Eng. Ondrej Stejskal
Logio s.r.o., Czech Republic
stejskal@logio.cz
Abstract #
This article describes a case study of application of decision model when setting up plan for renewal of assets. This particular case is dedicated to material handling equipment (AGV) while the principle is applicable for any type of asset. The first part of the article shows project situation and premises. The second part describes used methodology for project solution. In the third part the methodology is applied and the project results are presented. The article is finished by conclusion of the project as well as evaluation of usage of the methodology.
Keywords: asset management, asset renewal, replacement cycle, asset lifecycle costing, cascading decision making
1. Situation #
In the plant the company operates a fleet of 418 units of AGV (Automated Guided Vehicle). The fleet consists of groups of different generations of technology and ages between 3 and 12 years. Due to technological development the plant is facing situation where availability of spare parts for older generations is not further guaranteed by suppliers.
Given the problem described – ongoing technological developments and the operational aging of the fleet, it is necessary to set up a systematic approach, starting with preparation of a Plan for conceptual renewal of the fleet. The Plan is result of comparison of possible scenarios for asset replacement in terms of minimal costs and other criteria when applicable.
Number of AGV unites: 418 pcs
Types of AGV: 12
Planning horizon: 10 years
Interest rate: 9%
Considered costs:
- Renewal costs (AGV unit price)
- Flat rate maintenance service (prev. + correct. maintenance, operation)
- Overhauls
- Spare parts
- Re-certification and upgrades of current AGV types
- Purchase of batteries
Before start of any action, it is essential for all project “stakeholders” to understand and confirm “Why” it is done (so called the Purpose or “Higher Goal” of the project). In this case it is Ensuring stable operation of driverless handling equipment that meets the needs of logistics and production processes and Optimization of costs of driverless handling equipment.
In compliance with the project Purposes the detailed project goals are set. The difference between those two is that project goals have to be done within the project scope and outputs has to be delivered in accordance. The project goals are:
Phase I: Collection and Analysis of input data
Phase II: Definition of Replacement Scenarios of the fleet
Phase III: Comparing scenarios and drawing up an optimal Replacement Plan
Figure 1: Illustration of AGV handling equipment (source: SSI SCHAFER)
2. Necessary cooperation within the company #
A vital condition for project success is preparation of input data in a correct extent and structure. The key data requirements are as follows:
- A list of individual FTS units with Unique Identifier (the identifier must also be included in all subsequent data sets in order to assign the AGV unit to the data), including information for each unit:
- Place of operation (hall/route)
- Date of entry into service or, where applicable, date of manufacturing
- Shift regime of operation
- Type of ownership (own, leasing with maintenance, full-service, etc.)
- Mileage km and moto-hours from commissioning to present by months
- History of maintenance performed on individual AGV units with a description, labour consumption, materials and maintenance costs from the time of commissioning to the present
- List of charging stations with the assignment of AGV operating groups
- Consumption of energy, oils/other operating media for individual AGV from the time of commissioning to the present
- List of critical parts for which the support ends incl. term and BOMs
- List of offered used / retrofitted AGV from partners incl. price and technical condition
- Electricity costs for charging at individual charging stations
- Cost of CAPEX + estimation of price indexation and residual value (including costs of acquisition and commissioning, disposal of old equipment, administration, etc.)
- OPEX costs + estimation of cost development for individual AGV units monthly from commissioning to the present
- Other relevant costs (if any)
- Discount rate for calculating the NPV of the future cash flow
3. Project Methodology #
In this chanter the key steps to set up Replacement plan are described. As the project constrains and premises always vary, the solution is always individual according to situation (asset type, extent and quality of input data, key user/decision maker preferences, etc.)
The key attributes to establish Replacement plan are validated replacement criteria and rules. By setting prioritization and sequence of criteria the cascading decision tree is made to create Replacement scenarios. Replacement criteria of the case are shown in Table 1.
Table 1: Replacement criteria and rules
| No. | Criterion | Description |
| 1 | Replacement Model | Optimal replacement cycle – the local minimum function of the total average annual costs. If there is no local minimum, the replacement cycle is 8 years (according to the information from manufacturer). |
| 2 | End of Spare parts (SP) support | The lifespan can be extended (the year of replacement shifted) by cannibalization of other units or by re-certification of missing part.AGV will never be replaced before the end of SP support defined by manufacturer/dealer. |
| 3 | AGV Cannibalization | Principle: the lifespan of a certain group of AGV is extended at the expense of other AGV units, which are put out of service, and from which SP are cannibalized.Cannibalization parameters: on the basis of SP historical consumption, the future consumption of SP for lifetime extension of AGV kept in service is determined. And finally, the number of discarded (renewed) AGV units is calculated and as well as extended life of the remaining AGV. |
| 4 | Re-certification (upgrade) | Some types of AGV can be upgraded – the critical spare part is replaced by newer model and type of AGV is upgraded so that SP unavailability is eliminated. The CAPEX for upgrade and extended lifespan of the particular AGV type is set by manufacturer. |
| 5 | Overhauls | According to the manufacturer, Overhaul is carried out every 4 years of operation of a unit. The second option is implementation of Overhaul based on the difficulty of the route where AGV is operated.The year of replacement is shifted by certain number of years from the last performed Overhaul, which prevents replacement of unit where significant cost to Overhaul was recently invested. |
| 5 | Routes | Some AGV types cannot be replaced individually, the replacement has to be done by the whole group in one time. |
| 6 | Cost of AGV per km/hour of operation | Types of AGV are evaluated in terms of cost depending on mileage and shift regime. AGV with higher costs are then, replaced as a priority within the concept of cannibalization or within the route. |
| 7 | Technology update | The current concept does not consider a change of technology from AGV to AMR, as well as the transition from lead-acid to lithium-ion batteries due to the returnability of investment. |
| 9 | CAPEX limit and distribution | The limit of CAPEX is not set. Requirement of the plant is to distribute CAPEX evenly over the years. |
| 10 | Limit for Overhauls | Yearly limit for number of performed Overhauls and related costs is defined. |
| 11 | Capacity of manufacturer | Maximal production of 500-700 unites of AGV yearly is defined. |
3.1 DATA ANALYSIS
Clearing of data, segmentation of the fleet (age, mileage, costs), investigation of corelation of costs with age/mileage, AGV type, place of operation etc. Structure and quality of the data reflects future constrains and deviations of results. Results should be always presented together with this information.
Figure 2: OPEX development by years
3.2 REPLACEMENT COST MODEL
Searching for cost optimum of the replacement interval. For every single AGV unit the minimum average annual cost of ownership is analysed.
Figure 3: Replacement cost model calculation
3.3 REPLACEMENT SCENARIOS
The defined replacement criteria are reflected in the cascading decision tree. The result is single year of replacement for each unit, not yet containing all the constrains and mainly not yet taking into account the corporate processes for replacing assets (technological, economic and administrative constraints).
Figure 4: Cascading decision tree
3.4 REPLACEMENT PLAN
Replacement plan is factual replacement schedule based on all constrains. Cannibalization and re-certification are also only incorporated at this point for its non-standardized approach that would bring an enormous complexity when used in phase of establishment of Replacement scenarios.
Figure 5: Three steps for determination of Replacement plan
4 Project results #
The final Replacement plan is based on mathematical determination of the replacement year + application of cannibalization of spare parts (longer AGV life-time and optimized CAPEX), optimization of AGV replacement distribution in years (CAPEX optimization), optimized recertifications and upgrades (levelling-out CAPEX for recertification) and optimized number of overhauls in years (levelling-out CAPEX and capacities for overhauls).
Other principles applied within the establishment of the Replacement plan are:
As result of data analysis the cost optimum of 8 years of AGV lifetime given by the manufacturer was verified. AGV with the highest cost per km/h of operation and on the most demanding routes is replaced first.
If the AGV is close to the planned replacement, the overhaul is not performed. If an overhaul has been performed, the lifetime of the AGV is automatically extended by 4 years.
The utilization of maintenance capacity, new AGV production capacity and annual CAPEX spendings are maximized.
Each AGV is kept in operation until the end of SP support. The lifetime of AGV is extended by obtaining SP from decommissioned AGV or by upgrading SP with ending support.
Main Result: Longer average lifetime of AGV unit (+3 years). Thanks to the systematic approach around 10 M EUR less will be paid in total costs in the given time horizon.
Example of graphical outputs of the project are seen on Figures below.
Figure 6: Complex costs analysis and segmentation for single AGV unit
Figure 7: Segmentation of the AGV fleet by age
Figure 8: Example of Cost Replacement model without minimum annual average costs – it does not show progressivity over time, and it is close to linearity. For this reason, the annual average cost curve (red) has no local minimum. The optimum in this case is to replace the AGV unit never and to continue preventive and corrective maintenance until total obsolescence.
Figure 8: Number of replaced units of AGV in years
Figure 8: Distribution of cumulative CAPEX + OPEX in years
Conclusion #
Based on the course of the project, it can be stated with certainty that when managing a large number of assets, such as a fleet, it is hardly imaginable to achieve any optimal solution for asset replacement without a systematic approach and a data-based solution.
On the other hand, the data-based solutions have pitfalls indeed and poor data quality can be critically misleading. Therefore, it is necessary to evaluate how a given shortage of data can contribute to the overall result. The typical data pitfalls of such projects are missing cost history, missing links of costs to individual asset units (group payments and flat rates under framework contracts), work orders missing causes of repairs and used spare parts, data is not consistent – duplicates, missing connection between data sources through unique identifiers. The flat rates and framework contracts can cause misleading information for a cost model that does not show the progressivity of maintenance costs with age and thus the average annual cost function has no local minimum. Based on experience, the data requests must be submitted in advance and in clearly defined structure and extent. The responsible person has to be appointed to hand the data over on time and to be ready for subsequent consultation and explanation of the data as it is never done on the first iteration.
Another challenge may be indexing costs in the future, especially in economically unstable times, when for longer planning horizons, the deviations can cause significant differences in total costs or a change in the order of variants for decision-making.
The methodology of creating a Replacement Plan has proven its robustness in adapting to the conditions and requirements of the individual cases. When there are validated and cleaned data, the Cost model is appropriate base to get understanding of asset replacement cycle. The final realistic Replacement plan is then completed by sequence of applying replacement criteria with given prioritization.
