Warehouse Order Picking Optimization

Submitted by:

Noura Alorainy – 161653

Reem Aljammaz – 161409

Sarah Aldaher – 151106

IE490: Industrial Engineering Capstone Project

Advisor: Dr. Abdullah Alrashdan

22 March 2020

ABSTRACT

Nicene, an online site and for beauty cosmetics, has experienced significant growth in its market demand ever since the launch in 2017. Therefore, they aim to improve the growing warehouse operations. Thus, optimizing the manual order-picking process. This project applies different engineering tools that will streamline the overall order-picking process—starting from building a simulation model to test the proposed approach, then using a decision matrix with the weighted scoring model to reduce order-picking errors. Furthermore, lastly, applying the Engineering Standards and Mechanical Specifications of the new conveyor belt system.

LIST OF FIGURES

Figure 1. Current Simulation Model via Any Logic

Figure 2. Cause-and-Effect Diagram of Mismatched Orders

Figure 3. Proposed Simulation Model via Any Logic

Figure 4. Warehouse Packages Dimensions

Figure 5. Recommended Conveyor Belt

Figure 6. Elbow Height

Figure 7. Sitting Chair

Figure 8. Anti-Fatigue Mats

Figure 9. Footrest

Figure 10. Suggested Conveyor Belt Dimensions

Figure 11. Current Picking List with Unnecessary Elements Highlighted

Figure 12. Improved Picking List

LIST OF GRAPHS

Graph 1. Current Process Flowchart

Graph 2. Existing Warehouse Maximum Output Capacity

Graph 3. Existing Warehouse Maximum Output Capacity via Any Logic

Graph 4. Improved Process Flowchart

Graph 6. Proposed Warehouse Maximum Output Capacity via Any Logic

LIST OF TABLES

Table 1. Time Study (System Observations of Pickers)

Table 2. Time Study (System Observations of Scanners and Packers)

Table 3. Causes and Preventive Methods (Area of People)

Table 4. Causes and preventive methods (Area of Method)

Table 5. Causes and preventive methods (Area of Environment)

Table 6. Proposed System Outputs via Any Logic

Table 7. Proposed System Outputs via Any Logic

Table 8. Weight of Each Criterion

Table 9. Cost of Implementation Scorning Model

Table 10. Reliability Scorning Model

Table 11.  Error Proofing Degree Scorning Model

Table 12. Evaluation of Methods (Pick List Errors)

Table 13. Evaluation of Methods (System Error)

Table 14. Evaluation of Methods (Shelving errors A)

Table 15. Evaluation of Methods (Shelving errors B)

INTRODUCTION

Background

Nicene is an online mobile site application for beauty cosmetics, hair, and skincare products. Since the launch of the online store in 2017, Nicene managed to become one of the largest e-commerce sites in Saudi Arabia and the GCC with its vision and strive to become the leading pioneers in providing products with the highest quality standards and most top delivery speed.

Project Objectives

The project aims to optimize the order-picking process by:

• Streamline the order-picking process, thus increasing productivity

• Reduce the number of mismatched orders
• Introduce ergonomic solutions in the workplace

Project Scope

In this project, we aim to streamline the overall manual operations related to the order-picking process at Nicene warehouse in Riyadh by introducing a new order-picking approach, applying error-proofing methods to minimize mismatched orders in the warehouse, and considering ergonomic solutions in workstations.

PURPOSE STATEMENT

At Nicene warehouse, there have been several errors linked to the order-picking process. The order-picking system currently applied is the Picker-to-Order, where workers pick a single order in every trip and then place it in the scanning station for registration and inspection. After that, rules are moved to a two-worker station for final packaging before dispatch. It has noticed that this approach causes the workers to crowd in aisles, leading to further clutter of orders in scanning stations. Although laws picked manually, it has also noticed that several mismatched orders being shipped daily due to the lack of sufficient quality checks on each request.

Model Approach

To further understand how operations ruined in the warehouse, the below process flowchart illustrates the top-down understanding of how the process works.

METHODS

In this project, multiple methods used to analyze and improve the order-picking process in the warehouse. They were starting from building a simulation model to testing the proposed approach, then using a decision matrix with a weighted scoring model to reduce order-picking errors. Furthermore, lastly, applying the Engineering Standards and Mechanical Specifications of the new conveyor belt system.

Streamline the Order-Picking Process

Simulation provides a way to assess whether the current process, Picker-to-Order, is delivering the optimum output rate among the different stations and inspects areas of the unbalanced number of workers who need redistribution. Besides, the newly proposed solution, Pick-and-Sort, will be tested through simulation since it would be hard to run multiple experiments on-site.

To understand how the warehouse is operating daily, numerous visits were conducted on different days to collect data of the three main stations; pickers, scanners, and packers. The tables below show data on the output rate of different workers during the day. The data were collected based on a random fashion and demonstrated on the time-study tables below.

Graph 2. Current Warehouse Maximum Output Capacity

These findings tested for verification on a simulation model that designed via Any Logic software concerning these main stations:

• Picking
• Scanning
• Packing

The model considers determining a discrete event model with processing time triangularly distributed Scanning Packing stations.

Reduce Order-Picking Errors

To understand the causes of inconsistent orders, there have been many warehouse visits that included:

1. Visual inspections of the working environment, process, and workers.
2. They were interviewing the warehouse manager to learn about hidden causes that might result in an inconsistent order.
3. Analyzing the system with the team to identify the causes of different rules.

A cause-and-effect diagram used to organize potential causes:

Figure 2. Cause-and-Effect Diagram of Mismatched Orders

After identifying the causes, we started to deliberate, brainstorm, and search for potential prevention methods and solutions. For a more organized approach, the possible solutions are listed based on the area of where the cause might have occurred.

Area 1: People

Table 3. Causes and Preventive Methods (Area of People)

The preventive method of Training workers on the purpose of the process and the source of human error will not be considered for further analysis since the solution is straightforward.

Area 2: Method

Table 4. Causes and preventive methods (Area of Method)

Area 3: Environment

Table 5. Causes and preventive methods (Area of Environment)

The preventive method of implementing practical housekeeping will not be considered for further analysis since the solution is straightforward.

ANALYSIS

Streamline the Order-Picking Process

Graph 3. Current Warehouse Maximum Output Capacity via Any Logic

As seen in Graph 2&3, the simulation run via Any Logic represented similar readings and, therefore, verified that the conceptual model had reflected accurately in the digital representation.

The findings of the current process show that although the total number of workers is 60 in the warehouse, the process is not running smoothly. First of all, the pickers who take long trips along the warehouse to collect orders, have up to 14% higher output rate in comparison with the scanners. These caused by the cluttered laws at the scanning station, where scanners take quite a long time to scan and check every item before moving orders to the next station. Secondly, the final packers are finishing up the final packaging of rules with a 19% lower rate than the first packers, who both are supposedly performing in close rates.

Model and Simulate the Pick-and-Sort System

The Pick-and-Sort approach no longer requires an order to be picked in its entirety in one picker tour. Instead, orders are picked in batches then placed in a sorting area, where sorters assemble individual items into customer orders. This approach will be tested via Any Logic to facilitate the order-picking process as well as introducing a new inspection unit, which is the sorting station. Furthermore, the new layout will contain four stations: picking, sorting, scanning, and packing and a conveyor belt connecting the scanning and packing stations. Lastly, the below process flowchart illustrates the top-down understating of the new proposed system.

Graph 4. Improved Process Flowchart

Have studied a new approach, an experiment was conducted in the warehouse to collect data on workers’ output rates. Then, these readings included in the new layout design that consists of:

• Picking
• Sorting
• Scanning
• Packing

To find an optimum number of worker needed per day for the process, Work Cell Balancing equation used:

• Cycle time =
• Total time required to perform a task =

• Number of workers needed =

Based on the equation showed above, the optimum number of workers needed daily is 48 workers who will be distributed based on the new layout spacing as shown:

Figure 3. Proposed Simulation Model via Any Logic

Picking and sorting are the two stations that will have the most modification in the model. View Table 6&7, a conveyor, will be added in the scanning and packing stations to expedite the

operation.

Table 6. Proposed System Outputs via Any Logic

Table 7. Proposed System Outputs via Any Logic

Graph 6. Proposed Warehouse Maximum Output Capacity via Any Logic

The new process starts by picking a batch of 5 orders, then place them in sorting station; it has a quite close output rate with around 3% difference. Also added conveyor, scanners are now working in one parallel with the packers and have a close output rate with only a 2% difference.

Reduce Order-Picking Errors

To further analyze the preventive methods and ultimately choose a bias-free way that fits in the warehouse requirements and constraints. A Decision Matrix with Weighted Scoring Model used to arrive at a well-analyzed choice.

First: Choosing the Most Relevant Criterion.

The criteria used to evaluate each method are:

1. Low Cost of Implementation
2. High Reliability
3. High Error Proofing Degree

Second: Giving each criterion A Weight.

Each given a weight based on importance to the warehouse manager at Nicene

Table 8. Weight of Each Criterion

Third: Creating Scorning Model for Each criterion.

1. The Cost of Implementation:

The cost set as a benefit criterion, where the higher the score values are, the better it is. Giving the lowest score if the price is the top and highest score if the price is low

Table 9. Cost of Implementation Scorning Model

2. Reliability of The Method:

The reliability set as a benefit criterion, where the higher the score values are, the better it is. Giving a high score if automated or low rating if conventional/manual

Table 10. Reliability Scorning Model

3. Degree of Error Proofing:

The Degree error-proofing set as a benefit criterion, where the higher the score values are, the better it is. Giving the lowest score to detecting the error, then the average rating for preventing the failure and highest score for the method that detects then prevents the crash.

Table 11.  Error Proofing Degree Scorning Model

Forth: Scoring and Calculating the Total Score for Each Method.

The Weighted Sum Model equation used to calculate the final score for each method suggested, and it is defined as follows:

Where,

WJ denotes the relative weight of the importance of the criterion.

a_ij was the performance score of the preventive method when it evaluated in terms of the standard.

A_i^WSM-score is the total score for the preventive process.

The best preventive method is the one that has the highest total score among the alternatives.

The preventive methods are listed based on the area where it will facilitate a solution. Since all of the calculating done using excel, the total scores for each suggested preventive method are as follow:

Area 1: People

• Cause1: Human Errors due to picklist

Table 12. Evaluation of Methods (Pick List Errors)

We can see the highest total score among the alternatives is for making pick lists clear and easy to read.

Area 2: Method

• Cause1: System Error

Table 13. Evaluation of Methods (System Error)

The highest total score among the alternatives is for reengineering the process by adding extra inspection units.

• Cause2: Shelving errors

A.

Table 14. Evaluation of Methods (Shelving errors A)

B.

Table 15. Evaluation of Methods (Shelving errors B)

We can see that for both shelving errors, the highest total score among the alternatives is for radio frequency systems.

RECOMMENDATIONS

Streamline the Order-Picking Process

Pick-and-Sort System

This approach has balanced and streamlined the process, as seen in Graph 6. The output rate in all stations is quite similar, with an additional inspection unit at the sorting station that will improve quality checks in Table 13. The new conveyor belt system will also declutter scanning and packing stations, thus organizing the warehouse layout.

Apply Engineering Standards

Improve workstations to engineering standards, and ergonomic recommendation is as follow:

1. Ergonomic Standards

1. The first aspect is the adjustment of the conveyor belt height up to the 95th percentile (Martin Helander).

• Another height adjustment based on the standing elbow height, which should be the 95th percentile. For a male, it is 120 cm. The conveyor height should also be adjusted at 10-5 cm below elbow height, effectively making it 115 cm (Martin Helander).
• The height selected ranges from 65 to 120 cm, as instructed by OSHA requirements (United States Department of Labor).
• Since the majority of conveyor belts have a fixed height, it recommended adjusting the strap to a height that is comfortable for tall workers. A provision should also be for short workers, such as using work platforms or chairs (OSHA).

Figure 6. Elbow Height

Figure 7. Sitting Chair

The sitting chair can be on the conveyor belt for short workers. It should have padding that is about 3 inches thick (Working in A Standing).

1. An anti-fatigue mat should reduce cases of tiredness as well as feet problems as a result of extensive working hours on hard floors in the warehouse. (OSHA)

• The facts warehouse, mats of the tensile strength of 32 lbs/in, and compression of 0.30 lbs/sq.in should use.
• Anti-fatigue rugs have a compression feature that helps in minimizing body fatigue after working for an extended period.

Figure 8. Anti-Fatigue Mats

1. Room for adequate leg and knee clearance (OSHA)

• The suggested approval is 720 mm.

1. Provision of footrests (OSHA)

Figure 9. Footrest

1. Conveyor Should be accessible from both sides of the production line (OSHA)
2. Small shelves or containers must be within reach (OSHA)

Figure 10. Suggested Conveyor Belt Dimensions

1. Essential OSHA Work Practices and Administrative Controls:

• Develop and implement safe operating procedures for conveyors and conduct periodic inspections to ensure compliance.

• Only trained individuals should be allowed to operate conveyors and only trained, authorized staff to perform maintenance.
• Employees working with or near conveyors should be polished regarding the location and use of emergency stopping devices and the proper procedures for conveyor operation.
• Employees should be forbidden to ride on conveyors
• Employees should be educated to lubricate, align, and maintain conveyors when the conveyor is stopped. If this is impractical, advise workers to perform this work at a safe distance from any in-going nip points or pinch points. Installing extended oiler tubes and adjusting screws will help in these instances.
• Employees working with or near conveyors should be prohibited from wearing loose clothing or jewelry, as well as require them to secure long hair with nets or caps.
• Perform servicing and maintenance under an energy control program per 29 CFR 1910.147.

Engineering Standards and Mechanical Specifications of the Conveyor Belt System

1. Width

Since the most significant package in the warehouse is 407mm the suggested conveyor will be 500mm as the recommended width by (Monk Conveyors) ranges from 300-600 mm for small parcels,

Figure 4. Warehouse Packages Dimensions

1. Material for the belt (Materials)

It is recommended for the conveyor belt to be constructed from the following documents:

• A carcass that includes steel cords and textile plies
• Covers made from PVC
• Extra protection materials for impact and edge protection

1. The thickness of the belt (Good Year)

The thickness of the conveyor belt covers both the steel cord and fabric to suit the warehouse application, and their width depends on the type of the belt and application. Different regulations apply to round, single strand, and vertical and inclined belts. In our case for light packages, a 2-ply belt suggested, and it has a thickness of 5/32”.

1. Structural Components: (Materials)

It can be made from aramid fibers because they offer excellent impact resistance, low elongation, toughness, and resistance to damage.

1. Motor Size (Bodine Electric, 2013)

The majority of the products moved in the warehouse weight a maximum of 10 kg. The conveyor needs to be able to handle up to 90kg at once. Thus, the motor required should

• Reflected acceleration torque of about 200 oz-in
• Torque friction of 100 oz-in
• Reflected breakaway torque of 120 oz-in

1. Speed: (Bodine Electric, 2013)

The conveyor belt should have an adjustable speed, and it is recommended to be set to approximately 60 rpm.

1. weight (Good Year)

For a 2 Ply with a thickness of 5/32”, the pressure should be 0.075 PIW

Figure 5. Recommended Conveyor Belt

Reduce Order-Picking Errors

New order picking-list

One of the main things that help in a successful picking process is a clear and easy picking list. Changing the picking method from single order picking to batch order picking. Many elements in the current picking list that make a list confusing and hard to read.

Redesigning the picking list by removing features that the picker does not need and enhancing the elements that are useful for the selector.

Figure 11. Current Picking List with Unnecessary Elements Highlighted

The unnecessary elements are as follow:

• The barcode of the individual order, since in the new process, the personalized orders will be sorted in the sorting station.
• The date of the order placed and the zone that it will deliver.
• The ID of the product in the warehouse management system.

The improved picking list will restructure the contents of the current picking list and add elements to ease the picking proses. These changes will be as follow:

• Putting the SKU as the first column is essential since it is the first thing the picker will look for on the list.
• The second and third columns will be the Name and Image of the product, so the selector can visually compare the outcome.
• The third column would be QuantityQuantity, so after the visual inspection, the picker would pick the QuantityQuantity needed for the product.
• The last column would be the Checking list so that the selector would check the products selected.

Having these elements in this order will facilitate a more accurate picking process.

Figure 12. Improved Picking List

Radio Frequency Scanner

A wireless handheld device that uses a radio frequency wireless network to communicate with a WMS (Warehouse Management System) to transfer information.

RF scanners are usually used in warehouses to pick products in which the order or picklist will appear on the screen and shows the picket that he needs to choose and where to find the item. The picker then picks the theme, scans it to check if it is the right one, and automatically updates the selected item in the WMS. Once the order fulfilled, the information sent through the RF scanner gets updated to the WMS, and the picker receives his next order to pick.

Benefits

• Reduce paper usage in the warehouse and provides a paperless option, so selectors do not have to waste time handling different paperwork for each order.
• Streamline operations dramatically by reducing manual picking errors
• The automated system will update and confirm orders which will lead to a more accurate order fulfillment
• Live results presented in the screen and system will cause Less room for error
• Provides more precise inventory control and status updates since picking information are in real-time
• Continuously monitoring stock, providing business with real-time stock data.
• Support multiple order picking methodologies such as single, batch, or wave picking.
• Allows for workforce productivity reporting and benchmarking.

PROJECT CONSTRAINTS

The obstacles we have faced during the project are as follow; first, the warehouse layout has been built with a fixed narrow 1-meter space between rack aisles to allow for maximum storage capacity by setting up high-density storage to utilize space. This layout, however, imposes significant operating constraints for picking options. Second, Language diversity is one of the biggest obstacles in the warehouse in which most of the workers cannot read or speak neither English nor Arabic, which are the two main languages used in the warehouse for either communicating with managers on using the WMS. Third, The Nicene warehouse is relatively new, and with the growing market demand, their picking process regularly changes, therefore we have minimal historical data on the current process.

REFERENCES

Wheeler, C., 2020. Picking Error Problems – Suggestions And Solutions. [online] Newcastlesys.com. Available at: <https://www.newcastlesys.com/blog/bid/349401/picking-error-problems-suggestions-and-solutions> [Accessed 1 March 2020].

Romaine, E., Ciervo, B., Lodwig, D., and Gionet, B. (2020). 12 Ways To Improve Order Accuracy Of Your Operation | Conveyco. [online] Conveyco. Available at: <https://www.conveyco.com/improve-order-accuracy/> [Accessed 1 March 2020].

Six-Sigma-Material.com. 2020. Mistake Proofing. [online] Available at: <https://www.six-sigma-material.com/Mistake-Proofing.html> [Accessed 3 March 2020].

ŠKERLIČ, Sebastjan & Muha, Robert. (2017). Reducing errors in the company’s warehouse process. Transport Problems. 12. 83-92. 10.20858/tp.2017.12.1.8.

Cisco-eagle.com. 2020. 10 Ways To Reduce Warehouse Inventory Errors Due To Order Picking | Cisco-Eagle. [online] Available at: <https://www.cisco-eagle.com/blog/2012/06/25/10-ways-to-reduce-inventory-errors-due-to-order-picking/> [Accessed 16 April 2020].

A Guidebook For The Evaluation Of Project Delivery Methods.,2009. Washington, D.C: Transportation Research Board, pp.78-86.

OSHA. “Safeguarding Equipment and Protecting Employees from Amputations” 2007

[online] Available at: <https://www.osha.gov/Publications/osha3170.pdf> [Accessed 3 April 2020].

United States Department of Labor. n.d.  “Baggage Make-Up Room.”

[online] Available at: <https://www.osha.gov/SLTC/etools/baggagehandling/baggage_makeup.html> [Accessed 3 April 2020].

“Working in A Standing Position – Working in Standing/Sitting Positions: OSH Answers.” Ccohs.Ca, 2020,

[online] Available at: <https://www.ccohs.ca/oshanswers/ergonomics/standing/sit_stand.html>.  [Accessed 3 April, 2020].

Helander, Martin, and Martin Helander. A Guide to Human Factors and Ergonomics. CRC  Taylor & Francis, 2006.

[online] Available at: <https://dl.epdf.pub/download/a-guide-to-human-factors-and-ergonomics.html?hash=1feec7582054d24ff4b625dbf4b12e66&captcha=b46dd14db41406f5ff3d08c8cfe6ac40> [Accessed 3 April 2020].

Good year rubber products.com. n.d. BELTING PRODUCTS AND SERVICES CATALOG [online] Available at <https://www.goodyearrubberproducts.com/files/Apache/ApacheBeltCatalog/ApacheBeltCatalog.pdf> [Accessed 9 April 2020]

Bodine Electric Resources. ( 2013). How To Select And Size Gearmotors For Conveyor Applications. [online] Available at: <https://www.bodine-electric.com/core/files/bodineelectric/resources/bodine-gearmotors-sizing-gearmotors-for-conveyor-apps-07470035.pdf>.

Bodine Electric Gearmotor Blog. (2013). How To Select And Size Gearmotors For Conveyor Applications.  [online] Available at: <https://gearmotorblog.wordpress.com/2013/11/20/gearmotors-for-conveyors/>[Accessed 9 April 2020].

Materials, Conveyor et al. “Conveyor Belt Material Manufacturers    Suppliers.” Conveyorbelting.Net, 2020

[online] Available at: <https://www.conveyorbelting.net/conveyor-belt-materials/>.  [Accessed 3 April, 2020].

Monk Conveyors. n.d.  “Standard Belt Conveyors.” 

[online] Available at:  <https://www.monk-conveyors.com/pdf/Downloads/belt%20conveyors.pdf>. [Accessed 3 April, 2020].

> [Accessed 1 March 2020].