Predictive maintenance accuracy

An analysis is only as good as the data therefore, the equipment used to collect the data is critical and determines the success or failure of a predictive maintenance or reliability improvement program. The accuracy as well as proper use and mounting determines whether valid data are collected. 

Another method used by some plants to acquire data is hand-held transducers. This approach is not recommended if it is possible to use any other method. Hand-held transducers do not provide the accuracy and repeatability required to gain maximum benefit from a predictive maintenance program. If this technique must be used, extreme care should be exercised to ensure that the same location, orientation, and compressive load are used for every measurement. Illustrates a hand-held device. 

This limitation prohibits the use of most microprocessor-based vibration analyzers for complex machinery or machines with variable speeds. Single-channel data-acquisition technology assumes the vibration profile generated by a machine-train remains constant throughout the data-acquisition process. This is generally true in applications where machine speed remains relatively constant (i.e., within 5 to fO rpm). In this case, its use does not severely limit diagnostic accuracy and can be effectively used in a predictive-maintenance program. 

Accuracy Decisions on machine-train or plant system condition will be made based on the data acquired and reported by the predictive maintenance system. It must be accurate and repeatable. Errors can be input by the microprocessor and software as well as the operators. The accuracy of commercially available predictive maintenance system varies. While most will provide at least minimum acceptable accuracy, some are well below the acceptable level. 

It will be extremely difficult for the typical plant user to determine the level of accuracy of the various instruments that are available for predictive maintenance. Vendor literature and salesmen will assure the potential user that their system is the best, most accurate, etc. The best way to separate fact from fiction is a comparison of the various systems in your plant. Most vendors will provide a system on consignment for periods up to thirty days. This will provide sufficient time for your staff to evaluate each of the potential systems before purchase. 

The data logger or microprocessor selected by your predictive maintenance program is critical to the success of the program. There is a wide variety of systems on the market that range from handheld overall value meters to advanced analyzers that can provide an almost unlimited amount of data. The key selection parameters for a data acquisition instrument should include the expertise required to operate, accuracy of data, type of data, and manpower required to meet the program demands. 

It must also be noted that the type of maintenance program adopted in the facihty may have a significant impact on the frequency of many tasks. For example, a program that includes predictive maintenance information may reduce the frequency of tasks requiring hardware checks for wear or accuracy. Likewise, a good reliability-centered maintenance (RCM) program (see Chap. 2.2) will reduce the frequency of most tasks while increasing the frequency of a lesser number of tasks performed on mission-critical equipment. 

As noted earlier, an accurate selection of sources requiring control can best be achieved through the application of the Air Quality Display Model for each AQCR of interest. It is estimated that the total one-time cost of such an exercise at a level of accuracy necessary to predict source sulfur dioxide control requirements for each of the 245 AQCR s would be about 10 million with an additional annual maintenance cost of 500 thousand to allow for emission inventory changes. For comparison. 

The accuracy of any model to predict VOC concentrations in test chamber experiments mostly depends on the accuracy of the source (term ER in Eq. 1) and sink submodels (terms A and D in Eq. 1) incorporated into the lAQ model. In other words, a realistic estimate of human exposure to VOCs emitted from indoor materials and products requires knowledge not only of their emission rates but also of their adsorption/ desorption capacity or the buffer effect on VOC concentrations in indoor air. Small environmental test chambers are increasingly used in order to characterize the emission of VOCs (i.e. the source term in Eq. 1) from materials and products present indoors, whether they are used to realize the building environment, maintenance work (includ-... 

The stability of proteins refers to the maintenance of a defined three-dimensional structure with specific thermodynamic and functional properties. High-resolution structures in the crystalline state and in solution have reached a stage at which the atomic coordinates of proteins can be compared with an accuracy down to root mean square deviation (r.m.s.d.) values less than 1 A. However, even this precision does not allow the fi ee energy of stabilization to be calculated from the coordinates, nor does it allow predictions with respect to the dynamics of functionally relevant local interactions in active or regulatory sites of homologous proteins. The fluctuations between preferred conformations of native proteins involved in such functionally important motions may very well show amplitudes and angles of up to 50 A and 20°, respectively. ... 

In almost all design situations, air leakage rates cannot be accurately predicted. The only real accuracy necessary is that extremes be avoided. The designer is in reality specifying a maintenance level for the system when he estimates air leakage. All vacuum systems will have a finite leakage which... 

It is therefore important to realise that standard tests are not devised to give direct prediction of end-use performance. In general, the reverse is the case, in that if a particular grade of a particular polymer is found to perform satisfactorily in a given end-use, it can then be characterised with reasonable accuracy by standard tests, and the latter tests then used to ensure the maintenance of the reqnired end-nse qnality. 

Random failures occur, as the name suggests, randomly and are the result of degradation mechanisms within the system. Often evaluated by means of failure rates (e.g. failures per hour of operation) or due to physical causes involving a range of mechanisms (e.g. lighting or problems during manufacture, installation or maintenance). Generally it is possible to quantitatively predict, with reasonable accuracy, failure rates for this type of failure. 

---