- 10 October 2007 -
Specification mistakes and wrongly supplied equipment
Specifying and procuring pumps or their spare parts may seem to be an easy job but appearances can be deceptive. Sometimes those in charge are not as painstaking as they should be and make simple mistakes, leading to useless spare parts or unnecessary equipment being acquired, as mechanical engineers Rafael Ocampo and Braudilio Ruiz show.
During the course of its useful life, every pump will need the replacement of various worn and broken parts, and it is the duty of the user to procure the right spare parts in order to make the life of the pump as long and reliable as possible. However, those in charge of spares procurement do not always have all the necessary knowledge and skills to specify correctly every manufacturing detail of the required part. The result is an inadequate material or treatment, an incorrect dimension, or a part that will not fit properly with existing parts.
Besides, due to procurement pro-cedures, sometimes the spare parts are obtained from a different supplier from the original manufacturer, and this supplier, of course, is not likely to have available every design and manufacturing detail necessary to manufacture the spare parts exactly as required. In this case, the person in procurement will have to ensure the availability of all the information needed. Also, although luckily this is not very frequent, the supplier can be careless in the process of designing an assembly of several pieces of equipment from different manufacturers, resulting in a defective set.
In the following case studies, several examples will be discussed to show how such mistakes can influence costs and affect pump availability.
Wrong alignment and rotation direction
A replacement fire water pump was required and ordered from a supplier who was not the manufacturer of all the equipment involved. A complete set of internal combustion engine, speed reducer and pump was offered by the supplier, fulfilling the requested pump characteristics. Plant engineers were satisfied with the pump features and therefore the set was ordered and supplied as a package, all installed on a steel bedplate and supposedly tested at the manufacturer's workshop.
After installation and during pre-commissioning checks, a terrible misalignment was discovered between the speed reducer and the pump. Angular misalignment was visible even to the naked eye and radial misalignment was greater than one millimetre. Although a flexible zigzag spring coupling was used, vibrations during the test runs should have been large enough to attract the attention of the engineers in charge of the test, if this really was carried out. Of course, this misalignment was corrected immediately.
During commissioning, the engine and the close-coupled speed reducer were run uncoupled from the pump and achieved the specified operating parameters. However, when running the complete set, the pump failed to deliver the required flow and discharge pressure. A detailed revision of the system was made and an unthinkable mistake was found. The rotation direction of the engine coincided with the mark on its body and also matched the rotation direction of the pump as marked on the pump casing. Nevertheless, the supplier did not realize that the speed reducer installed in between was a one-stage geared speed reducer so its outlet rotation direction was the reverse of that of the inlet. Therefore, the pump was running in the wrong rotation direction and hence could not reach the design parameters.
To install a replacement pump matching the direction of rotation of the existing engine and speed reducer would have involved changing the large suction and discharge pipes. It was agreed that the delivery of a new complete set would be paid for by the supplier who was responsible for the mistake in designing the assembly.
Inadequate motor characteristics
Four new screw pumps were required to replace old, existing fuel-oil pumps. They were ordered from the supplier of the original pumps but with the request that the new pumps should be able to handle a more viscous fluid. The supply of complete sets of motor pumps was agreed. Installation and pre-commissioning checks proceeded as usual and no difficulties were encountered. However, during the commissioning test each one of the four pumps failed to fulfil its duty. None of them was even able to start rotating.
It was confirmed that these positive displacement pumps of the rotary three-screw type were designed and manufactured for the required discharge pressure, capacity and efficiency so the pumps themselves ought not to have had any trouble reaching the required process parameters and the power consumption should have been as calculated. Why then did they not run?
The next step was to check the energy supply and the electric motors' characteristics. Energy parameters were checked and confirmed to be as required. Also, when the four motors were run without a load they reached the no-load current and rated speed stated by the manufacturer and the direction of rotation was counterclockwise, as requested.
The motors' rated power, number of poles, voltage, frequency, rated torque, rated current and power factor stated on the nameplate were as required but, still, the loaded motors did not run. Then, a very carefully check of the motors' design data was carried out and it was found that starting torque had been designed as 100% of the rated torque, the same as for the original pumps. But the supplier did not take into account that the more viscous fluid would impose a very different duty on these pumps and, although the power required to run the pumps would be the same, the starting torque required would be higher. For this reason, the motors were unable to break the initial inertia of the set increased by the higher viscosity of the new fuel.
Four new electric motors had to be ordered and paid for by the supplier since they were responsible for the mistake in specifying the motor characteristics. The rated torque of the new motors remained the same but the starting torque was specified as 190% of the rated torque; they operated successfully.
Spare impellers and shaft sleeves
Spare impellers and shaft sleeves were needed for six high-pressure, nine-stage, feed water pumps in two identical twin power plants.
Impellers
The original impellers were high carbon 12-14% chromium stainless steel castings with wear surfaces hardened to increase wear resistance. Prices for spare impellers from the original manufacturer of the pump were very high and a substitute cheaper supplier was found. Procurement personnel did not do their homework very well and failed to fully specify the impellers' features, stating only 'stainless steel casting' as the construction material to use. After delivery, manufacturing specifications and other documents from the supplier remained with the procurement department and never reached any technical department on site.
During the assembly process, a routine dimensional check was carried out and the impellers fitted very well with the 13% chrome steel wear rings whose diameter was maintained as originally designed. Cold clearances were checked and found to be around the design figures. The pumps were assembled and put back into operation.
During the next maintenance check, severe seizing was found on the wear surfaces of the impellers and wear rings wherever a new impeller had been fitted but not on the old impellers that had remained installed. A full dimensional check was performed on each new impeller and they were found to be exactly within the tolerance range for every measurement. Nevertheless, the colour difference between the old and new impellers after operation attracted the attention of the engineers and they considered a possible difference in construction material. A simple magnetic test showed that the impellers were not made of magnetic material and the engineers contacted the procurement department to check the documents. There the cause of the problem was found.
The supplier had not been terribly painstaking either and, starting from the insufficient specification of 'stainless steel casting', decided to use AISI 304 stainless steel to manufacture the impellers, although maintaining the geometry and dimensions as requested.
The thermal expansion of austenitic steels is about 1.5 times greater than that of ferritic or martensitic stainless steels so, during operation at 180°C, the thermal expansion of the new impellers was larger than the calculated expansion for 13% chrome steel and the clearance between impellers and wear rings disappeared, causing the damage.
Thermal expansion calculations for the new material were made and a new diameter for the wear surface of the impellers was determined. Wear rings on all the new impellers were machined to the required dimensions and no further problems occurred.
Shaft sleeves
The original specification for the shaft sleeves stated 12-14% chrome steel, hard chromium plated. Again procurement personnel did not do their homework very well, failing to fully specify the shaft sleeves' features, stating only 12-14% chrome steel and forgetting about the surface hardening treatment. As with the impellers, a routine dimensional check was made of the shaft sleeves during the assembly process and they fitted very well.
During the next maintenance check, however, severe seizing was found on the sleeves' wear surfaces. On subsequent investigation of the worn parts and the documents from the manufacturer, it was found that the hard chromium plating had never been specified nor manufactured. Therefore, the surface of the shaft sleeves was soft enough to be easily damaged by the sliding friction generated by movement against the stainless steel rings of the hydraulic shaft seals.
New shaft sleeves with the proper hard surfacing treatment had to be ordered.
Conclusions
It is true that every human being can make a mistake but it is also true that you should think twice and speak once. It could be added that you should think thrice and write once.
An engineer should evaluate his or her personal responsibility when designing or specifying parts, equipment or systems because mistakes can represent big losses in terms of equipment costs and plant shutdown time and, in some cases, could even endanger lives. This is especially true for plant engineers who, ultimately, are responsible for the plant running and sometimes are also beneficiaries of the profits. But, apart from the economic issues, there are ethical reasons, the sense of shame and the possibilities of advancement in the engineer's career to be taken into account. So, every effort is worthwhile in order to avoid engineering mistakes.
The discussed examples show very clearly that those engineers in design, specification or procurement roles can make serious mistakes that have an impact on equipment and plant availability and cost.
If you are designing spares, equipment or systems, check more than once that you are using the proper calculation methods and the right data and that the final result matches what is intended. The mistakes discussed in the first and second case studies would have been avoided with only a little more checking.
If you are specifying spares, equipment or system components, check every detail more than once against what is required. If existing equipment has been modified, even with slight changes, original spares probably will not fit. The mistakes discussed in the third case study would have been avoided by being only a little more careful while processing the information to be used in the procurement.
If you are procuring spares, equipment or system components, make sure you have at your disposal all the correct and required information to carry out the inquiry; establish good communications with the final user, pass the information from the manufacturer to the plant engineers and obtain their approval for every detail before signing the contract. If the supplier is going to be different from the original one, a very detailed revision of the specification will be necessary and every manufacturing detail should be agreed before proceeding.
All these efforts during the preparation stages are fully repaid at the production stage.
Contact
Rafael Ocampo
E-mail: rocampod@gmail.com
Pump Industry Analyst
Sealing Technology
World Pumps © Copyright 2008, Elsevier Ltd, All rights reserved.