Development process

In the proposal document, the team had decided to develop a robotic arm that would take the product from the manufacturing line to the packaging line.  After refining the requirements with the customer there was need for additional systems to be incorporated into this design.  Referring to the decision matrix below in Table 1, you can see that the conveyor system scored the highest, meaning it is the most sensible solution for the requirements and needs of the customer.

    Table 1.  Statistical Analysis Results

 

A robotic arm could not accomplish all aspects that the customer was looking for in the design solution.  For instance, a robotic arm would not have the capacities to solve the inspection, rejection, stacking and sorting issues which the customer found to be of utmost importance.  Using the decision matrix to evaluate the robotic arm and the conveyor system ideas against criteria of importance to the client, it was realized that the conveyor system would be the best solution for our application needs.  With the help of the decision matrix, the team decided to scrap the idea of a robotic arm and has decided to go with a conveyor system which will include various subsystems.  These subsystems are to include:

• Inspection
  • Detect defects in cones such as cracks, holes, missing material and absence of a cone
• Rejection
  • Discard nonconforming products into existing scrap system
• Counting
  • Detect presence or absence of cones for stacking purposes
• Downtime Storage
  • Capable of handling two hours worth of product in the event of a packaging line malfunction

 

Statistical Analyses

The time between the release of each batch of cones from the oven once each batch touched the conveyor was recorded thirty times for the analysis.  Due to the increased human error in approximating the time at which the cones reached a fixed point, the time between batches in the above manner was preferred.  This method yielded better results and produced less human error.

Figure 1.  Calculation of Standard Deviation

Since the sample size was greater or equal to 30, a z-score was used in obtaining the results.  For the calculations, please refer to Figure 1 above for further clarification on how the results were obtained. 

The calculations resulted in a mean batch time of approximately 2.34 seconds with a standard deviation of 0.1180 seconds and a variance of 0.0139 seconds².  Using a confidence interval of 99%, the mean was expected to fall between approximately 2.28 seconds and 2.39 seconds.  These results are summarized in Table 2.

Table 2.  Statistical Analysis Results

Possible errors in the statistical analyses may involve: human error with attaining the times (visualizing the cones), human response time when using the start and stop on the stopwatch, accuracy of the stopwatch, consistency of the machinery and also the trimmer.  However, all of these things have minimal impact on the results.

For calculation purposes, the time it would take for a stack of six cones to accumulate was needed in order to generate the flow rate for the production of the stacks of cones.  This resulted in a flow rate of 1,512 stacks of cones produced per hour, assuming perfect conditions and no rejected product.  The flow rate calculations can be further explored below in Figure 2. 

Figure 2.  Flow Rate Calculations

Joy Cone occasionally experiences downtimes of up to two hours, for which they would like to have an overflow storage system for the stacks of cones.  It was determined that an overflow system would need to store about 2000 feet of stacked cones for the given time period.  If spacing is to be involved between each cone stack, this value could begin to exceed nearly 4000 feet.  For further details into the calculation of the storage capacity for the overflow storage system, please refer to Figure 3 below.

Figure 3.  Overflow Storage System Capacity Calculations

 

Pro Model Analyses

ProModel is a robust simulation-based decision making tool that is mostly used in the manufacturing industry.  This program helps the user to simulate a manufacturing environment for decision making purposes.  The program can be used to improve performance for various different manufacturing applications.  Statistical analyses of the model can be produced from the ProModel simulations.  Our team used ProModel as a simulation environment for the purposes of visualizing and verifying our final design. 

Figure 4.  Joy Cone Layout for ProModel Simulations

Figure 4 above represents the Joy Cone layout as modeled in the ProModel program.  Each subsystem is labeled as either C or D, depending on the oven location.  Since the information of most interest was in the simulation and the data collected through this program, we felt that it was unnecessary to represent the correct icons for the locations. 

Table 3.  ProModel Simulation Results

Name	Scheduled Time (MIN)	Total Entries (# of Entities)
To Packing	120	3521
Oven C	120	3597
Counter C	120	3529
To scan C	120	3597
Scanning C	120	3596
To Count C	120	3530
Reject C	120	66
Oven D	120	3602
To Scan D	120	3602
To Reject D	120	71
To Count D	120	3529
Merge Conveyor	120	1764
Scanning D	120	3601
Counting D	120	3528
Overflow Buffer	120	3517
To Shipping	120	2148
Packing	120	2140

The ProModel simulation was run for a time period of two hours in order to collect data on the number of entities the accumulation buffer location saw during that time period.  According to Table 3 above, the statistical results output by ProModel showed that the accumulation buffer saw 3517 stacks of six cones during a two hour time frame.  This result was used to help choose the size that the actual accumulation buffer would need to be for our system.  We chose to use the ProModel data because we felt that having a large buffer size would help account for any uncertainty in the production line.  As a result, it was found that an accumulation buffer size of 3600 stacks of six cones was sufficient for a storage buffer capacity.  Since two spiral conveyors are projected to be used, it can be concluded that each spiral conveyor should store 1800 cones.

 Since our design was conceptual, there was no physical testing performed.