1. Collect basic information related to the equipment: Criticality Analysis, Failure Reports, Operational Context, ACR and Maintenance Costs.
2. Analyze information collected.
example-of-methodology-definition-policies-maintenance-spare-parts-for-an-oil-pumping-system3. Rank criticality of the equipment corresponding to the systems with the greatest operational impact (for the example, these are the Turbo pumps for the oil pumping service). In this case, the information taken from the Operational Context indicated the following:
Census:
2 SOLAR Turbo pumps of 7,500 HP each (# 1 and # 6) (Subsystem A).
8 RUSTON Turbo Pumps of 5,400 HP each (# 2, # 3, # 4, # 5, # 7, # 8, # 9 and # 10) (Subsystem B).
•
Operational philosophy: Continuous and parallel operation of 1 SOLAR +2 RUSTON.
The following passive redundancies are then defined:
• 100% for the SOLAR Subsystem.
• 300% for the RUSTON Subsystem.
From the records of the equipment, information was obtained regarding: Hours of Operation and Number of Stoppages, indicated in the following table:
Given the information indicated in the table, the criticality (preliminary) of each of the equipment is estimated, considering:
Criticality (CR) = Failure Frequency Value (linked to shutdowns) * Impact
The impact for both Subsystems was considered with the minimum value = 1, due to the high redundancy.
Failure Frequency = Number of Accumulated Shutdowns / Total Hours of Operation
The most critical turbo pumps are:
• Sub system A: TB # 1.
• Sub system B: TB # 10.
4. Define the location of each equipment in its Lifecycle. For this, the Coefficient of Variation (CV) is used:
Coefficient of Variation (CV) = Deviation of Times Between Failures (Dtef) / Average Time Between Failures (TPEF)
The TEF corresponding to the most recent 18-month period were taken. considering it a more suitable period to estimate the current operating condition.
To statistically process the TEF data, the SPSS program was used, obtaining the following results:
The graphs will later support the estimation of the PF intervals, as a reference for the frequency of predictive monitoring.
The CV makes it possible to measure the degree of dispersion associated with the TEFs of each TB and to compare them. On the other hand, it indicates the possible location of the equipment in its life cycle, under the following criteria:
CV> 1 Equipment in start-up phase, CV = 1 Equipment in normal operation phase and CV <1 Equipment in wear phase (or condition chronic failure).
The results of the CV were:
By relating the CV of each TB, with its corresponding Accumulated Hours of Operation, its operating condition can be inferred. The results are indicated in the following table:
Sub system A
Sub system B
5. Define adequate Maintenance Policies, according to the estimated condition:
Sub system A
Sub system B
6. With the information generated, a maintenance plan must be prepared that includes the tasks specifically required by each team. For the example, it is observed that the priority actions would be the corrective maintenance to be carried out on TB # 1, # 6 and # 10 respectively (sequence of priorities considering their capacity and CV in relation to accumulated Hours of Operation). Next, it is appropriate to carry out corrective maintenance to TB # 4, # 7 and # 8. As for TB # 2, it is convenient to plan in the medium term, its repowering or replacement, according to the results of the technical economic analysis correspondent.
For TB # 3, 5 and 9, a basic condition-based maintenance policy is deemed adequate. This implies the following decisions:
• What monitoring to apply?
In this case, it was found that the organization responsible for maintenance has the equipment indicated: Contact temperature and vibration meter (SKF).
- Temperature meter by laser beam. Thermographic camera.
The aforementioned equipment is suitable for monitoring the condition of TB.
• How often?
To estimate the appropriate frequency, a short-term forecast is made, based on the analysis of the TEF distribution, considering as a premise that by adopting an approximation to the normal distribution (Gaussian curve), the following conditions are met:
- TPEF-S <TEF <TPEF + S With 68.3% probability
- TPEF-2 S <TEF <TPEF + 2 S With 95.4% probability
- TPEF-3 S <TEF <TPEF + 3 S With 99.7% probability
The results were:
• TB # 3 Given:
TPEF: 104 hours
S: 53.1 hours Also
analyzing the frequency distribution of TEF, it was considered 130 hours of operation, as an adequate period between predictive monitoring.
• TB # 5
• Given:
• TPEF: 120.37 hours
• S: 48 hours
• Also analyzing the distribution of TEF frequencies, 145 hours of operation was considered as an adequate period between predictive monitoring.
• TB # 9
• Dice:
• TPEF: 107 hours
• S: 61 hours
• In addition to analyzing the TEF frequency distribution, 50 hours of operation was considered as an adequate period between predictive monitoring.
7. The Maintenance Plan will in turn support the Logistics Plan to provide timely spare parts and consumable materials.
8. For the Logistics Plan, you must first define what to buy?, for the corrective maintenance of TB # 1, # 6 and # 10 with higher priority and for TB # 4, # 7 and # 8, with lower priority. To estimate what to buy ?, it is pertinent to carry out an ABC Analysis, considering the VAU and subsequently develop a Criticality Matrix.
9. The Maintenance Plan should clearly and comprehensively define (applying a functional envelope taxonomy), the maintenance actions to be carried out, according to the criticality of the equipment. In the example, it is appropriate to apply:
In relation to TB # 1, 6 and 10, decisions regarding logistics involve:
• What to buy?
• How much to buy?
Its definition will be based on the corrective maintenance plan. A suitable methodology is the following:
Example of spare parts required for Turbo Pumps:
• Hydraulic part:
• Mechanical part (TAG):
To update the prices of the spare parts, it is convenient to apply F = P * (F / P, i, n) = P * (1 + i) ^ n
I = Average inflation per period (annual)
For the example, I = 6% is considered.
In the Compound Interest Table corresponding to i = 6%, the values of the required factors are obtained:
In relation to TB # 4, 7, 8, 3, 5 and 9, it must be defined for the next annual period:
• ¿ What to buy?
• How much to buy?
• When to buy?
These decisions should consider:
• Medium-term requirements, to carry out the corrective maintenance planned in that term.
• Spare parts that should be available in stock (based on the Criticality Matrix).
In this last aspect, a suitable methodology is:
• How much to buy? and When to buy?
Given that demand and replenishment time are highly random, it is convenient to apply in combination, the methods indicated:
- Reorder Point.
The reorder point (PR), it is convenient to link it to the average inventory (IP), which considers the availability of spare parts, when ordering.
IP = (Economic Order Quantity (CEP) / 2) + Z * Deviation (DS)
- Periodic review.
Estimation of what, how much and when to buy ?, in the case of the pumping system:
Preparation of Criticality Matrix (What to buy?):
CRITICITY MATRIX
They must be available in stock. Respond to: What to buy?
Guarantee availability upon request, through agreements with Suppliers.
Process when required. If the technical analysis showed a high degree of difficulty for acquisition, evaluate substitution by equivalent and more commercial components.
Next, the Economic Order Quantity (CEP) is defined for each of the parts I1, II1, III1 and I2
CEP = (2 * Demand in period t * Cost of the order / Cost for inventory management in period t) ^ 0.5
Cost for inventory management for the period of 1 year = (between 0.2 and 0.3) * Unit price of the item Example: Axial bearing
Data:
Demand in the last annual period = 3
Cost of placing the Order: 15,000.00 $ Mex
Unit price: 60,000.00 $ Mex Cost for inventory management = 0.3 * 60,000.00 = 18,000.00 $ Mex
Result:
CEP = (2 * 3 * 15,000 / 18,000) ^ 0.5 = approximate to 2
Finally, the Reorder Point (PR) is estimated, equivalent to the Average Inventory (IP):
IP = (Economic Order Quantity (CEP) / 2) + Z * Deviation (DS)
Z corresponds to the value obtained from the tables of the Standard Normal Curve, for the Service Factor (%) assigned
DS = (Total Replenishment Time * Standard Forecast Error in Demand ^ 2 + Forecast of Monthly Demand ^ 2 * Standard Deviation in Delivery Time) ^ 0.5
Data:
CEP = 2
FS: 95% Z = 1.65
Maximum time for replenishment, as agreed with the supplier = 2 months (this premise allows the Standard Deviation in delivery time is made = 0)
To estimate the Standard Error of Forecast in Demand (ES), the demand for the indicated spare part in the last 6 years was analyzed:
Applying the Excel program, ES = 1.28
Result:
DS = ((2/12) * 1.28 ^ 2) ^ 0.5 = 0.53
IP = (2/2) + 1.65 * 0.53 = approx 2
If the data is not available or if the Service Factor is relatively low (for example considering that there is a high redundancy, assume FS = 0.7), a variant of the Delphi Method (opinion) could be applied to estimate PI and CEP.
CONCLUSION: it is appropriate to place an order for 2 Axial Bearings, when the availability of 2 is reached in the warehouse.
