·Home ·Table of Contents ·Industrial Plants and Structures

Development and Implementation of Reliability centered Maintenance using Vibration Analysis: Experiences at Rourkela Steel Plant S.Parida, N.R.Kotu, M.M.Prasad Research & Development Centre for Iron and Steel, Steel Authority of India Limited (SAIL), Rourkela Centre, Rourkela 769 011, India S.Vanam, V.S.Vishwanath CMMS/CBMS Department, Rourkela Steel Plant, Steel Authority of India Limited (SAIL), Rourkela - 769 011, India Contact

ABSTRACT

In Steel industry, maintenance cost accounts for 10 - 15% of production cost. Maintenance affects the target, quality and profitability of the plant. Implementation of modern concept of Condition Based Maintenance can appreciably reduce maintenance costs and enhance reliability of machine performance and quality of the output.

In many steel mills worldwide, failures of critical equipment have frequently led to loss of production, high maintenance and operation costs. As the capital cost of modern machines is very high, the requirement of continuous and prolonged operation becomes inevitable, to sustain economic operation. There has been a trend towards adoption of various forms of Condition Based Maintenance Systems which contribute to better health of machines, reduced maintenance costs, efficient use of personnel and improved system efficiencies.

When a large number of machines are involved, a majority being rotating equipment, Vibration Analysis suits ideally the requirement of Condition Monitoring, inspite of relatively higher initial costs in terms of instruments. It is easy to record and transfer data, ensure accuracy of data for trend analysis and fault diagnosis. Further, its versatility and portability permit effective planning and implementation of monitoring schedules.

Rourkela Steel Plant (RSP) has a healthy maintenance practice right from its inception in 1959. Condition Based Monitoring through Vibration Analysis has been developed and is in use at RSP, since April '93, as a key technique for implementation of Predictive Maintenance Program. The coverage has grown over three folds from 40 to 140 critical equipment in the last three years. In addition, for the first time in SAIL plants, single plane in-situ balancing was introduced with tangible benefits in terms of preventing unplanned stoppages, man-hours saved on dismantling and re-assembly of equipment, and savings on cost of components. The number of major breakdowns prevented by timely diagnosis of faults has increased from a mere 5 to 70 cases, resulting in substantial financial savings.

The paper describes the efforts made during recent years, by Rourkela Steel Plant, Steel Authority of India Limited (SAIL), in close association with Research & Development Centre for Iron and Steel (RDCIS), SAIL, to implement Vibration Analysis and related methods as effective Non-Destructive Testing tools to implement Total Productive Maintenance Program in an integrated Iron & Steel works. The paper further discusses a few typical case studies briefly.

INTRODUCTION

A conventional integrated steel plant like RSP, with a capacity of 2.0 Mt ingot steel per annum, has a vast array of equipment. The plant is a conglomeration of smaller units, each in itself complete and self-contained. These constitute Coke Ovens, Coal Chemicals, Sinter Plants, Blast Furnaces, Steel Melting Shops, Continuous Casting, Tonnage Oxygen Plants, Plate Mill, Hot Strip Mill, Cold Rolling Mill, other secondary mills, Captive Power Plants and a host of other departments.

Every piece of equipment needs special care and attention, characteristic to it. The Coal Chemicals unit has Gas Boosters and Exhausters that handle coke oven gas, a highly inflammable commodity, whereas Sinter Plant Blowers and Waste Gas fans handle air containing highly abrasive sinter dust. The Turbo Alternators of Captive Power Plant require round the clock vigilance involving a variety of parameters. Seemingly innocuous Forced Draft & Induced Draft fans of the Reheating Furnaces also assume significance because of their criticality in application. Failure of these fans may lead to cut down of Hot Strip Mill/ Plate Mill production by 33 - 50 percent.

BACKGROUND

Under the current business environment, cost competitiveness of steelmakers has assumed a priority role. As a global phenomenon, effective maintenance management has been accepted as the key to corporate strategy for reduced costs. This has led to integration of maintenance management function with production and business problems, not just equipment problems. With this realisation that maintenance management can cost 35 - 40 percent of revenues and, in most cases upto 15 percent of unit production cost, steel companies are increasingly opting for new maintenance technologies which can be effectively and relatively easily implemented for reduced costs and increased profitability. According to industry estimates, a 10 percent reduction in maintenance costs translates to a 30 percent increase in profitability.

To maintain an efficiently operating unit and avoid failure of critical equipment, especially modern steel industry equipment, the focus has clearly shifted over the years, from Breakdown Maintenance, i.e. repairing the equipment after it malfunctions, to Preventive Maintenance, i.e. fixing the equipment according to planned maintenance schedule. The next trend was Computerised Maintenance Management Systems (CMMS), and the latest trend encompasses asset management and maintenance, supported by various methods of Condition Based Maintenance Systems (CBMS) and in-service inspection processes. CBMS or Predictive Maintenance methods are an extension of preventive maintenance and have been proved to minimise the cost of maintenance, improve operational safety and reduce the frequency and severity of in-service machine failures. The basic theory of condition monitoring is to know the deteriorating condition of a machine component, well in advance of a breakdown, for proactive maintenance.

There are many machinery parameters that can be measured, trended and analysed to detect imminent failure or onset of problems. Common among them are :

1. Machinery vibration
2. Lube oil analysis including wear debris analysis
3. Infrared thermography
4. Ultrasonic testing
5. Motor current analysis
6. Shock pulse measurement, etc.

Additionally, operational characteristics such as flow rates, heat, pressure, tension, speed and so on can also be monitored to detect problems. In case of machine tools, product quality in terms of surface finish or dimensional tolerances is often an indication of problems. As all these techniques have value and merit, the application of any particular technique depends on the suitability and ease of implementation.

SELECTION OF VIBRATION ANALYSIS AS MOST REWARDING TOOL

Among the listed machine performance characteristics, the one characteristic that is common to practically all machines is vibration. Even machines in the best of operating condition will have some vibration, which may be regarded as normal or inherent. Whenever machinery vibration increases beyond safe limits, the usual reason is some mechanical trouble - unbalance, misalignment, worn gears or bearings, looseness, etc. The vibration data collected from site or logged on-line, are analysed with the help of dedicated software to indicate probable faults in machine components. Corrective actions are then taken to avoid failure.

Vibration is a powerful information source for estimation of technical condition of equipment's, more so in the case of rotating equipment. Considering the vast array of equipment in a steel mill, majority being rotating machinery with either oil or grease lubrication systems or using anti-friction or journal bearings, vibration analysis ideally fits the requirement of predictive condition monitoring. Although the start-up costs are relatively higher due to investments in instruments, dedicated software and skill development of maintenance personnel, the accruing benefits offset these hindrances during implementation. The advantages include ease of operation for recording and transferring data, accuracy in handling data for diagnosis and applicability to a wide range of plant operations. The portability and versatility of instruments permits effective planning and scheduling of monitoring activities and if needed, planning of shutdowns for maintenance. According to survey data of American Society of Mechanical Engineers (ASME), upto 82 percent of malfunctions of the mechanical equipment can be detected with the help of vibro-monitoring and vibro-diagnostic methods.

SYSTEM DEVELOPMENT AND IMPLEMENTATION

Rourkela Steel Plant, started in 1959 with German collaboration, has a healthy maintenance practice right from the inception. Over the past four decades of steel plant operation, RSP, a unit of SAIL, has a systematic approach towards the maintenance of equipment and facilities. It has been long journey from the time when German technical experts were available to guide and help in designing a suitable system and methods to the present decade when SAIL, RSP engineers on their own developed the concept of Computerised Maintenance Management System (CMMS). The procedures being practiced now are not only well documented but also time tested. Consequently, among all SAIL units, RSP could win the coveted UNDP/ UNIDO financial assistance amounting to US $ 1 million for computerisation of maintenance system during the year 1987-89. RSP has developed its own "in-house" software with joint efforts of maintenance and Computers & Information Technology personnel. The package has four modules, with necessary sub-modules and functions.

Fig 1: Flow chart for implementation of PMP based on vibration analysis

After successful completion and implementation of UNDP aided CMMS project, Phase-II of the same project was envisaged with CBMS as an important component. It started as an extension of CMMS in April 1993, and is nearly seven years old. In fact, vibration monitoring is a key technique through which the Predictive Maintenance Programme (PMP) is being done at Rourkela Steel Plant (Fig. 1). There are as many as 100 departments, who are maintaining various machines under them. After an initial survey and detailed discussion with all the departments, 40 critical machines were chosen and at present more than 140 critical machines are being monitored under this programme.

STEPS IN IMPLEMENTING PREDICTIVE MAINTENANCE
The logical steps in implementing a PMP are

• Detection,
• Analysis or Fault Diagnosis, and
• Correction.

Detection involves measuring and trending vibration levels at marked locations on each machine included in the programme on a regularly scheduled basis. The objective is to reveal significant increases in a machine's vibration level to warn of developing problems. Analysis helps to pinpoint specific machinery problems by revealing their unique vibration characteristics. Corrective action is taken after specific problem has been detected and identified by planning and scheduling all activities to ensure that machine downtime is kept to the absolute minimum (Fig. 1).

Critical machines of steel plant like Gas booster, Exhauster, Turbo-Alternators, etc. are continuously monitored by use of permanently mounted vibration transducers. By current standards, breakdown maintenance has been totally eliminated, and in no case it is more than 10% of the maintenance actions.

CASE STUDY I - GCP ID FAN # 1, SMS-II

In a process industry like an integrated steel plant, unscheduled breakdown of critical machines in the stream can cause total production loss and other subsequential losses. One such critical machine was Gas Cleaning Plant (GCP), Induced Draft Fan # 1 of SMS-II (Fig. 2). This machine used to be stopped frequently due to very high vibration levels and many times for unforeseen breakdowns. In the year 1998, there was a total forced outage of 25 days attributed to this Fan.

Fig 2: Schematic setup of GCP ID Fan #1, SMS-II

The normal assumption of failure was unbalance due to deposits on the impeller. This led to high vibration amplitudes (Table I and Fig. 3) and consequently resulted in breakdowns. After every break down, the machine was inspected to identify the primary damages as well as secondary damages. The cost involved for such failures were on account of,

• Cost of lost production,
• Cost of maintenance including replacement of damaged parts,
• Cost of emergency spares procurement,
• Cost involved by compromising quality standard while doing emergency repairs,
• Cost of additional overheads, etc.

Table 1: Machine vibration & spike energy readings before (17th May '99) and after (21st May '99) repairs

Fig 3: Machine health trend report

Towards the end of 1998, a condition based maintenance programme was implemented in the shop. In the first phase, the fan was stopped before the vibration velocity exceeded 8.00 mm/s and sequential inspection and cleaning were taken up. It was found that almost at an interval of 40 days, the impeller was taken up for cleaning which resulted in breakdowns, leading to stoppage of nearly 14 hours. By this method, uncertain pattern of breakdown was totally eliminated.

In the second phase of the implementation, critical review of the behavioral pattern of the entire system was taken up as part of Dynamic Predictive Maintenance. A problem diagnostic study revealed that apart from the influence of unbalance caused by the form of deposits, there are also other major inaccuracies in the system. Detailed frequency analysis, phase analysis, coastline down pattern and orbit analysis was taken up on this equipment to have a very precise diagnosis of the system behaviour.

The following major defects were identified to be present in the equipment -


• Inadequate rigidity of the fan pedestals,
• Resonance condition,
• Hot alignment,
• Skew in the bearing (resulting from thermal growth), and
• symptoms of axial thrust.
After introduction of necessary corrective measures, the vibration amplitudes have dropped sharply below safe working limits (Table I and Fig. 3).

CASE STUDY II - WASTE GAS FAN, SINTER PLANT II

The Waste Gas Fan in Sinter Plant II (Fig. 4) was a frequent cause for concern. This fan sucks and exhausts air laden with the sinter dust into the atmosphere through a chimney. Due to the abrasive nature of the medium handled, the rotor vanes are subjected to non-uniform wear resulting in impeller unbalance.

Fig 4: Schematic setup of WG Fan #1, SP-II

The vibration readings taken before repair showed operation beyond alarm level (Table II and Fig. 5). Subsequently,
detailed spectrum analysis revealed


• Misalignment between fan and motor bearings,
• Deteriorating condition of fan non-drive end bearing,
• Oil condition has to be investigated, and
• Fan is unbalanced, and it needs dynamic balancing.

After opening the bearing housings, it was observed that the sleeve bearing on fan non-drive end was burnt and fan drive end bearing condition had also deteriorated. Two sets of white metal sleeve bearings were replaced. After boxing up of fan side bearings, proper alignment was done between motor and fan. The fan was then rotated and found to be dynamically unbalanced. IRD 885 Data Analyser/ Balancer was used to do in-situ balancing and the post repair vibration levels were well within the alarm level (Table II and Fig. 5). This sort of selective repair action is possible only because of vibration monitoring and accurate assessment of source of trouble.

Table 2: Vibration readings before repair (22nd October '98) and after repair & in-situ balancing (23rd October '98)

Fig 5 : Machine health trend report

Table 3: Vibration readings after in-situ balancing (3rd July '99)

Trending of vibration amplitude data at scheduled intervals indicated normal operation of the fan till 30^th June '99. On this day, the fan had tripped twice and diagnosis by CBMS techniques indicated that the fan impeller was unbalanced. It was dynamically balanced at 1500 rpm, and the vibration level reduced from 14.53mm/s to 2.94mm/s. The vibration recorded just after in-situ balancing of the fan on 3^rd July '99 is shown at Table III. Since then, the machine is running smoothly and uninterruptedly.

RESULTS

Fig 6:Growth through CBMS activities

With condition monitoring activities rightfully in place at RSP, there has been substantial growth in all aspects encompassing the maintenance system
(Fig. 6). The on-going process of acquiring newer instruments, software and exposure of personnel to advanced techniques has greatly helped in increasing the coverage from 40 to over 140 critical equipment in the last three years. There has been significant increase in the number of major breakdown prevention cases from 10 to 102, which is more than 10 times, during the period 1994-95 to 1999-2000. After the introduction of in-situ balancing, the savings on account of total maintenance cost have been encouraging. The savings have grown by 650 percent in 1999-2000 from an amount of Rs.20.3 million in 1994-95. More so, the trend line of savings, plotted with a fourth order polynomial, indicates rapid jump through CBMS activities in future. This amply justifies that the return on investment in vibration instruments & software, alongwith skill enhancement of maintenance engineers is highly attractive.

Fig 7: Equipment health condition at RSP

As a corollary, there has been vast improvement in the health of equipment
(Fig. 7) available for production processes in accordance with ISO 2372. There has been an evident shift in the percentage of equipment from category F towards category A, with a majority of equipment under category B and C. This signifies a direct correlation with improvement in reliability of equipment, since CBMS activities incorporate a reliability-based philosophy, such that it concentrates on mean time to symptom, and not failure.

CONCLUSIONS

The knowledge of machine vibration signatures and their time dependent behaviour are the basis of efficient condition monitoring of rotating machines. Combined with other methods as well as stand-alone system, it can be used to perform a multiplicity of inspection activities. It successfully diagnoses and locates faulty operation conditions at an early stage in order to prevent severe failures and to enable predictive and condition-oriented maintenance. Effective use of CBMS offers benefits in plant availability, optimised use of resources, reduced production costs, reduced downtime and as an offshoot, enhanced data collection and monitoring requirements leading to improved documentation. The implementation of CBMS has led to paradigm shift in emphasis from isolated equipment to the complete process of production. Vibration monitoring and analysis is ideally suited for such programs as Total Productive Maintenance (TPM) in which they may be used by both maintenance and production personnel.

ACKNOWLEDGEMENTS

The authors express their sincere thanks to all colleagues for their support and guidance during preparation of the paper and the management of SAIL, RDCIS and RSP for the kind permission to publish the work.

© AIPnD , created by NDT.net

|Home|    |Top|