The present invention relates to a method and system for condition monitoring of mechanical equipment, and more particularly to a method and system for early identification of a deteriorating condition in the equipment.

Background

Condition monitoring of mechanical equipment can be done a number of different ways, both with physical inspection and condition estimation based on measured parameters. For many applications, such condition monitoring is essential in order to identify a deteriorating condition before it leads to an actual fault or breakdown of the system. One such example is offshore drilling rigs, where operational reliability is critical and faults can have significant

consequences, both environmental and in terms of personnel safety. This applies in particular to such equipment which is operated under harsh conditions (e.g. exposed to weather or in contact with substances such as drilling mud) and/or subject to irregular loads or repeated load variations.

Presently, the most common maintenance strategy is calendar based, that is the equipment has to undergo a full maintenance cycle at predefined times.

This is costly, especially if the equipment can not be inspected/serviced on site.

Most of the cost is related to non-productive time caused by the mandatory service. Apart from the potential cost saving from performing maintenance only when needed, or from planning for maintenance at times that are convenient according to the rig's work plans, great savings can also result from condition monitoring and predictions based on the trend of condition indicators if early signs of deterioration can be used to prevent breakdown.

There is therefore a need to further improve known systems and techniques for this. The objective of the present invention is to improve the accuracy and reliability of condition monitoring systems and methods known in the art, and reduce or eliminate at least some disadvantages associated with conventional systems. Summary of the invention

The above and other objectives can be obtained by a method for monitoring the condition of a mechanical system having a first oil lubricated component, the mechanical system being subject to repeated, intermittent load cycles, comprising the steps of:

(a) identifying the start of a load cycle;

(b) measuring a load of the first oil lubricated component;

(c) measuring a temperature of a lubricating oil for the first oil lubricated component;

(d) determining a first system performance parameter for the load cycle, selected from a group of (i) a friction coefficient for the first oil lubricated component, (ii) a temperature increase coefficient for the lubricating oil for the first oil lubricated component, (iii) a temperature rise rate coefficient for the lubricating oil for the first oil lubricated component;

(e) repeating steps (a) to (d) for a plurality of load cycles;

(f) comparing a plurality of determined first system performance parameter values to identify a change in said first system performance parameter over the plurality of load cycles.

Advantageously, by determining performance parameters for individual load cycles and comparing the determined values over a plurality of cycles one can identify a deteriorating trend in the mechanical or lubricating performance of the system.

The first oil lubricated component is preferably a gear or a swivel

Preferably, the step of measuring a load of the first oil lubricated component includes measuring a power transferred through said gear.

Preferably, step of measuring a load of the first oil lubricated component includes measuring a hook load acting on the swivel. In an embodiment whereby the mechanical system further comprises a second oil lubricated component, the method further comprising the steps of

(b2) measuring a load of the second oil lubricated component;

(c2) measuring a temperature of a lubricating oil for the second oil lubricated component;

(d2) determining a second system performance parameter for the load cycle, selected from a group of (i) a friction coefficient for the second oil lubricated component, (ii) a temperature increase coefficient for the lubricating oil for the second oil lubricated component, (iii) a temperature rise rate coefficient for the lubricating oil for the second oil lubricated component;

(e) repeating steps (b2) to (d2) for a plurality of load cycles;

(f) comparing a plurality of determined second system performance parameter values for the plurality of load cycles.

Preferably, the first oil lubricated component is a gear and the second oil lubricated component is a swivel.

Preferably, the step of measuring a load of the first oil lubricated component includes measuring a power transferred through said gear and the step of measuring a load of the second oil lubricated component includes measuring a hook load acting on the swivel.

Preferably, the step of comparing a set of determined performance parameter values for the plurality of load cycles includes comparing a plurality of first system performance parameter values with a plurality of second system performance values.

Preferably, said friction coefficient is a function of a friction work in said first or second oil lubricated component and the load of said first or second oil lubricated component, said temperature increase coefficient is a function of a temperature increase in said first or second oil lubricated component and the load of said first or second oil lubricated component, and said temperature rise rate coefficient is a function of a rate of temperature increase in said first or second oil lubricated component and the load of said first or second oil lubricated component.

Said plurality of load cycles are preferably the first 2-5 cycles following a start of the mechanical system from a cold condition.

The mechanical system is preferably a top drive for a drilling rig.

According to yet another embodiment of the method for monitoring the condition of a mechanical system having a first oil lubricated component and a second oil lubricated component, the mechanical system being subject to repeated, intermittent load cycles, comprising the steps of:

(a) identifying the start of a load cycle;

(b) measuring a temperature of a lubricating oil for the first oil lubricated component;

(c) measuring a temperature of a lubricating oil for the second oil lubricated component;

(d) determining a first system performance parameter for the load cycle, selected from a group of (i) a first friction coefficient for the first oil lubricated component, (ii) a first temperature increase coefficient for the lubricating oil for the first oil lubricated component, and (iii) a first temperature rise rate coefficient for the lubricating oil for the first oil lubricated component;

(e) determining a second system performance parameter for the load cycle, selected from a group of (i) a second friction coefficient for the second oil lubricated component, (ii) a second temperature increase coefficient for the lubricating oil for the second oil lubricated component, and (iii) a second temperature rise rate coefficient for the lubricating oil for the second oil lubricated component;

(f) repeating steps (a) to (e) for a plurality of load cycles;

(g) comparing a plurality of determined first system performance parameter values and second system performance parameter values to identify a change in said first system performance parameter or said second system performance parameter over the plurality of load cycles. In this embodiment, the first oil lubricated component is preferably a gear.

The step of measuring a load of the first oil lubricated component preferably includes measuring a power transferred through said gear.

Alternatively, the first oil lubricated component is a swivel.

The step of measuring a load of the first oil lubricated component preferably includes measuring a hook load acting on the swivel.

Alternatively, the first oil lubricated component is a gear and the second oil lubricated component is a swivel. The step of measuring a load of the first oil lubricated component preferably includes measuring a power transferred through said gear and the step of measuring a load of the second oil lubricated component preferably includes measuring a hook load acting on the swivel. The step of comparing a plurality of determined first system performance parameter values and second system performance parameter values preferably includes comparing a plurality of first system performance parameter values with a plurality of second system performance values. The method may preferably further comprise the step of determining a third system performance parameter, the third system performance parameter being a function of the first system performance parameter and the second system performance parameter. The third system performance parameter is preferably a ratio of the first system performance parameter to the second system performance parameter.

Alternatively, the third system performance parameter is a difference between the first system performance parameter and the second system performance parameter.

Said plurality of load cycles are preferably the first 2-5 cycles following a start of the mechanical system from a cold condition.

The mechanical system is preferably a top drive for a drilling rig.

Brief description of the drawings

Figure 1 shows an illustration of a top drive for a drilling rig.

Figure 2 shows a typical plot of input power to a top drive during a drilling sequence.

Figure 3 illustrates the temperature increase in an oil bath following the start of a drilling load cycle.

Figure 4 shows the temperature of the gear box and swivel oil baths through a series of drilling sequences.

Figure 5 shows in more detail an extract of the data shown in Fig. 4.

Figure 6 shows a set of measured gear lubrication oil temperatures and a fitted first-order model.

Figure 7 shows a measured torque, a measured swivel oil temperature and a gear oil temperature for three drilling events.

Description of preferred embodiments

Preferred and exemplary embodiments of the invention will now be described in relation to a system used for earth drilling, however the invention is suitable for use with a wide range of equipment and is not restricted to the specific examples shown here. Figure 1 shows an illustration of a top drive for a drilling rig, having a gear box 1 and a swivel 2, both being oil lubricated. In a first embodiment of the invention, there is provided a method of using live and/or recorded data from a drilling machine to estimate the condition of the gear box and swivel. A typical drilling machine or top drive may be composed of one on more motors 3, electrical or hydraulic, with an associated gear box 1 for adapting the rotational speed and torque of the motors to the requirements for rotating the drill string with a suitable rotational speed and torque. The top drive is provided with connection points 5 on lifting yokes 6 for connection to hoisting wires in a hoisting or lifting system on the drilling rig. In its lower end, the top drive is provided with elevator links 7 and an elevator 8 for engaging the top end of a drill string. In use, the top drive provides rotational torque to the drill string by means of motors 3 and controls its vertical position in relation to the wellbore through the hoisting system.

The speed and torque of the drill string may be monitored by physical sensors coupled to a drilling control system, and are recorded for monitoring purposes. Alternatively, the speed and torque of the electric or hydraulic motor(s) 3 can be measured. The gear box 1 is lubricated by means of an oil bath, and the temperature of the oil bath is measured by physical sensors coupled to a drilling control system, and is recorded for monitoring purposes. To separate the axial load from the motor(s) 3 and gear box 1 there is a thrust bearing 2, also known as the swivel. This swivel 2 is lubricated by means of an oil bath, and the temperature of the oil bath is measured by physical sensors coupled to a drilling control system, and is recorded for monitoring purposes. The oil bath for the swivel 2 can be separate from the oil bath of the gear box, or it may be the same oil bath. As a vital part of the drilling process, a drilling fluid, often colloquially referred to as "mud", is circulated downward to the drill bit through the drill string and returned to the surface, normally through the annular space between the drill string and the well casing. Normally, this drilling fluid is also circulated through the main shaft 4 of the top drive, thus passing through both the gear box 1 and the swivel 2. A drilling sequence on an offshore drilling rig consists of drilling one stand (approx. 30 m) followed by a connection period with no drilling. During the drilling phase (also known as a drilling sequence or drilling event), the temperatures of both the gear box and the swivel will rise considerably, and these temperatures will drop during the period with no drilling (the connection phase or event). The temperature rise is due to the friction loss in the gear box and swivel, respectively. Figure 2 shows a typical plot of input power 10 to the top drive during a drilling sequence. The drilling sequences can be seen; these may take in the order of 100 minutes, while the connection phase may take 30 minutes. (However these values may vary significantly according to the relevant operating conditions.)

Upon start of a drilling sequence, the lubrication oil temperature in the oil bath for the gear box and the swivel will increase, and this increase will typically follow a first-order type response. Figure 3 illustrates this, showing the oil bath temperature development graph 20 against time. Figure 3 also shows the initial rise rate 21 for the first-order graph 20 and the final (steady state) response 22. Both these parameters are well-known for first-order functions and straightforward to calculate.

This type of heating model is uniquely defined by only two parameters: the initial heating rate before heat loss to the ambient is significant (the graph 21 ) and the steady state limit (the top asymptote, graph 22) where the heat loss is equal to the friction power input. With a constant heating over a drilling event, starting at temperature T ₀ , the temperature T(t) will increase slower as the heat loss increases. With an approximately linear dependency between temperature difference and heat loss, the solution of the inherent differential equation fits well with a simple exponential function in time:

T(t) = To + (T. - To)(1 - exp ( - (t / t _c ) ) )

When a steady state is reached, there is no net heat increase, P _h = P _f - Pi = 0 when T approaches T«. The final temperature limit T« will be given by T« -T ₀ = P, / I. The above subscripts h, f and I refer to heating, friction and heat loss to the ambient. P _h therefore means net power that goes into heating the component, P _f is the power dissipated in the component because of friction, and Pi is the power lost to the ambient because of temperature difference between the component and the ambient. Figure 4 shows temperature plots 1 1 , 12 of the swivel and gear box oil baths, respectively, through a series of drilling sequences interrupted by pauses, followed by a long cooling event. The total time shown is approx. 89 hours (=3.71 days) with 4 hours / grid. The cyclic behaviour of the temperatures during operation can be seen.

Figure 5 shows an extract of the data shown in Fig. 4, i.e. the temperature of the gear box and swivel oil baths, at the start of drilling and over 5.63 hours (30 min/grid). As the drilling starts, there is a rapid temperature increase in both oil temperatures. The plot in Fig. 5 shows 3 drilling events each followed by a cooling event when connecting new stands of drill pipe.

By providing temperature sensors for logging the swivel and gear box oil bath temperatures, and analysing the response in these parameters to a change in the load, a deteriorating condition or damage in the swivel or gear box can be identified at an early stage. One method of achieving this according to the present invention will now be described. In this embodiment of the invention, a method is provided comprising steps (a) to (f). The method will be described employed on both the swivel and the gear box oil bath, however it is to be understood that the method may be employed individually on either of these components, or on both.

During operation of a top drive system, a first step (a) comprises identifying the start of a load cycle, i.e. a drilling sequence. This can be done for example by measuring the power input to a motor (e.g. an electric or hydraulic motor) driving the top drive unit (see Fig. 2), by measuring the torque applied by the motor, or by measuring the speed of the top drive. Several of these variables may be directly available from a drilling control system.

In a second step (b), a load of the swivel and/or gear box is measured. The load for the gear box may be the power transferred through the gear box (which can either be measured directly, measured from the motor power, or calculated based on e.g. top drive rotational speed (rpm) and torque). For the swivel, which does not transfer any load per se, the load may be a "virtual load". A particularly useful parameter for estimating swivel load is as a function of rotational speed (rpm) and hook load, i.e. the weight carried by the top drive and the swivel. (This may be a length of drill string extending downhole below the drilling rig.)

Thus, in this example, for the gear box the relevant load is the rotational power equal to torque (in Newton-meter) times rotational speed (in radians per second), equals Watt. For the swivel, taking up the vertical load on the drilling machine the relevant power is the vertical load in Newton times the mean radius of the roller bearing times the rotational speed. The heating power equals the specific heat capacity of the oil and steel, multiplied by the mass of oil and steel, respectively. In a third step (c), the temperature of the lubricating oil in the swivel and gear box oil baths are measured continuously or periodically during the load cycle.

In a fourth step (d), a system performance parameter for the load cycle is calculated based on the measured load and temperature values. This system performance parameter may be: (i) a friction coefficient, (ii) a temperature increase coefficient, or (iii) a temperature rise rate coefficient for the lubricating oil for the first oil lubricated component.

A friction coefficient may be defined as the ratio of the instantaneous friction heating power divided by the relevant load of the same component (gear box or swivel). By monitoring the heating sequences, the relevant power inputs and, optionally, the ambient temperature plus the mud temperature and mud flow, one can use this thermal model to calculate an instantaneous (or, for example, averaged over a drilling event of approx. one hour) friction coefficient for each relevant component (gear box and swivel) for every drilling event.

The friction heating power added to the lubricating oil may be calculated or estimated based on the measured temperature increase of the oil, if necessary taking into account the heat capacity of the oil and/or relevant mechanical components which will be heated by the oil. If appropriate one may also take into account heat loss to the ambient, e.g. by estimating this as proportional to the instantaneous temperature difference between the oil and ambient temperature.

The friction coefficient is thus indicative for the work lost to friction in the component, which will increase in the case of a deteriorating condition of that component. A temperature increase coefficient can be found by measuring the temperature increase of the oil in response to the start of a drilling event. This may be in the form of a steady-state value of the lubrication oil temperature, if this

temperature reaches a substantially steady value during the drilling event, or, alternatively, the temperature increase coefficient may be equal to the temperature increase itself.

Alternatively, the temperature increase coefficient may be based on a temperature increase at a specific time, pre-determined following the start of a drilling event. This is shown in Figure 6, where the temperature of a set of measured gear lubrication oil temperatures 30 is read off at a time 40 minutes after the start of an event, as indicated by dashed lines 31 and 32.

Yet another alternative may be to define the temperature increase coefficient as a function of the temperature increase and the load, for example the

temperature increase divided by average load over that drilling event or measurement interval. Referring to Fig. 7, this may, as an example, be taken as the temperature increase divided by the average torque during the first 1 h of operation of a drilling event; for the first cycle in Fig. 7 this would give a temperature increase coefficient for gear oil temperature 42 of approximately 5 °C / 16 kNm = 0.31 °C/kNm.

The temperature increase coefficient is therefore indicative of the level of temperature increase in the lubricating oil for a component, in response to operation of the mechanical system. A temperature rise rate coefficient can be found by approximating the temperature response in each load cycle to a first-order graph (see Fig. 3) and calculating the initial rise rate 21 . Figure 6 illustrates a fitted first-order model 33 to the measured gear lubrication oil temperatures 30. Having a fitted model, one can calculate the time constant, initial rise rate, and/or the final value of the modelled temperature response. Either of these, or a combination, may form the temperature rise rate coefficient on their own, or in combination with the measured load for the relevant drilling event.

Alternatively, a simplified temperature rise rate coefficient can be calculated for example by taking the temperature increase from the start of the load cycle to a pre-determined time thereafter, for example 40 minutes, and calculating the temperature rise rate in that interval. For example, the data point indicated by dashed lines 31 ,32 in Fig. 6 would give a temperature rise rate in this interval of approx. 18.5 °C / 40 min. = 0.5 °C/min

The temperature rise rate coefficient is therefore indicative of the rate of temperature increase in the lubricating oil for a component, in response to operation of the mechanical system.

For systems or modes of operation where the load varies between cycles, the load is preferably taken into account when defining the temperature increase coefficient and the temperature rise rate coefficient. However where the load is substantially constant over a plurality of cycles, this may not be necessary.

The method further comprises repeating steps (a) to (d) for a plurality of load cycles. Figure 7 shows three drilling events: A, B and C, and a measured torque 40, a measured swivel oil temperature 41 and a gear oil temperature 42. One implementation of the method according to the invention may in this case be to: identify the start of a load cycle (A, B and C) by identifying the increase in the measured torque 40;

defining the load for the gear as the applied torque and measuring this to produce a measured torque signal 40;

measuring the temperature of the gear oil to produce a measured temperature signal 42;

defining the performance parameter for the gear as a temperature increase coefficient equal to the temperature increase 60 minutes after start of a drilling event divided by the average measured torque 40 over the same period (for cycle A this is equivalent to approx. 5 °C

temperature increase divided by approx. 17 kNm average torque);

repeating the above steps for cycles A, B and C (and, optionally, further cycles); and

comparing the values for individual cycles over time to identify a deteriorating trend.

Preferably more than 10 load cycles, or more preferably, more than 30 load cycles, or even more preferably, more than 50 load cycles are used in order to improve the accuracy of the predictions. Thus, over time, the performance parameter (friction coefficient, temperature increase coefficient or temperature rise rate coefficient) can be trended, and a reliable mean or median computed, with a more or less stable standard variation (sigma). If the component (gear box or swivel) for example has suffered abnormal wear, the friction work from the component will increase, which will generally lead to increased heat dissipation into the lubricating oil. Thus, the performance parameter will tend to increase gradually. When a stable upward trend is detected, exceeding the 4, 5 or 6 sigma levels, notifications to the maintenance responsible can be given, noting an early warning of a deteriorating condition. In a further embodiment according to the present invention, there is provided a method comprising the steps:

(a) identifying the start of a load cycle;

(b) measuring a temperature of a lubricating oil for the first oil lubricated component, such as a swivel,

(c) measuring a temperature of a lubricating oil for the second oil lubricated component, such as a gear,

(d) determining a first system performance parameter for the load cycle, selected from a group of (i) a first friction coefficient for the first oil lubricated component, (ii) a first temperature increase coefficient for the lubricating oil for the first oil lubricated component, and (iii) a first temperature rise rate coefficient for the lubricating oil for the first oil lubricated component;

(e) determining a second system performance parameter for the load cycle, selected from a group of (i) a second friction coefficient for the second oil lubricated component, (ii) a second temperature increase coefficient for the lubricating oil for the second oil lubricated component, and (iii) a second temperature rise rate coefficient for the lubricating oil for the second oil lubricated component;

(f) repeating steps (a) to (e) for a plurality of load cycles;

(g) comparing a plurality of determined first system performance parameter values and second system performance parameter values to identify a change in said first system performance parameter or said second system performance parameter over the plurality of load cycles.

Advantageously, the first and second system performance parameters are compared with each other. This has the benefit that a deteriorating trend can be identified in one system performance parameter relative to the other. For example, one would not normally expect both components to fail at the same time. Thus, by comparing a system performance parameter of one component to a system performance parameter of the other component, one can identify a deteriorating trend without the need to take external variables into account. In an advantageous embodiment, a third performance parameter is defined, the third performance parameter being a function of the first and second

performance parameters. In one embodiment, the third performance parameter may be a ratio of the first and second performance parameters. In another embodiment, the third performance parameter may be a difference between the first and second performance parameters.

This provides the advantage that a deteriorating condition can be identified based on the development in the third performance parameter. For example, one would expect the first and second performance parameters to be somewhat influenced by external conditions, such as ambient temperature, mud

temperature, etc. A third performance parameter, being, for example, a ratio between the first and second performance parameters will be significantly less sensitive to such external factors.

Thus, by treating for example the difference between or ratio of the two measured oil bath temperature increase coefficients as a new variable, improved accuracy can be achieved. Normally, the temperatures of the two oil baths will follow similar development curves, not equal but related. Both of them are unlikely to fail or change behavior significantly at the same time, but both will be somewhat influenced by e.g. ambient temperature. However, if one suddenly changes behavior and gets significantly warmer, relative to the other, it may be suspected to fail in the near future. The third performance parameter may thus identify such a condition, while eliminating noise from other factors.

An advantage of comparing two performance parameters with each other, either directly or by means of defining a third performance parameter, is that two components of a mechanical system can be used for condition monitoring or state estimation, even if these components have substantially different operating characteristics or functions. For example, in a top drive, the swivel only carries the load via a rolling motion whereas the gear transfers the rotating power through a gear system. Thus their friction coefficients will be substantially different, however they can be expected to respond similarly to changing operating conditions of the mechanical system.

An advantage of the method according to the present invention is that the operational characteristics of a system which is subject to regular and intermittent loads can be utilized to improve accuracy of the measurements and predictions. This is because there will always be measurement noise and random or cyclical variations, but by carrying out measurements over a large number of cycles one can reliably identify trends which would otherwise not be possible to measure on individual or a small number of cycles. Thus, for example, from the oil bath temperature increases and the power usage, a friction coefficient (e.g. friction heating power divided by input power) can be calculated and trended over time. Deviation from the long time average above a certain number of standard deviations will be the basis for early notification.