Search

Warehouse Design – Modelling and Simulation


1. Introduction


In my series of Blogs, we have looked at various aspects of warehouse design and how the application of automation can lead to significant improvements in warehouse efficiencies and the volumes processed. The evaluation of the alternative concepts is a vital component in the decision-making process. Any changes to a warehouse design, or any investment in automation, will come at a considerable cost, and these decisions should be based upon accurate modelling and simulation.


Modelling and Simulation will give you the confidence and clarity to make decisions, backed up by analysis. Different solutions and concepts can be compared, and the impact of forecasted changes to the volumes processed and anticipated changes to order profiles can be analysed.


The review process starts with detailed data analysis, this has been the subject of another Blog – so please refer to that for more detail (www.thlogisticsconsultant.com/post/warehouse-design-data-analysis). This Blog will cover 2 principles – Modelling and then Simulation, summarised as follows:

  • Modelling – supported with AutoCAD Layouts and spreadsheet modelling and creation of 3D Sketch Up Images and videos

  • Warehouse Simulation using K. Sight CLASS Simulation software (hereafter referred to as CLASS)

The next Blog to be published will focus on Demo 3D, this software is used extensively by the major automation suppliers. The Demo 3D software has simulation capability but is primarily used to test and simulate automated equipment, and not usually extended to cover all warehouse operations.


2. Modelling and Simulation


Overview


In previous Blogs we have looked at different storage solutions, including:

  • Block Stacking,

  • Shuttle Racks,

  • Very Narrow Aisle Racks,

we have also looked at automated solutions including Goods to Man systems such as:

  • Robots (Kiva / HIK Robots / Fetch),

  • Collaborative robots and Swarmbots

  • Autostore

The latest Blog discussed alternative option for Sorters – (of which there was 9 different types discussed). There are other automated and technology solutions that would also offer efficiency gains, and these too would be ideal candidates for evaluation using modelling and simulation, these include:

  • loading and unloading systems,

  • automated packing and labelling equipment,

  • Automated Storage and Retrieval Systems (ASRS),

  • Miniload systems,

  • Conveyor systems,

  • Robot picking solutions,

  • Mobile robots and AGV’s,

  • Pick to Light

  • Put to Light

  • Voice Picking

Note: the above concepts will be covered in future Blogs.

Evaluation Project Overview


For any modelling or simulation project the approach is similar. Typically, as a consultant the process is as follows:


1. Project scoping meeting.

The initial meeting usually involves the Key Stake Holders and the Project Sponsor. This meeting should include a site tour if possible to view the operation and to observe any key areas that may need further investigation, the reasons why the evaluation is required and the options to be evaluated, and at this point some high-level data might be supplied for context.


Following this meeting the consultant can document the scope of work to be completed and create a project proposal document. The documents will also list all of the expected outputs, these will include presentations, written reports, layout drawings, and results from any analysis performed including spreadsheet and simulation reports. The proposal document will also outline the commercial terms of the contract and will include a project plan.


2. Kickoff meeting with client


Once the project starts and the consultant has been engaged, the first task is the project Kickoff meeting. At this stage there is likely to be the opportunity to view the operation again, and to meet the key stakeholders and project team. Depending on the complexity of the project, the consultant may send a data request proforma in advance of the Kickoff meeting, or immediately after.


A typical Kickoff meeting agenda may be as follows:


Introductions

Client / Consultant – the background to the project is explained

Project– why is the project required? Typical issues to be addressed include:

  • Increased order volumes

  • Changes to order profiles

  • SKU range changes

  • Additional products / product groups

  • Site relocation

  • Site extension

Scope – what is in / out of the scope of the project

Approach – what is the process, what systems and analysis tools will be used

Roles – the project plan can be reviewed; the tasks can be allocated to team members as required and timescales discussed.

Next Steps – the frequency of meetings can be discussed, and the short-term tasks can be clarified. At this project status reporting and methods of addressing issues will be discussed.

Q&A

3. Data Analysis

Data analysis, this is covered off in detail in the Blog – Warehouse Design – Data Analysis.


For any operation that is looking to evaluate or implement any of the above options I would strongly advise that this process is not overlooked, and that all of the volumes processed through the warehouse are understood and documented.


Once the data has been analysed then this should be reviewed as follows:

  • Current average volumes – use actual volumes from an average week in previous 12 months, as late as possible

  • Current peak volumes – use an average week from a peek month

  • Future average – if there is a forecast for growth then the volumes should be uplifted, for example calculate expected volumes for year 2025 if there is a 5-year growth plan.

  • Future Peak volumes – as above, but to show average peak.


Comment: the results generated from an average week will help identify how an operation will operate for the majority of the year. Peak weeks will most likely be covered with additional short-term staff and equipment, and extra shifts may be utilised including weekend working. It is important to consider the overall effect of peaks and the growth over the planning horizon will impact on any concepts being considered.


Data Sign Off Meeting: Once the data has been analysed the consultant should hold a separate sign off meeting with the key staff to confirm the data has been interpreted correctly, and that any assumptions made are valid. It is highly likely that this part of the process will uncover some surprises for the clients and will probably result in some re-evaluation of the analysis.


The operational and system process should also be documented. The following shows some example reports:


Sample Project Analysis Report – Material Flow Diagram


Sample Project Analysis Report – KPI Listing


Any project will most likely result in a major rethink on the storage strategy or picking strategy. A common theme amongst recent projects is how to handle increased Ecom orders. Typically, many projects will look to achieve higher pick rates and how to handle increasing volumes, this will also drive a requirement to include more efficient packing and sortation. The recent introduction of new robotised systems has also driven increased scrutiny in warehouse operations.


There are so many different ways to store and pick products in a warehouse it is highly likely that numerous layouts and concepts may be evaluated.

Sample Layout created in AutoCAD



Warehouse Productivity Rates

It is vitally important that the task rates for the various warehouse processes are known. The task rates, and the actual performance against the rates should be verified. Picking typically accounts for 60% or more of the labour hours, and consequently is the area of highest importance.


If the task rates are known then these can be used to populate a resource model, if not, then the rates can be generated using work study techniques with a stop watch, or the rates can be set using a system such as MOST or MTM. MOST task rates can be generated relatively quickly, and very accurate results can be achieved.


Note: MOST is a system developed using methods-based techniques to determine the time it takes to perform tasks. The time that results from the most methods is for an operation to be performed at a standard rate of performance, (100% rating). Most uses a sequence of predetermined motion patterns called sequence models.

Note: always make sure that you use certified and qualified engineers to conduct this analysis. Check their qualifications – my certificates are displayed on my website.

Sample Project Analysis – MTM Task Rate Calculator



Warehouse Modelling


Once the data analysis stage has been completed, and we know the task rates for the operations in the warehouse we can now start building a resource model. This model is used to determine the level of resources, both operators and MHE to meet certain levels of volumes processed through the various stages of a warehouse operation.

Apart from fully automated warehouses, all warehouses will have many different types of manual operations. These operations could also include tasks with interactions with automation. Different types of automation will have different impacts on the tasks performed by the warehouse operators and equipment deployed. For each of the types of automation we can calculate the throughput rates and include this in the analysis.


If we are to introduce new tasks to the operation, i.e. loading a new automated piece of equipment, or a new picking methodology then we can generate new task rates using the MOST system.


The modelling should consider an average period of activity and also a peak level of activity, and any future volume assumptions should also be considered. Seasonality trends should be considered for both inbound and outbound activities.

Sample Project Analysis – Outbound Analysis


Sample Project Analysis Report – Day in the Life (DILO) Diagram




Conclusions relating to Modelling


Once the modelling process is complete, we will have:

  • Resource Model Outputs for each of the variants modelled.

  • Current / future – average / peak weeks capacity / resource calculations

  • Budget costs of each operation, these costs can then be entered into cost modelling tools to compare cost per unit / per order. Any investment in new automation could be put through a Return on Investment analysis tool.


We can now make informed decisions as what the next steps should be, backed with analysis.


Although the modelling of operations using a spreadsheet may not guarantee results to the same level of accuracy as a simulation tool such as CLASS, for the purpose of generating results quickly this level of analysis may be sufficient. This process usually identifies many surprises for the clients and helps them fully understand how new concepts may not generate the actual improvements anticipated. This level of analysis should be conducted in the first instance before moving onto Simulation or engaging fully with automation suppliers.


Should the modelling identify that there are significant gains resulting from the analysis of the concepts being evaluated, then the operation should be simulated using either CLASS or Demo 3D.

Sample Project Analysis – Resource Calculator


3d Sketchup

In addition to creating 2D layouts using AutoCAD, 3D Sketchup software is used to generate scaled images and videos, these can then be used for presentations and proposals.


Videos and 3D images can be a valuable communication aid to demonstrate changes, especially if an existing warehouse is to be extended, or if the warehouse layout including the racking and shelving is to be modified. Any new automation can also be included in the images and videos.

3d Sketchup


Note: see the Services section of my website for a sample 3D image and video.

https://www.thlogisticsconsultant.com/3dimagesandvideossketchup


K. Sight CLASS Simulation

Link to CLASS website:

https://cirruslogistics.com/products/class/warehouse-simulation/


The CLASS software been in development for over 25 years, and I personally have been a user for over 20 years. The software is used to create 2D and 3D warehouse layouts and to simulate warehouse operations. The software is used by most of the major 3rd Party Logistics providers and many major retailers and in-house logistics operations all over the world.


Users of CLASS attend a 3-day training course which covers the core skills required to complete detailed warehouse layouts and execute simulations. There are additional courses for more advanced topics and special areas such as site traffic modelling, (this extends the warehouse simulation to include all vehicles movements into / around and when exiting the site, this can identify potential bottlenecks in:

  • the gate house / security areas,

  • parking areas,

  • queuing areas,

  • capacity of the shunters.


Warehouse Layout including Site Traffic Routes and Processing areas


The CLASS simulation software is relatively quick to learn. The users will need to have a basic understanding of warehouse operations, equipment, and processes in order to understand the outputs from the simulation and to make logical changes and to understand the results generated. The software is now on version 16 and users familiar with Windows applications will find it easy to use.


Simulation Stage 1 – Creating the Physical Warehouse.


The process starts with the user creating the physical warehouse using a 2D scaled layout.


Note: if an AutoCAD drawing of the site is available this can be imported to CLASS as an image, which can be scaled to the exact dimensions of the warehouse.


The CLASS layout is created by adding design objects created on layers in the layout. The key dimensions of the warehouse can be specified within the CLASS layout. These may include wall dimensions, dock sizes, racks sizes, aisle widths. The relative X and Y coordinates of items from a known datum point can be specified, e.g. the racks could start at 4 metres from a known point.


Design objects can be positioned in the layout to the exact position required.

The design objects include:

  • Building walls

  • Docks / doors

  • Marshalling areas

  • Racks / shelving units

  • Mezzanine Floors

At the time of creating the design objects various details, such as rack heights, number of levels, dock door heights, wall heights can be specified on screen, by means of setting a Y value. By setting the height of various items the warehouse can be viewed in 3D.


Note: X, Y and Z coordinates are used in drawing software, X = horizontal axis, Y = vertical axis, and Z = 3d height.


The software can generate 2D elevation drawings of racks including dimensions. The dimensions can be set in metric or imperial. Front / Side Elevations can be generated to include dimensions as required. Warehouse layouts can also be generated in CAD format if required.

Rack Elevation


Once the warehouse design has been created the warehouse can be viewed in 3D, various screen shots can be captured and fly through videos can be created.


3D Image of Warehouse


Simulation Stage 2 – Setting the Operating Volumes and variables.


Following the creation of the warehouse the volumes to be processed through the warehouse need to be entered. The software is designed to simulate an operation over a 7-day cycle.


The key data generated through the Data Analysis process will be used to drive the simulation. In the previous section we discussed creating 4 sets of data, generally the “current average” week is used to test and validate the simulation model. This model is considered to be the Base Case model, the initial results from the simulation can then be compared with current operations. The base model can then be calibrated if required.


The data required to populate the simulation is summarised as follows:

  • Inbound data – number of trucks, grouped logically – to include the number or pallets / items per vehicle load. The historic or forecasted arrival times are also assigned to the loads.

  • Outbound data – number of trucks set to leave – size of the load – the number of orders shipped, number of pallets picked and from which locations in the warehouse.


Additional information is required to support the picking activities in the warehouse, to include the average number of lines per order and number of units picked per order. To drive the simulation the process time for the picking operations is input, e.g. 10 seconds per line allowance for scanning / item verification and 5 seconds per pick unit.


Additional information is required to allow for replenishment, to include average travel distances and a time allowance for the consolidation of stock at pick face.


Finally, to allow the simulation to run the number of warehouse operators, the hours worked, the areas they are assigned to and the duties they perform are entered.


The simulation also requires MHE with the appropriate truck types and numbers entered. The truck operating speeds can be set (note: the software has accurate real-life speeds for the most common trucks pre-set). The simulation will allow for lifting times, acceleration, and deceleration times – and will also account for delays due to factors such as congestion in aisles.


In addition to warehouse trucks, the simulation can also include conveyors, AGVs, Automated Storage and Retrieval Systems, Miniload systems, Robotic forklifts, the Amazon / Kiva-type Goods to person picking systems. The processing rates for automated equipment is predefined.


Complete CLASS Model – ready to execute the simulation



When all of the data has been populated the operation is ready to be simulated. The image shows the various simulation objects summarised below:

  • Inbound Flows in green, put away routes and picking routes set.

  • Outbound Flows in Blue.

  • Labour Teams and shifts in Yellow.

Simulation


The final stage in the process is to execute the simulation. The simulation can be viewed on the screen and the progress through the 7-day cycle monitored using the on-screen timer. Any bottlenecks can be instantly viewed as the simulation progresses through the week, such as congestion on docks.

Upon completion of the simulation a suite of reports can be obtained which shows the various status of tasks.


Key Simulation outputs / Reports.


  • Inbound / Outbound vehicle reports

  • Dock utilisation

  • MHE utilisation

  • Labour Utilisation

  • Task Completion – orders picked

  • KPI reports – task rates generated by the simulation for tasks such as picking

  • Order completion times against targets

KPI Dashboard



Modifying the simulation


During the data population stage, it was stated that a base model should be created using the “current average” week volumes. This model can be calibrated, and the operator performance can be adjusted, or other adjustments made so the base model reflects the real life - actual operation.


The simulation can now be modified to allow for other “what if” types of modelling –

  • What if the volumes increase by 20% - (there is a quick uplift function that allows for these types of quick analysis)?

  • What if the truck types are changed for certain tasks?

  • What if we changed the shift start times, the number of operators on each shift?

  • The layout and operation could be changed to introduce some form of automation and the simulation run again.

Each time parameters are changed the results in the dashboard will change.


Simulation – Next Steps


At this stage, the simulation process is complete. Various concepts and layouts can now be compared. The current and future average and peaks weeks can be simulated, and the results analysed, any potential bottlenecks or operational problems can be identified in advance of the problems occurring in real life.


It is possible to evaluate manual operations against automated options. The data extracted can then be used to populate cost models if tendering for business, (especially suitable for 3PL’s), or used in Return on Investment models to justify any investment in layout, system, change in process or automation.


Peak operational resource requirements can be scheduled in advance, short term hire equipment can be ordered, and labour recruited and trained in advance of peak periods.

1. Case Study

3D image showing the Demo warehouse



Whilst working at Cirrus as a Consultant I conducted an evaluation of alternative storage concepts. CLASS comes pre-installed with a demo project; this is a fully functioning made up warehouse of a 7-day operation. There is a VNA reserve bulk storage (on the left of the image), Wide Aisle picking area, a mezzanine shelving area, and a conveyor system. The VNA racks are used as a bulk storage area for full pallets, the pallets are then either sent to the despatch area for shipment, or sent to replenish pick pallets in the wide aisle section of rack (on the right of the image). The wide aisle area is a case picking area and the pallets from VNA replenish the pick faces.


The VNA area has a total of 19,200 pallet spaces, and it is assumed that for planning purposes the maximum occupancy of the area assuming 95% utilisation is 18,240. The purpose of the exercise was to compare how different rack types would impact on the storage space required, and the labour and MHE required to meet the same operating volumes.


The base Demo model was modified to show a total of 8 different rack types, each rack type had a different utilisation factor applied (utilisation factors were discussed in detail in the Warehouse racking Blog www.thlogisticsconsultant.com/post/warehouse-design-warehouse-racking – these are variable based upon stock holding and storage profiles).

Sensible assumptions were made about the height and number of storage levels for each rack type based upon the MHE used to access the racks. Block stacks were considered to be a maximum of 4 levels, the VNA area 8 levels, the Wide Aisles 6 levels. For each rack type the appropriate aisle width was applied based on the MHE used. The following table summarises the footprint required for the bulk storage area for each variant. The total pallet spaces required to achieve the target number of pallets is also shown:


Rack Type – Storage Space Comparison


Example Layout – Narrow Aisle Articulated APR - 10,824 sq. metres

Example Layout VNA - 6120 sq. metres



Exercise Results


For each of the 8 alternatives the simulation was run and in each case the results were totally different. The put away and pick rates were different, the number of employees and trucks required to meet the outputs were different. The travel distances varied considerably.


The outputs were populated into a cost model, and comparisons between concepts quickly identified different:

  • Labour costs

  • MHE costs

  • Space costs (Rent plus Rates, Utilities, Maintenance etc)


cost to install each rack type could also be compared.

The purpose for sharing this story is that the exercise based upon some high-level assumptions, using a fictitious warehouse, generated some really interesting results. The time required to compare one concept to another was very quick.

4. Conclusion

Any investment in new technology and automation, new layouts or new equipment should be based on a sound understanding of the current operation and future growth projections. Detailed modelling and simulation should be used to validate and test any future plans. The value of going through this process of data analysis and detailed design, followed by modelling and simulation should not be underestimated. The process will require a little investment but will reduce the likelihood of very costly mistakes.


If you are considering making changes to your layout, equipment, increasing volumes processed or evaluating any form of automation then please let me help.I can coordinate and provide modelling and simulation.Before committing to any major investment in automation the analysis of the current operation and assessment of alternative concepts should be assessed by impartial, independent experts.If I can help you with any other aspect of warehouse operations design, please do not hesitate to ask for

164 views0 comments

© 2020 Tony Hughes Consulting Ltd

  • Twitter