Planning the Test

In setting up any performance test, the following principal steps should be taken  

It is of great advantage that the person who will analyze plant data participate in the test program. Many factors should be evaluated during the test to determine the validity of the test data. Evaporator tests can be costly, especially if the data obtained are not sufficient to meet the test objectives. Adequate test personnel must be available to gather all necessary data over a relatively short period. Testing a sextuple-effect evaporator can require several hours for one data set at one operating point. 

Diagnosis of plant data depends heavily upon actual physical inspection of the equipment, checking on conditions not apparent from any panel board or instrument readings. Included are  

The test team must be completely familiar with the evaporator system. Flow sheets and instrument diagrams must be understood. Physical location of measuring devices must be known and evaluated. 

Accuracy of the data required must be established. Errors in the interpretation of results may arise either from errors of measurement or from factors that are not normally subject to measurement. The effect of these errors depends to a large extent on the intended use of the test results. If test data are used to predict performance of the same evaporator under slightly different conditions, most of the normal errors are self-compensating. If test data for one effect of an evaporator are to be used in the design of a new evaporator, even minor errors can be of great importance. Accurate data is necessary when comparing actual peformance to that expected. 

---

To plan the tests correctly and economically requires knowledge of the performance of the material at the start. It is easy to select temperatures which lead to failure in times either too short to satisfy the standard, or too long to wait for. Adding a contingency for this adds considerably to the cost of testing. 

Prior to the scheduled shutdown of the No. 4 Unit, a test on the turbo-generator was planned. The test was intended to confirm the operation of a modification made to the generator control circuits. This would enable it to continue to provide power during its... 

The next box is 9.2, covering safety validation planning, which tells us to plan the tests and records needed to demonstrate that the equipment will meet the SRS. 

Programs to develop MCFC technology are also under way in Europe. Ansaldo SpA (Italy) is setting up faciUties to produce 1-m cells in an automated process, and their goal is to test 100-kW stacks in 1994. The 100-kW stack is also to be tested by IBERDROLA in Spain as part of a complete power plant system. Two Dutch companies. Stork and Royal Schelde, have joined with the Dutch government to form Brandstofcel Nederland (BCN), which plans to test a 50-kW MCFC and two 250-kW MCFC stacks in 1994. 

In test planning, the number to be placed on test n and the number of failures rmust be deterrnined. The operating characteristic curves in Reference 18 can be used to specify the test, and to control the errors. 

Success Testing. Acceptance life tests ate sometimes planned with no failures allowed. This gives the smallest sample size necessary to demonstrate a rehabiUty at a given confidence level The rehabiUty is demonstrated relative to the test employed and the testing period. 

As an example in estimating wear rate in a nylon bushing consider a 10-rnm diameter shaft miming 900 rpm (0.47 m/s) under 0.5 x 10 N/m (70 psi) load. The P/ of 0.235 X 10 N/m -m/s(6510 psi-fpm) and if = 0.24 x 10 m /N for filled nylon in Table 9 gives a wear rate of 0.20 mm/1000 h. Since Pp test results vary widely, these wear estimates are only guides. For maximum utiHty, the test materials, finishes, temperature, load, speed, and lubrication should dupHcate as nearly as possible those in the planned appHcation. 

If the experiment is conducted in stages, precautions must be taken to ensure that possible differences between the stages do not invaUdate the results. Appropriate procedures to compare the stages must be included, both in the test plan and in the statistical analysis. For example, some standard test conditions, known as controls, may be included in each stage of the experiment. 

The hterature abounds with instructions as to the proper manner in which to accomplish the necessaiy and MIC-safe testing procedures. The engineering personnel planning these test operations should avail themselves of that knowledge. 

Duration of Test Although the duration of any test will be determined by the nature and purpose of the test, an excellent procedure for evaluating the effect or time on corrosion of the metal and also on the corrosiveness of the environment in laboratory tests has been presented by Wachter and Treseder [Chem. Eng. Pi og., 315-326 (June 1947)]. This technique is called the planned-interval test. Other procedures that require the removal of sohd corrosion products between exposure periods will not measure accurately the normal changes of corrosion with time. 

Intent Plant personnel, supplies, and budget are reqiiired to successfully complete a unit test. Piping modifications, sample collection, altered operating conditions, and operation during the test require advance planning and scheduling. Analysts must ensure that these are accomplished prior to the actual test. Some or all of the following may be necessary for a successful unit test. 

Owing to the greater test uneertainties assoeiated with field testing as eompared to planned shop testing, the warranty and guarantee should be modified to take into aeeount the nonideal test set-up. 

A eommon failure in pressure measurement is the uneertainty of the eonfiguration of statie-pressure taps penetration through the pipe wall. This failure is another early-planning eoneern, sinee proper taps are easy to provide prior to plaeing the maehine in serviee, but inspeetion of the taps after operation has eommeneed is a luxury rarely afforded the test team. 

Part of the planning should include the evaluation of test uncertainty. This evaluation can be limited to a common sense approach based on available instrumentation and the locations relative to the ideal. A more sophisticated study can be made in which instrumentation accuracy and the impact of any inaccuracy on the measured parameters is evaluated. This is a complex task with the need being based on the motivation for the test. If the test is being performed to settle a dispute, a formal understanding of the uncertainty should be developed. Methods for evaluation of test uncertainty are found in ANSI/ASME PTC 19.1 [11]. 

The most difficult part of a field test is the flow meter, if it wasn t planned in the construction phase. There is no way to simulate a meter run if you don t have the proper pipe length. Figure 10-8 is an example of the requirements. An ASME long radius flow nozzle is preferred by the author, though a short throat venturi will do. The probability is that an orifice is all that will be available. It should be examined before and after the test to verify not only the bore diameter, but the finish. The bore should... 

Note that filter aid selection must be based on planned laboratory tests. Guidelines for selection may only be applied in the broadest sense, since there is almost an infinite number of combinations of filter media, filter aids, and suspensions that will produce varying degrees of separation. The hydrodynamics of any filtration process are highly complex filtration is essentially a multiphase system in which interaction takes place between solids from the suspension, filter aid, and filter medium, and a liquid phase. Experiments are mandatory in most operations not only in proper filter aid selection but in defining the method of application. Some general guidelines can be applied to such studies the filter aid must have the minimum hydraulic resistance and provide the desired rate of separation an insufficient amount of filter aid leads to a reduction in filtrate quality — excess amounts result in losses is filtration rate and it is necessary to account for the method of application and characteristics of filter aids. 

The measurement procedure must have been tested previously under realistic conditions and should give representative results of worker exposure. Thus, it should not be influenced by unusual work schedules, where tasks may not be well defined or planned. The procedure should include... 

The design verification plan should be constructed so that every design requirement is verified and the simplest way of confirming this is to produce a verification matrix of requirement against verification methods. You need to cover all the requirements, those that can be verified by test, by inspection, by analysis, by simulation or demonstration, or simply by validation of product records. For those requirements to be verified by test, a test specification will need to be produced. The test specification should specify which characteristics are to be measured in terms of parameters and limits and the conditions under which they are to be measured. 

Test plans to be produced which define the sequence of tests, the responsibilities for their conduct, the location of the tests, and test procedures to be used... 

The test sample to have successfully passed all planned in-process and assembly inspections and tests prior to commencing qualification tests... 

---