Optimization of the cycle time to

increase productivity at Ruashi

Mining
I.N.M. Nday1* and H. Thomas1
*Paper written on project work carried out in partial fulfilment of
BSc.Eng (Mining Engineering) degree
Affiliation:
1University of the Witwatersrand,

South Africa.
Synopsis
The optimization of the cycle time to increase productivity at Ruashi Mining was an on-site project
investigation conducted from December 2017 to January 2018. The main objective of the investigation
Correspondence to: was to determine the actual cycle time components. These comprise the queuing time, loading time,
I.N.M. Nday hauling time, and dumping time of ADTs (articulated dump trucks). Data was also collected on two
hydraulic shovels to identify where these could be optimized. The actual cycle time obtained was
compared to the theoretical cycle time to ascertain the constraints, which affect productivity. By applying
Email: Systems Thinking to the cycle time, man-made constraints of the actual cycle time were identified and
[email protected] recommendations to mitigate them suggested to optimize the cycle time. However, rainfall is a natural
constraint that adversely affects the actual cycle time, and which cannot be mitigated.
In addition, both the theoretical and actual cycle time were analysed, as well as the theoretical and
Dates:
Received: 6 Feb. 2019 actual productivity for the year 2017. Based on the findings, conclusions and recommendations were
Revised: 5 Apr. 2019 made to optimize the cycle time.
Accepted: 16 Apr. 2019
Published: July 2019 Keywords
opencast mining, cycle time, optimization, productivity.

How to cite:
Nday, I.N.M. and Thomas, H.
Optimization of the cycle time to
increase productivity at Ruashi
Mining.
The Southern African Insitute of Introduction
Mining and Metallurgy The aim of the project was to increase Ruashi Mining’s productivity by 39%, from about 7.537 Mt of
DOI ID: broken material mined to 12..367 Mt targeted in 2017.
http://dx.doi.org/10.17159/2411-
Background
9717/624/2019
Ruashi Mining is a copper (main commodity), cobalt (by-product), and acid producer located in the
ORCiD ID:
Democratic Republic of Congo (DRC), 10 km from the city of Lubumbashi. Figure 1 shows the location
I.N.M. Nday
https://orchid.org/0000-0003- of the Ruashi Mining open pit.
1802-3609 Ruashi Mining started production in 1911 and in 2005 the mine was registered as a DRC
company owned by Metorex Group Limited. In November 2013, the Chinese company Jinchua Group
International Resources Co. Ltd. bought a majority stake in the business. Ruashi Mining currently
produces 38 000 t of copper and 4400 t of cobalt per annum. The company has contracted NB Mining
(a contract mining company) to transport its ore. To optimize the productivity of the mine, considering
ore flow delivery at Ruashi Mining, the cycle time is the indicator used in this project. The cycle time
components were recorded on site using a stopwatch.
Problem statement
There are many ways in which the productivity of a mine can be optimized, one way being by
decreasing the actual cycle time. At Ruashi Mining there are constraints affecting and extending the
cycle as shown in Figure 2. The constraints need to be identified and eliminated or mitigated to optimize
the cycle time, which will increase productivity in terms of the tonnage of broken rock transported per
annum.

Literature survey
To structure this project, a Master’s project on the optimization of the haulage cycle model for open pit
mining using a simulator and a context-based alert system was reviewed.
By utilizing mining software, simulations of an actual system in a process can be generated.
Parameters of the simulation can be modified several times with the aim of looking for the ultimate
optimization model. In this way it is possible to save costs, resources, and time.

The Journal of the Southern African Institute of Mining and Metallurgy VOLUME 119 JULY 2019 631  ◀

July Journal.indb 631 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining

Figure 1—Location of Ruashi Mining in the southern part of the DRC (Google Earth, 2018)

utilization. LPM minimizes the number of ADTs required for the

shovel utilization without ADTs queuing and is equivalent to
maximizing the overall production rate.
Lane (2018) stated that a fundamental understanding of the
mining system is vital to maximizing operational output. Systems
Thinking is a management discipline used to comprehend
Figure 2—Excess minutes in the cycle time per excavator and ADT type different components constituting a system and to examine how
these components link and interact to contribute to the overall
Furthermore, an As-Is and a To-Be model were developed system efficiency.
to represent the haulage route of an open pit mining operation The mining system comprises different activities that interlink
so that the cycle time can be optimized. Values generated with and when the optimum performance of each activity is obtained,
the As-Is model were subsequently adjusted in the To-Be. the targeted throughput can be produced. An underperforminig
This model could be used for other mines, taking into account activity compromises the total throughput. Therefore, it is
their parameters such as the production, geology, and haulage imperative to thoroughly understand the mining system to
equipment. identify a constraining activity in that system and optimize it to
The main goal of the project was to develop a methodology improve the mining output. It is worth noting that, optimizing a
to improve productivity, reduce unproductive time, increase truck non-constraint activity will not have as great an impact on the
and shovel utilization, and reduce the queuing time for shovels overall throughput as optimizing a constraint ( Goldratt, n.d.).
and the crusher by applying simulation techniques, without The Theory of Constraints is a methodology for identifying
incurring any additional investment. the most important constraint that compromises the mine
In addition, the As-Is model allowed for the representation throughput and then systematically improving that constraint
of a mine operating at a particular time, and the To-Be model until it is no longer the limiting factor. Sets of tools provided by
showed how the operation would look after certain adjustments. the Theory of Constraints to maximize the mine throughput are:
(Vasquez Coronado, 2014). 1. T he five focusing steps (used to identify and eliminate
Ercelebi and Bascetin (2009) developed a model based on constraints) as shown in Figure 3.
closed queuing network theory that can be utilized to determine 2. The thinking processes (used to analyse and resolve
the optimal number of trucks working with each shovel in the problems).
system. In addition, using a linear programming model they
These are designed with cause-and-effect tools. Firstly,
determined how the ADTs should be dispatched to the shovels.
identify the cause of undesirable effects, and secondly eliminate
The case study was conducted at the Orhaneli open pit coal mine
them without creating new ones. Thinking processes answers
in Turkey. Ercelebi and Bascetin stated that productivity of the
three questions:
equipment is an important factor in the profitability and that
shovel-truck systems are the most common in open pit mining. ➤ What needs to be optimized?
The aim of developing a model was to maximize productivity ➤ What should it be optimized to?
and hence increase production, which in turn will result in cost ➤ What actions will lead to the optimization?
reduction and increased profitability. 3. T hroughput accounting (a method for measuring
At Ruashi Mining ADTs do not arrive at the shovel at performance and guiding management decisions)
consistent times, nor does it take the same amount of time for (Goldratt, n.d.).
shovels to load trucks. Therefore, the randomness of arrival times Catoca Mining Society (CMS) is an opencast diamond
of ADTs at the shovel results in queuing time or the shovel being mine which was assessed with the purpose of productivity
idle while waiting for ADTs to arrive. optimization. The production rates at the mine were neither
A linear programming model (LPM) assumes no ADTs constant nor did they follow the designed or planned output. As
queuing under ideal conditions with maximum shovel a strategy to overcome the issue, the mine increased the unit

▶  632 JULY 2019 VOLUME 119 The Journal of the Southern African Institute of Mining and Metallurgy

July Journal.indb 632 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining
Mining, under NB Mining. The purpose was to investigate the
mining operations, to identify constraints and to collect cycle time
data to determine the total actual cycle time for optimization to be
realized.

Interpretation of data
The interpretation of the data (analysis and results) included
statistical calculations to find the average loading, hauling, and
manoeuvring time (dumping and queuing times). The actual
cycle time was then compared to the theoretical cycle time and the
excess times were determined for each cycle time component.
Figure 3—Five focusing steps to identify and eliminate constraints
Systems Thinking applied to optimize the cycle time
Approach to optimize the cycle time
numbers for both excavators and haulage to operate at maximum
efficiency. The operational conditions such as the weather, By definition, Systems Thinking is a management discipline used
breakdowns, and maintenance were to be considered because to comprehend different components constituting a system and
these conditions could disrupt the hauling and loading operation. examine how these components link and interact to contribute to
Achieving shovel productivity is dependent on having the overall system efficiency (Lane, 2018). Figure 4 shows the
adequate ADT coverage to minimize shovel delays. In cycle time system with individual activities.
addition, the availability of the shovel and ADT system is a The cycle time system consists of different activities; namely,
crucial parameter to consider, as it dictates the utilization and the loading time, travelling time with ADTs loaded, queuing time
subsequently the output (productivity) of the system. at the run-of-mine (ROM) pad, dumping time at the ROM pad
The parameters affecting the loading/haulage cycle were stockpiles, hauling time with ADTs unloaded, and queuing time
identified and simulated accordingly to determine the optimum at the loading point. For an optimal cycle time to be achieved,
operation cycle. Furthermore, the results have shown that for the abovementioned activities must be performing efficiently.
a distance of 2.7 km, six ADTs should be used for optimal Therefore, to optimize the cycle time, the constraining activity
efficiency (Vemba, 2004). must be identified and mitigated.
Mkhatshwa (2009) investigated the Mamatwan opencast
manganese mine in South Africa, which was failing to reach its Constraint identified in the cycle time system
daily production target. This was due to factors such as road By considering the cycle time as a system, and knowing that
conditions, and poor matching of ADTs and shovels. Mkhatshwa every system has a constraint, it is imperative to identify the
further stated that properly designed haul roads result in constraint(s) in the cycle time system to be able to optimize that
minimum traffic congestion and thus increase the production system.
efficiency of the ADTs. From the on-site investigation and data analysis, the
Mkhatshwa also stated that by correctly matching the following constraints were identified:
equipment, production can be improved by minimzsing the ➤ Work ethics of the operators (operators stopping the ADTs
loading time and obtaining 95% loading of ADTs. In practice, it is to discuss trivial matters, e.g. sport, for a short period of
impossible to achieve a 100% load for the ADTs. time)
➤ Mine road conditions
On-site investigation ➤ Condition of equipment
The on-site investigation included the time spent at the Ruashi ➤ Long queuing time of the ADTs at the shovel.

Figure 4—The cycle time system with individual activities

The Journal of the Southern African Institute of Mining and Metallurgy VOLUME 119 JULY 2019 633  ◀

July Journal.indb 633 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining
To optimize the cycle time, these activities must be mitigated, Influence of rainfall
starting with the most important and proceeding to the least
The southern part of the DRC has a tropical climate, with
important. Furthermore, to optimize the present cycle time, there
the rains from mid-November to the beginning of April.
is a need to be aware of the theoretical cycle time, which would
Consequently, open pit mining operations are affected negatively
be compared to the actual cycle time.
during this period.
On-site data analysis ➤ ADT hauling speed is decreased
The data collected on-site was analysed and used for the ➤ Mud accumulayes on the mine roads and in the open pit
calculation of the actual cycle time to be optimized. The analysis ➤ During heavy rain, the operations stop due to safety
included: concerns and access issues. The Ruashi open pit contains
➤ Queuing time talc, which becomes slippery when wet. Consequently, it
➤ Hauling time is dangerous to operate mining equipment under such
➤ Dumping time conditions.
➤ Production data. However, Ruashi Mining continues to operate during this
The production data collected revealed a failure to reach the period because of ore stockpiled at the ROM pad. Figure 6 shows
daily production target as well as the monthly target of excavated the monthly production for the year 2017–2018, highlighting the
material (ore and waste) to be delivered to different destinations. effect of rainfall on production.
To add value to Ruashi Mining, there must be consistent ore The monthly production achieved (orange trendline) is below
delivery to the ROM pad. the targeted production (blue trendline) from mid-November
Figure 5 shows Ruashi Mining’s monthly production for the 2017 to the beginning of April 2018.
year 2017. The targeted production was met or surpassed in only
4 out of the 12 months. As a result, the 2017 Ruashi Mining Theoretical cycle time calculation
annual production target was not achieved. The average cycle time is obtained as follows:

Figure 5—Monthly production versus target at Ruashi Mining

Figure 6—Effect of the the rainfall season on monthly production at Ruashi Mining

▶  634 JULY 2019 VOLUME 119 The Journal of the Southern African Institute of Mining and Metallurgy

July Journal.indb 634 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining
➤ Total resistance force = Grade + Rolling resistance =
[1] 21.63% (approx. 22%).
Figure 7 depicts the rim pull chart used to determine the
speed of the BELL B40D ADTs. The overall B40D weight was
Theoretical queuing time results
66 851 kg, which is used on the left-hand side graph. An arrow
When the cycle time is ideal, there are no constraints intercepts the 22% total resistance force and another arrow
compromising an operation. By reviewing the paper by Ercelebi intercepts the gear in red on the right-hand side graph. A speed
and Bascetin (2009), it could be concluded that in an ideal cycle of 7 km/h is determined from the graph, so that the total hauling
time there is no queuing of the ADTs at the shovel. time (going and returning) could be determined:
Theoretical loading time results
The loading time is the sum of the times (in minutes and
seconds) for the number of passes. For this project, the loading
times were recorded for the two Hitachi shovels, namely the
Hitachi EX 1200-6 and Hitachi EX 870 LCR-3. The hauling time that the B40D should take to travel from the
Struck capacity of B40D 18.5 m3 open pit to the ROM pad and return to the pit is 25.11 minutes.
Struck capacity of B50D 22 m3 Figure 8 depicts the rim pull chart for the B50D; the grade and
EX 870 LCR-3 capacity 5.5 m3 the rolling resistance remain constant as for the B40D.
EX 1200-6 capacity 6.2 m3 The B50D speed was determined to be 6.5 km/h from the
B50D rim pull chart, as seen in Figure 8. The overall B50D weight
Equation 2. Loading time was 79 920 kg. The blue arrow is used to illustrate how the speed
was determined.
[2]

➤ EX 1200-6 = 28.3 seconds = 0.4717 minutes

➤ EX 870 LCR-3 = 29 seconds = 0.4833 minutes.
Theoretical hauling time results The hauling time that the B50D should take to travel from
the open pit to the ROM pad and return to the open-pit is 27.05
[3] minutes.

Theoretical dumping time results

➤ ADTs are required to travel on a flat surface at 40 km/h
➤ ADTs are required to travel up the ramp of 10° (grade = The dumping time at the ROM pad includes the raising and
10°; tan (10°) × 100% = 17.63%) at a speed of 12 km/h, lowering of the bucket.
and down the ramp at a speed of 20 km/h ➤ Dumping time B40D: 13.0 + 7.6 = 20.6 seconds = [4a]
➤ Assume the rolling resistance (based on the road 0.3433 minutes
conditions observed from the on-site investigation) to ➤ Dumping time B50D: 11.2 + 9.9 = 21.1 seconds =
[4b]
be 4% 0.3517 minutes.

Figure 7—Rim pull chart for the B40D

The Journal of the Southern African Institute of Mining and Metallurgy VOLUME 119 JULY 2019 635  ◀

July Journal.indb 635 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining

Figure 8—Rim pull chart for the B50D

Table I
Theoretical cycle time results
Average time Average time
40 t 50 t 40 t 50 t
Queuing time (min) 0 0 Queuing time (min) 0 0
Loading time (min) EX 870 LCR-3 0.4833 0.4833 Loading time (min) EX 1200 0.4717 0.4717
Hauling time (min) 25.11 27.05 Hauling time (min) 25.11 27.05
Dumping time (min) 0.3433 0.3517 Dumping time (min) 0.3433 0.3517
Average theoretical cycle time (min) 25.937 27.885 Average theoretical cycle time (min) 25.925 27.873

Table II
Actual cycle times for different shovels and payloads
Average time Average time
40 t 50 t 40 t 50 t
Queuing time (min) 03:36.0 03:36.0 Queuing time (min) 03:36.0 03:36.0
Loading time (min) EX 870 LCR-3 02:01.0 02:12.2 Loading time (min) EX 1200 01:43.6 02:00.5
Hauling time (min) 21:59.4 25:59.5 Hauling time (min) 21:59.4 25:59.5
Dumping time (min) 00:25.2 00:24.3 Dumping time (min) 00:25.2 00:24.3
Average theoretical cycle time (min) 28:01.6 32:11.7 Average theoretical cycle time (min) 27:44.2 32:00.0

Theoretical cycle time results It is worth noting that the loading time data was collected for
two different shovels, and that affects the actual cycle times.
The theoretical cycle time is the ideal cycle time. Table I shows
the theoretical cycle time results after calculating different
Theoretical cycle time versus actual cycle time analysis
components making up the total theoretical cycle time.
Different shovels give different loading times, depending on The main differences between the theoretical and actual cycle
their capacities when loading ADTs of different payloads. This times are in the input assumptions.
will result in different theoretical cycle times, as can be seen in The purpose of calculating the theoretical (ideal) cycle time
Table I. was to determine the benchmark to optimize the actual cycle
time. In addition, the theoretical cycle time was calculated with
Actual cycle time calculations the data obtained from the manufacturer’s specification booklets,
The actual cycle time was calculated from the data collected whereas the actual cycle time was calculated with data collected
on site using a stopwatch. A summary of the actual cycle time on site. Figure 9 shows the chart comparing the actual and
components is shown in Table II. theoretical cycle time of a Hitachi EX 1200-6 shovel.

▶  636 JULY 2019 VOLUME 119 The Journal of the Southern African Institute of Mining and Metallurgy

July Journal.indb 636 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining

Figure 9—Actual versus theoretical cycle time for the EX 1200-6

Figure 10—Actual versus theoretical cycle time for the EX 870 LCR-3

As expected, the actual cycle time is longer than the (planned and achieved) in 2017. By knowing the planned tons
theoretical cycle time. For the B40D, the difference in cycle time production per ADT and the number of trucks operating on the
is 1.52 minutes and for the B50D the difference is 4.13 minutes. mine, the calculations could be done as follows:
By mitigating the constraints identified, the cycle time will be
[5]
optimized, and the actual cycle time will be reduced to a time
equal to or better than the theoretical cycle time. Figure 10 shows
the actual versus theoretical cycle time for a Hitachi shovel EX [6]
870 LCR-3.
From Figure 10, with a Hitachi 870 LCR-3, the difference in
the cycle time of the B40D is 2.08 minutes and that for the B50D
[7]
is 4.24 minutes. There is also a need to optimize this cycle time
as the gap (largely caused by the queuing time) between the
theoretical and the actual cycle time is significant. ADTs and shovels have surpassed the required working
hours, which has become a constraint as identified during the on-
Production calculation site investigation. Consequently, unpredictable breakdowns occur,
which affect the machine availability and utilization as shown in
Production data interpretation
Table III.
The production data to be analysed was collected on the mine. As a result, there was a 39.06% shortfall in production in
Shift data: 2017–2018, with 4 830 403.91 t of unmoved material being the
➤ Two shifts per day outcome of the longer cycle time due to the constraints.
➤ Eleven hours per shift (ideal scenario)
➤ Nine effective working hours per shift, i.e. 18 effective
Conclusions
working hours per day. There are two types of constraints identified in the production
mining cycle. There are man-made constraints identified through
Planned versus actual production per annum Systems Thinking analysis (i.e. the mine roads and condition
Equations [5] to [7] are used to calculate the annual production of equipment, lack of work ethic from the operators, bucket fill

The Journal of the Southern African Institute of Mining and Metallurgy VOLUME 119 JULY 2019 637  ◀

July Journal.indb 637 2019/07/31 2:34 PM

 Optimization of the cycle time to increase productivity at Ruashi Mining

Table III
Comparative production calculation results (predicted actuals) for each truck type matched with each shovel

Annual planned production Annual actual production

EX 870 LCR-3 EX 1200-6 EX 870 LCR-3 EX 1200-6
B40D B50D B40D B50D B40D B50D B40D B50D
Payload (t) 40 50 40 50 40 50 40 50
Bucket fill factor (FF) 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9
Availability (Av) 0.8 0.8 0.8 0.8 0.75 0.75 0.75 0.75
Utilization (Uti) 0.7 0.7 0.7 0.7 0.6 0.6 0.6 0.6
Cycle time (min) 25.94 27.89 27.44 27.87 28.02 32.12 25.93 32
Cycles per hour 2.31 2.15 2.19 2.15 2.14 1.87 2.31 1.88
Production (t/h) 46.63 54.21 44.08 54.25 34.69 37.83 37.49 37.97
Available number of ADTs 11 5 11 5 11 5 11 5
Production of all ADTs 512.94 271.06 484.90 271.26 381.58 189.13 412.34 189.84
Effective shift hours per day 22 18
Tons produced per day 11 284.63 5 963.43 10 667.76 5 967.71 6 868.52 3 404.42 7 422.14 3 417.19
Number of days per annum 365 357
Tons produced per annum 4 118 888.51 2 176 651.13 3 893 730.61 2 178 213.13 2 452 062.53 1 215 378.27 2 649 702.74 1 219 935.94
Total tons produced 12 367 483.39 7 537 079.47
Production shortfall (t) 4 830 403.91
Difference (%) 39.06

factor of the shovel, and the queuing of ADTs at the shovel). data must have been caused by the shovel bucket filling
There is also a naturally occurring constraint, which is the rainfall factor. NB Mining must reinforce its operator-training
from mid-November to the beginning of April. This constraint programme so that the operators have the competency to
cannot be mitigated. operate the shovels at full efficiency
From the analysis of the results, the actual cycle time of ➤ The following formula can be used to calculate the number
a B40D ADT can be reduced by 8% (from 28.02 minutes to of trucks required per shovel. This will reduce the queuing
25.94 minutes) when used with the Hitachi 870 LCR-3 shovel. time of the ADTs at the shovel. For the purpose of this
Moreover, the actual cycle time for the B50D ADT can be reduced calculation, the actual hauling and loading cycle times are
by 13.2% (from 32.12 minutes to 27.89 minutes) using the used because they are based on the data collected on-site.
Hitachi 870 LCR-3. When excavating with the Hitachi 1200-6
shovel, the actual cycle time can be reduced by 5.5% (from 27.44
minutes to 25.94 minutes) for the B40D ADT and by 12.9%
(from 32 minutes to 27.87 minutes) for the B50D ADT. [8]
Finally, a reduction in cycle time increases production as
it can achieve a higher tonnage per hour and per shift. The
production can be increased with the same resources and fixed Future studies can be conducted on:
costs by optimizing the productivity through Systems Thinking.
Eventually, the increase in the financial performance of the mine ➤ Application of Systems Thinking in mining to optimize a
may extend the life of mine (LOM) as it will allow access to mine’s throughput
previously uneconomical reserves. Ultimately, the positive impact ➤ Adding value to a mine by saving costs on diesel when
on the country will be tangible through royalties, taxes, and the implementing ’stop-start systems’ on ADTs.
economy.
References
Recommendations Goldratt, E. Not dated. Theory of Constraints – Lean Manufacturing. https://www.
leanproduction.com/theory-of-constraints.html [accessed 29 March 2018].
➤ Scheduled mine roads and equipment maintenance. Mine Vasquez Coronado, P.P. 2014. Optimization of the haulage cycle model for open pit
roads must be regularly maintained, especially during mining using a discrete-event simulator and a context-based alert system. The
the rainy season, to allow smooth running of the mine University of Arizona. http://hdl.handle.net/10150/321594
operations Ercelebi, S. and Bascetin, A. 2009. Optimization of shovel-truck system for surface
➤ Empowering the workers would be an effective way to mining. Journal of the Southern African Institute of Mining and Metallurgy,
vol. 109. pp. 434–437.
overcome their poor work ethic, by demonstrating the
Lane, G. 2018. Systems Thinking for mining. Industrial and Research Seminar II.
impact their work has on the overall mine operation and
University of the Witwatersrand, Johannesburg.
output. Proving to workers how losing minutes discussing
Mkhatshwa, S. 2009. Optimization of the loading and hauling fleet at Mamatwan
trivial matters adversely affects the daily, monthly, and open pit mine. Journal of the Southern African Institute of Mining and
yearly cycle time, therefore, the productivity (Lane, 2018) Metallurgy, vol. 109. pp. 225–226.
➤ It was observed from the on-site investigation and proven Vemba, M. 2004. Loading and transport system at SMC – Optimization. Journal of the
by data analysis that the inconsistency in the loading time South African Institute of Mining and Metallurgy, vol. 104. pp. 141–146. u

▶  638 JULY 2019 VOLUME 119 The Journal of the Southern African Institute of Mining and Metallurgy

July Journal.indb 638 2019/07/31 2:34 PM