With the focus on oil analysis as a key performance indicator (KPI) for the health of a piece of equipment, are you sure your oil analysis is pointing you in the right direction? Many factors can affect the results of an oil sample, and one of them is how the sample is taken. Personnel can be trained to draw an oil sample, but do they understand how to collect the sample properly?
Coolant in the Transmission
Consider an incident involving a 40-ton rough-terrain crane located 1,500 miles away. The oil analysis results showed coolant in the transmission. When problems like this were detected, a second sample was taken before any serious work was performed. The second sample confirmed coolant in the transmission. The on-site mechanic gave the crane a thorough examination and detected nothing out of the ordinary that would cause the coolant to be present.
The crane was taken out of operation, and a third sample again showed coolant in the transmission. Because of growing concern, the crane expert was dispatched to the job site to examine the crane. He found nothing that should cause such a problem. In just one week, thousands of dollars were spent and an excessive amount of downtime resulted. The crane expert approached the lubrication technician and watched him pull a sample.
Upon watching the technician pull samples, the crane expert solved the mystery of coolant in the transmission fluid. In his ritual of sampling, the technician first pulled a coolant sample utilizing the drop-tube method. He then wiped the outside of the tubing so it could be reused for the transmission fluid sample, which was pulled second.
It should be noted that this particular lubrication technician was a recent hire, and had received limited instruction on collecting a sample.
When does a change in machine health necessitate observation of the lube tech’s performance? The example above gives us a good idea; however, it is not recommend to waiting until three samples are analyzed. Although the lubrication technician understood the need to pull a sample, there was a breakdown in the communication and understanding of how to do so properly. This incident demonstrates that merely instructing someone on what to do does not guarantee his or her overall understanding.
When implementing an oil analysis program or training new personnel, there is a learning curve and growing pains that accompany either task. Oil analysis results can be utilized for determining not only machine health, but also the competence of the lubrication technician (with further scrutiny).
When implementing a new oil analysis sampling program, utilizing the results for training (as a KPI) is tricky because a history on the equipment does not exist. Over time, however, trending may emerge if samples can be tracked by the person responsible for collecting them. If one lube tech consistently samples higher silicon and aluminum levels on average than other lube techs, this warrants closer examination.
Depending on the ratio and geographical location, silicon and aluminum suggest dirt in the oil, or possibly a poor sampling technique that allows for dirt to ingress at the sample bottle. In critical applications, it is imperative to rule out dirt ingression into the oil reservoir before looking into the lube tech’s oil-sampling technique.
Monitoring the performance of a new lubrication technician is more straightforward. With a history available, both the machine health and the sampling procedure can be verified if a sudden spike occurs in the oil analysis results. Ideally, these two conditions should happen simultaneously, but the type of spike and criticality of the equipment in question would dictate the specific action. A spike in the particle count would cause concern regarding the sampling technique, whereas a spike in water would indicate a machine issue. If the problem is found with sample collection, additional training should be made available to the lube tech.
There are various ways to evaluate the oil analysis data for comparison purposes depending on the situation and the end goal. Such approaches include the following:
Result trends for equipment sampled by multiple people
Comparison between historical data and new data collected by a new lube tech
Random occurrences of contamination specific to a particular lube tech
Overall comparison of discrepancies for samples collected by a single lube tech
This list provides a starting point. Establishing a guideline is difficult because no two sites are the same, but trends and statistics can be calculated to meet the needs of any site. The end goal is to utilize the information to determine how to enhance the oil analysis program as a whole. POA