulated across the overall system.
For regulated applications, amixon® can provide qualification-relevant documentation and support Design Qualification, Installation Qualification and Operational Qualification, or DQ, IQ and OQ. The technical design can be aligned with project-specific requirements such as EU GMP, FDA 21 CFR Part 11, EHEDG, 3-A Sanitary Standards, ATEX or ASME. Responsibility for process and cleaning validation as well as for the validation of electronic systems remains with the operator.
amixon® states that it develops and manufactures centrally in Paderborn. The high depth of manufacture, qualification-relevant documentation, pilot-plant trials and long-term spare-parts supply support a life-cycle view in which not only the initial purchase but also cleaning times, product losses, maintenance and adaptability feed into the OEE assessment.