SE547740C2

SE547740C2 - A method for controlling auto-cycling of a hydraulic torque wrench system and a hydraulic torque wrench system for performing the method

Info

Publication number: SE547740C2
Application number: SE2430483A
Authority: SE
Inventors: Ola Åkerbladh
Original assignee: Atlas Copco Ind Technique Ab
Priority date: 2024-09-24
Filing date: 2024-09-24
Publication date: 2025-11-18
Also published as: SE2430483A1; WO2026068133A1

Abstract

A method is provided for controlling auto-cycling of a hydraulic torque wrench system (1). The method comprises: setting a target pressure (Pmax) for a tightening; receiving a signal indicative of said hydraulic pressure; and detecting if any one of certain conditions a)-c) occurs during a trial cycle (TCO) performed as an initial cycle of the tightening. The method further comprises: in response to detecting condition a) or c), estimating an exit pressure of the performed trial cycle and continue the tightening by further auto-cycling; and/or in response to detecting condition b), commanding termination of the tightening. With the present method, an estimation of which state the joint is in at the beginning of the tightening is provided and action may be taken accordingly.

Description

[0001] The present invention generally relates to the field of hydraulic torque wrench systems. In particular, the present invention relates to controlling auto-cycling of such hydraulic torque wrench systems.

[0003] Background of the invention

[0005] Hydraulic torque wrench systems are typically used for high torque applications with large bolts e.g. in wind towers and to join flanges in the oil and gas industry. Such systems comprise a wrench coupled to a hydraulic pump. In the wrench, a piston is arranged in a cylinder to alternatingly advance and retract by the application of hydraulic pressure from the pump on either side of the piston. The piston is in turn mechanically connected to a pawl and ratchet mechanism that translates the advancing movement of the piston to a rotating movement of an output interface (such as a square or socket) of the wrench, whereby the fastener is tightened. The pawl and ratchet mechanism is further arranged to refrain from translating the retracting movement of the piston to the output interface.

[0007] When manually cycling such systems, an operator has to operate one or more actuators on a pendant to manually initiate advancement and retraction of the piston. It may be difficult for the operator to be sure of when the tightening is done, since this is decided by the operator by simply watching when the fastener stops moving. To be sure that the tightening is done, the operator typically cycles over and over again to be sure, which takes a lot of time. Hence, such manual cycling is both cumbersome and time consuming. Therefore, control devices have been developed that automatically controls the cycling of the system, which in the art is referred to as auto-cycling. When auto-cycling, the control device need not just keep track of when the target torque is reached but also when to switch from advancing to retracting the piston. In order to do so, such control devices are typically provided with input from different sensors in the system. First and foremost, the hydraulic pressure that acts on the side of the piston and causes the piston to advance (and tighten the joint) is sensed and used as an input. Sometimes, also flow sensors are used to monitor the hydraulic flow in and out of the cylinder to keep track of what happens in the cycles. For cost reasons, it is however desirable to reduce the number of sensors to a minimum.

[0008] In order for auto-cycling to work properly, the control device may preferably keep track of a number of parameters during the tightening. For example, the control device may need to know when to switch from advancement to retraction of the piston and it may be desirable to estimate the angle of the tightening. Limiting the number of sensors to a minimum, this is a challenging task.

[0010] A particular challenge, which has not yet been solved in the prior art, is for the control device to estimate which state the joint is in at the start of the tightening operation. The joint may e.g. be in the rundown phase (free running), or already fully or partly tightened (the latter may be referred to as a rehit and partial rehit, respectively). To estimate in which state the joint is in may be useful for further control of the tightening.

[0012] Summary of the invention

[0014] It would be advantageous to achieve a method and a control device overcoming, or at least alleviating, the above-mentioned drawbacks. In particular, it would be desirable to enable a method and a control device able to estimate the state of the joint in the beginning of a tightening performed by means of auto-cycling. It would also be desirable to enable a method and control device able to act according to the estimated state of the joint for enhanced further control of the tightening.

[0015] To better address one or more of these concerns, a method and a control device having the features defined in the independent claims are provided. Preferable embodiments are defined in the dependent claims.

[0016] Hence, according to a first aspect, a method, performed by a control device, for controlling auto-cycling of a hydraulic torque wrench system is provided. The auto-cycling comprises regulating a hydraulic pressure of the system to alternatingly advance and retract a piston of the system for driving an output interface (such as a square or socket) of the system for tightening a joint.

[0017] The method comprises:

[0019] - setting a target pressure for a tightening;

[0020] - receiving a signal indicative of said hydraulic pressure (over time); and - detecting if any one of the following conditions occurs during a trial cycle, after any play in the mechanical connection between the piston and the output interface has started to close, the trial cycle being performed as an initial cycle of the tightening:

[0021] a) the gradient of the signal exceeds and stays over a predetermined gradient threshold until the target pressure is reached and

[0022] the time it takes to then retract the piston (in the trial cycle) is above a predetermined retract time threshold,

[0023] b) the gradient of the signal exceeds and stays over the predetermined gradient threshold until the target pressure is reached and

[0024] the time it takes to then retract the piston (in the trial cycle) is below the predetermined retract time threshold, and

[0025] c) the gradient of the signal first exceeds the predetermined gradient threshold and then drops and stays below it for a predetermined amount of time. The method further comprises:

[0026] - in response to detecting condition a) or c), estimating an exit pressure of the performed trial cycle and continue the tightening by further auto-cycling; and/or

[0027] - in response to detecting condition b), commanding termination of the tightening.

[0029] According to a second aspect, a control device is provided. The control device is configured to perform the method according to the first aspect.

[0031] The inventor has realized that the above defined conditions a)-c) are representative of three common states that the joint may be in at the beginning of the tightening. Hence, by detecting whether any one of these conditions prevails in the very first cycle of the tightening, an estimation of which state the joint is in is provided and the method and control device may act accordingly, either by estimating an exit pressure of the initial cycle accordingly or terminate the tightening. The exit pressure of a cycle means the pressure installed in the joint at the end of the advance phase of the cycle. The exit pressure will depend on the state the joint is in. Knowing the exit pressure of the initial cycle is useful for further control of the tightening. Hence, improved further control of the tightening is provided.

[0033] In particular, condition a) may indicate that the joint is in a rundown phase at least in the beginning of the trial cycle. Here, the fastener of the joint runs freely from start, either all the way until the piston reaches its' dead end or until the fastener reaches snug, whereupon torque starts to build up. The pressure will therefore first be steady (with a gradient around zero) as the fastener runs freely, and if snug is reached, it will increase gradually. As the piston hits the very end of its stroke, the pressure gradient will drastically rise and end up above the predetermined gradient threshold and stay there until the target pressure is reached. This may be detected/identified according to the present method. To be able to distinguish between a rundown and rehit (condition b), will be described below), the time it takes for the piston to then (fully) retract may be measured/determined. If that takes more than the predetermine retract time threshold, that may indicate that the piston has travelled a full stroke length and that the joint is in a rundown phase. This may be detected/identified by the method and an exit pressure of the initial cycle may be estimating accordingly.

[0034] Condition b) may indicate that the joint is in a rehit state. This means that the joint is already tightened to the target torque or more. Here, the pressure will either directly (as soon as any gap in the mechanical connection between the piston and the output interface is closed) start to rise with a gradient exceeding the predetermined gradient threshold (in case the joint is tightened to the target torque/pressure or more), or after some time in the trial cycle (if the joint is tightened to at torque slightly below the target torque/pressure), and continue to do so until the target pressure is reached. To be able to distinguish the rehit from a rundown (condition a)), the time it takes for the piston to then (fully) retract may be measured/determined. If that takes less than the predetermine retract time threshold, that indicates that the piston has not travelled a full stroke length, but rather just a part of a full stroke, or even barely not moved at all (except for closing the play in the mechanical connection between the piston and the output interface). This in turn may indicate that the joint is in a rehit state. This may be detected/identified by the method, whereupon it may be commanded that the tightening should be ended (that is, not continued after the trial cycle) accordingly, since the joint is estimated as already fully tightened.

[0035] Condition c) may indicate that the joint is in a partial rehit state, meaning that some torque is already installed in the joint. The characteristic of the pressure curve in this state is that the pressure first may rise with a steep gradient, exceeding the predetermined gradient threshold, as the piston stands still and the pressure builds towards a level able to start moving the piston. As the piston starts to move (and install torque in the joint), the gradient of the pressure curve will decrease below the predetermined gradient threshold again. As the pressure gradient has been below the predetermined gradient threshold for the predetermined amount of time, condition c) is detected and an exit pressure is estimating accordingly.

[0037] In the present specification, when referring to "end-of-stroke" or "the end of the stroke of the piston", it is meant the end of an advance stroke of the piston.

[0038] In the present specification, the pressure in the second (i.e. following) cycle (roughly corresponding to the exit pressure of the first cycle) may sometimes be referred to as the entry pressure of the second cycle. Hence, the entry pressure of one cycle is the pressure when the fastener starts to turn and the tightening start to progress, whereas the exit pressure of a cycle is the pressure installed in the joint at the end of the advance stroke/phase of the cycle, such as when the piston reaches the very end of its' stroke, whereby the progression of the tightening is paused until the entry pressure of the next cycle is reached.

[0040] In the present specification, if not otherwise defined, when referring to "pressure" it is meant the pressure on the advance side of the hydraulic system. In other words, the pressure acting on the side of the piston that can push the piston in the advancing direction.

[0042] Estimating the exit pressure of the trial cycle in response to the detection of condition a) may be made in different ways.

[0044] For example, according to an embodiment, estimating the exit pressure in response to condition a) may comprise, for each one of a plurality of data points of the signal:

[0045] - for a first interval ranging forward from the data point, estimating a first average gradient; and

[0046] - for a second interval ranging backwards from the data point, estimating a second average gradient (dP2).

[0048] The exit pressure of the trial cycle may then be estimated based on the estimated first and second average gradients.

[0050] The present embodiment is advantageous in that the actual exit pressure of the trial cycle can be rather accurately detected/estimated. For example, the present embodiment may estimate a relatively accurate exit pressure irrespective of if the trial cycle was a complete rundown cycle or if it was partly rundown and partly an installation of torque (i.e., if the fastener became snug during the trial cycle).

[0051] By analysing the first and second average gradients for a particular data point, it may be possible to understand how the pressure changes in an area around that data point. For example, it may be possible to identify that, around a specific data point, the pressure gradient suddenly gets steeper. That may be an indication that the exit pressure of the cycle is reached.

[0053] For example, the first and second intervals may each range a predetermined number of data points (forwards and backwards, respectively) from the data point.

[0055] The average gradients may not necessarily be an arithmetic mean gradient of the interval, but rather an approximation/representation thereof.

[0057] For example, linear regression may be applied to the data points in the interval, whereby the gradient (such as a k-value) of the resulting linear function may be determined as the average gradient. Other analysing techniques for deriving the first and second average gradients may also be envisaged.

[0059] According to an embodiment, the method may further comprise, for each one of the plurality of data points, determining a score indicative of a difference between the first average gradient and the second average gradient. The estimation of the exit pressure may then be based on that score.

[0061] The score may e.g. be a calculated difference between the first and second average gradients itself, a ratio between the first and second average gradients or any other parameter/value indicative thereof.

[0063] The score may thus indicate how the first and second gradients are related to each other to thereby give input in order to identify the point in the curve where the piston reaches the end of the stroke, whereby the exit pressure is reached.

[0064] According to an embodiment, the exit pressure may be estimated (and accordingly set in the software of the control device) as the pressure of the one data point in a set of at least some of the plurality of data points (such as a subset of the plurality of data points) having a determined score indicative of the largest difference between the first average gradient and the second average gradient. This set of at least some of the plurality of data points may comprise only data points having positive first and second average gradients. In other words, the set may be exclusive from data points having at least one of the first and second average gradients being negative.

[0066] The present embodiments are advantageous in that the exit pressure can be estimated with a high degree of accuracy since each one of the data points is evaluated with respect to its' first and second average gradients. It may therefore be possible to find the one data point where the difference between the gradient before and after is the largest, which is a clear indication of that the exit pressure has been reached at that point. Inaccurate estimation of the exit pressure may cause higher risk of erroneous detection of the end of the stroke in the upcoming cycle in certain cases, such as when the end of the stroke of the upcoming cycle is very close to the entry pressure of that cycle.

[0068] Sometimes, small jumps/notches may occur in the pressure signal. In order not to mistake data points in such jumps/notches for a data point at the exit pressure, the data points having at least one of the first and second gradients being negative may be discarded from the set. The data point at the true exit pressure will have both first and second average gradients being positive.

[0070] According to an embodiment, the estimation of the first and second average gradients for the data points may be made with start from a data point subsequent to the termination of advancement of the piston of the first cycle (such as from when an end of stroke is detected or from a peak pressure of the cycle) and then for data point by data point backwards.

[0072] The present embodiment is advantageous in that computing resources are saved, as the exit pressure as a rule is found in the end of the pressure build-up phase of the cycle. So, by starting from behind, only a relatively small part of the signal needs to be examined to estimate the exit pressure.

[0074] According to an embodiment, evaluating the signal in order to estimate the exit pressure may be made retrospectively, after the pressure has peaked in the first cycle. Since the estimated exit pressure is not needed until the next cycle starts, the analysis of the signal to derive the exit pressure can be made after the torque build up phase, such as during the retraction phase, or even after the retraction phase, but before the next advance phase of the piston starts.

[0076] According to an alternative embodiment, the method may further comprise, in response to detecting condition a),

[0077] - estimating the exit pressure of the trial cycle to zero;

[0078] - in the next cycle, estimating the time it takes until the end of the stroke is reached; and

[0079] - in response to the estimated time being below a predetermined advance time threshold, perform another (which may be referred to as a second) trial cycle and detect if any one of conditions a)-c) occurs. The actions in response to the detected condition of the second trial cycle may be the same as those of the initial trial cycle.

[0081] Hence, in case condition a) is detected, the exit pressure may be estimating to zero by default. However, in condition a) -cases where the actual exit pressure of the trial cycle is more than zero, such as if the fastener has reached snug during the trial cycle and torque has started to get installed, an exit pressure set to zero may get consequences. In particular, if the exit pressure estimated in the trial cycle is used as a bar for when to enable end-of-stroke detection in the next cycle (as will be described further below), the first steep part of the curve resulting from the pressure building to a level able to start to advance the piston and tighten the joint may be mistakenly detected as an end-of-stroke. As a consequence, retraction of the piston will be commanded way too early, such as before the fastener has started to turn again. In order to detect end remedy that, the present embodiment may further comprise, in the very next cycle, estimating the time it takes until the end of the stroke is reached. If the estimated time is below the predetermined advance time threshold, that may indicate that a way too early end-of-stroke detection has been made, which in turn is an indication that the actual exit pressure of the trial cycle was more than zero. Therefore, another trial cycle may be commanded. This means that, in the following cycle, monitoring of any one of the conditions a)-c) may be performed again. Now, as some torque has been installed in the joint during the first trial cycle (which was partly rundown and partly installation of torque), condition c) will be detected in the commanded second trial cycle and the exit pressure estimated accordingly.

[0082] This embodiment may be less complex and require less computing power than detecting the actual exit pressure as described in the previous embodiment.

[0083] The time to be compared with the predetermined advance time threshold may be estimated (such as measured) from any suitable starting point identifiable by the control device in the beginning of the cycle following the trial cycle, such as from the very beginning of the cycle (that is, as soon as advance pressure is commanded/applied).

[0085] The predetermined advance time threshold may be based on the time it takes for the piston to make a full advance stroke under no load, e.g. as measured during a learning cycle performed before the tightening starts for calibrating the system.

[0086] The actions in response to the detected condition of the second trial cycle may be the same as those of the initial trial cycle.

[0088] According to an embodiment, the method may further comprise:

[0089] - in response to detecting condition c), commanding retraction of the piston in the trial cycle without awaiting end-of-stroke to be reached (that is, retraction may be commanded before the end-of-stroke of the trial cycle is reached, preferably as soon as condition c) is detected); and

[0090] - estimating an exit pressure of the performed trial cycle based on the value of pressure just before the retraction of the piston in the trial cycle.

[0092] Hence, once the method has detected that condition c) prevails, which may indicate that the joint is in a partial rehit state, the trial cycle may be prematurely terminated (that is, before the drastic pressure increases at the dead end of the stroke). The exit pressure may then simply be estimated/set based on (such as to) the pressure detected just before termination of the advance stroke. In other words, the peak pressure of the trial cycle may be used as basis for estimating the exit pressure. The present embodiment is advantageous in that it provides a simple and accurate way of estimating the exit pressure of the trial cycle in case the joint is estimated as being in a partial rehit state. As the trial cycle is terminated when the piston advances and applies torque to the joint, the pressure at the termination will be the exit pressure of the trial cycle.

[0094] According to an embodiment, the method may further comprise using the estimated exit pressure of the trial cycle as an input value in continued control of the tightening. Hence, further control of the auto-cycling may be based on the estimated/set exit pressure of the trial cycle.

[0096] For example, the method may further comprise enabling detection of an end-ofstroke in the next cycle (that is, the cycle directly following the trial cycle) only after the pressure in the next cycle has exceeded the estimated exit pressure of the trial cycle. Hence, the estimated exit pressure of the trial cycle may be used as a bar for enabling end-of-stroke detection in the following cycle.

[0098] By awaiting the pressure to build up to the same level as the estimated exit pressure of the previous cycle before enabling end-of-stroke detection, the very first part of the signal/curve of the second cycle is never examined/monitored for the purpose of detecting the end of the stroke. That first part of the advance pressure signal/curve typically contains at least one steep increase in pressure caused by the pressure building up to a level able to start advancing the piston, and then eventually to turn the threaded fastener of the joint. That steep pressure increase may be erroneously recognized as an end-of-stroke gradient. With the present method, the risk of such erroneous detection of the end of the stroke is reduced as the detection of the end of the stroke is enabled only in response to the pressure in the second cycle reaching the estimated/ determined value of the exit pressure of the first cycle. End-of-stroke detection may thus be disabled as long as the pressure of the second cycle is below the exit pressure of the previous cycle. Consequently, the need for using a time factor to delay the enabling of the end-of-stroke detection is reduced, or even eliminated. With the present method, only the relevant part of the curve, i.e., the part after the fastener of the joint has started to turn, is analysed in order to identify the end of the stroke. Accordingly, more accurate end-of-stroke detections are enabled.

[0100] According to an embodiment, the method may further comprise using the estimated exit pressure of the trial cycle as an input value for estimating a parameter of the tightening.

[0102] The estimated parameter may e.g. be used as a control parameter during the tightening and/or it may be saved as a result when the tightening is completed.

[0103] For example, the estimated parameter of the tightening may be an angle of the tightening. Hence, the estimated exit pressure may be used for estimating the tightening angle.

[0105] For example, the estimated exit pressure of the trial cycle may be utilized to estimate an entry pressure (the pressure at which the fastener starts to turn) for the next cycle. This entry pressure may then be utilized to estimate the angle.

[0106] According to an embodiment, a pump unit comprised the system and arranged to provide said hydraulic pressure is arranged to operate in at least two different stages, and wherein each stage is associated with a respective predetermined gradient threshold used to detect conditions a)-c).

[0108] Hence, the pump unit may be a multistage pump able to adapt the hydraulic flow to the current pressure in the hydraulic system. As an example, a pump unit having three stages will operate in the first stage up to a certain pressure threshold and then shift to the second stage. As the pressure exceeds a second pressure threshold, the pump will shift to its third stage. In the different stages, the hydraulic flow from the pump will be different. So, with the present embodiment, each stage of the pump may be associated with a specific gradient threshold. If the stage of the pump shifts, the gradient threshold that is used for comparison may shift accordingly.

[0110] According to an embodiment, the predetermined gradient threshold may have been set based on a pressure gradient occurring when advance pressure is applied and the piston is not moving. Such a pressure gradient may e.g. be an end-of-stroke gradient occurring when the piston has reached the end of its stroke. For example, the pressure gradient of the signal, which the predetermined gradient may be based on, may be detected during one or more learning cycles performed prior to the tightening for calibrating the system. For example, the predetermined gradient threshold may be equal to, or set slightly below, the end-of stroke gradient. As an example, the predetermined gradient threshold may be set to a value comprised within the range of 70-90% of the end-of stroke gradient.

[0112] According to an embodiment, the predetermined retract time threshold may have been set based on a retract time for a complete stroke of the piston.

[0114] For example, the retract time for a full stroke, which the predetermined retract time threshold may be based on, may be measured during one or more learning cycles performed prior to the tightening for calibrating the system. For example, the predetermined retract time threshold may be equal to, or set slightly below, such a measured retract time.

[0116] According to an embodiment, the detecting of the conditions a)-c) may be based on a derivative of the signal.

[0118] The present embodiment is advantageous in that the derivative of the signal may be easier/more distinct to analyse as compared to the original signal.

[0120] In the present specification, "the derivative" of the signal means any representation of the rate of change of the pressure signal. The derivative of the signal may be provided by calculation and/or filtering.

[0122] According to an embodiment, the method may further comprise (prior to the step of detecting if any one of conditions a)-c) occurs), based on the signal, detecting when the play in the mechanical connection (such as in pawl and ratchet mechanism) between the piston and the output interface has started to close, and in response to that detection, perform the step of detecting if any one of conditions a)-c) occurs.

[0123] The very first part of the pressure curve of each cycle may have certain characteristics that may be detected. For example, a first peak (or bump) may occur in each cycle as the pressure builds for overcoming the static friction between the piston and the cylinder and any remaining retract pressure. Then, a short period of steady pressure will follow as the piston starts to move and, without any significant resistance, close the gap between the pawl and the ratchet. Only then, the force from the piston will be transferred to the output interface. Such characteristics of the curve may be detected in any suitable way. For example, it may first be detected that the value of the pressure exceeds a certain threshold, then, if the pressure of the next data point is higher than that of the previous, that pressure is saved as a maximum pressure. When the pressure of the next data point is less than that of the maximum pressure, a counter is incremented. If a predetermined number of data points (e.g. ten) in a row are less than the maximum pressure, that is detected as an indication that the initial bump in pressure has been overcome, which means that the play in the mechanical connection between the piston and the output interface has started to close. It may be advantageous to await this initial bump in the pressure before starting to monitor the signal for detecting any one of the conditions a)-c) in order to avoid the initial bump and the following steady pressure period to be erroneously detected as condition c).

[0125] According to an embodiment, a hydraulic torque wrench system is provided comprising a control device according to the second aspect.

[0127] According to an embodiment, a computer program is provided comprising instructions which, when the program is executed by a computer (such as the control device), cause the computer to carry out the method as defined according to the first aspect or any one of the above-described embodiments.

[0128] Computer-readable storage medium comprising instructions which, when executed by a computer (such as the control device), cause the computer to carry out the method as defined according to the first aspect or any one of the above-described embodiments.

[0129] These and other aspects will now be described in more detail in the following illustrative and non-limiting detailed description of embodiments, with reference to the appended drawings.

[0131] Figure 1 shows a hydraulic torque wrench system according to an embodiment.

[0132] Figure 2 shows a time-pressure curve of three learning cycles according to an embodiment.

[0134] Figure 3 shows a time-pressure curve of a tightening operation according to an embodiment.

[0136] Figure 4 shows an example of a rundown trial cycle according to an example.

[0137] Figure 5 shows an example of a rehit trial cycle according to an example.

[0139] Figure 6 shows an example of a partial rehit trial cycle according to an example.

[0140] Figure 7 shows an enlargement of the curve of two of the cycles of the tightening operation of Figure 3.

[0142] Figure 8 shows a further enlargement of the curve of Figure 7.

[0144] Figure 9 shows an enlargement of the curve of three of the cycles of the tightening operation of Figure 3.

[0146] Like reference numerals refer to like elements throughout the description.

[0148] Detailed description of embodiments

[0150] A hydraulic torque wrench system 1 according to an embodiment will be described with reference to Figure 1. The system 1 comprises a wrench 2 hydraulically connected to a pump unit 4. The wrench 2 may comprise a piston 5 arranged to reciprocate inside a cylinder 8 as hydraulic pressure is applied via hydraulic connections 6, 7 on either side of the piston 5. The linear movement of the piston 5 may be translated via a pawl and ratchet mechanism (not shown) to a rotational movement of an output interface 9, such as a square or socket, of the wrench 2. The piston 2 advancing and retracting one time may be referred to as a cycle.

[0152] The pump unit 4 may comprise a single- or multistage pump. For example, the pump unit 4 may comprise a piston pump. For example, the piston pump may comprise several pistons driven by a motor. In a first stage of the pump, all of the pistons of the pump may be engaged so as to produce an as high hydraulic flow as possible. As the pressure in the system rises, the resistance for the motor increases and one piston at the time may therefore be disengaged so as to enable the motor to continue to drive the engaged piston(s), however at a lower flow rate. This may be referred to as the pump being operated in several stages (each stage corresponding to a unique number of pistons being engaged). Other types of single- or multistage pumps may also be envisaged.

[0154] The hydraulic torque wrench system 1 may be governed by a control device 3. The control device 3 may be configured to control auto-cycling of the system 1. The control device 3 may be integrated in the pump unit 4 (as illustrated in Figure 1) or provided as a separate device. The control device 3 may be a single device or distributed in several devices, such as cloud based. The control device 3 may comprise a memory and processing means as well as input and output interfaces for communicating with the pump unit 4.

[0156] The system 1 may further comprise one or more pressure sensors 11, 12. For example, one sensor 12 may be arranged to sense the pressure acting to advance the piston 5 in order to tighten the fastener. That pressure may be referred to as the advance pressure. Another sensor 11 maybe arranged to sense the pressure acting to retract the piston 5. That pressure may be referred to as the retract pressure. Data from the one or more sensors 11, 12 may be communicated to the control device 3.

[0158] The system 1 may further comprise a user interface, which e.g. may comprise a control pendant 10, enabling a user to actuate the system and/or input different parameters needed for the operation of the system 1.

[0159] Before starting a tightening operation, the user may enter a target torque, which may be received by the control device 3 and translated into a target pressure. Alternatively, the user may enter the target pressure directly. The target pressure is the threshold/limit, that the system aims at installing in the joint. The target pressure is the threshold/limit, that, when reached during advancement of the piston, may cause the control device 3 to end the tightening operation. Optionally, the user may further set a pressure relief valve (not shown) of the system 1 to a pressure slightly higher than the target pressure. While the target pressure may be referred to as a digital limit that the system 1 will aim for, the set pressure of pressure relief valve may be referred to as the physical limit that the pressure can reach.

[0161] Further, the user may initiate, e.g. by the actuation of a button on the pendant, a learning sequence for the system 1. The learning sequence may comprise one or more idling cycles for the purpose of calibrating the system 1. When the learning sequence is performed, the wrench 2 may preferably not be applied to any fastener, but rather just held in the air.

[0163] Figure 2 shows a time-pressure curve for the advance pressure (e.g. based on a signal from the sensor 12) of a learning sequence according to an embodiment. In this example, three learning cycles LC1-LC3 are performed and the target pressure Pmax is set to 400 bar. In the beginning of each cycle LC1-LC3, the pressure rises quickly before the piston starts to move. A bump 21 (small peak) may occur when the static friction between the piston and the cylinder is overcome and as the counter pressure on the other side of the piston is overcome to start advancing the piston. Then, a section 22 follows without any pressure increase as the play in the mechanical connection (such as in the pawl and ratchet mechanism) between the piston and the output interface is closed and as the piston then advances through the cylinder. Then, as the piston encounters the end of the advance stroke and therefore stops, the pressure starts to increase at point 23. A slight decrease in the gradient of the curve may occur at a point 24 if the pump unit shifts to its next stage (e.g. by disengagement of one of the pistons of the pump). The pressure continues to rise until the target pressure Pmax is reached, whereupon the control device may command retraction of the piston and the pressure therefore quicky falls down to (at least close to) zero again.

[0165] A gradient after the end-of stroke is reached, preferably for each cycle L1-L3, may be detected (and saved) by the control device 3. Then, the control device 3 may determine a reference end-of-stroke gradient based on the detected gradient(s). The reference end-of-stroke gradient may alternatively be referred to as a predetermined gradient threshold. The reference end-of-stroke gradient may e.g. correspond to, or be slightly lower than, the detected gradient. For example, the reference end-of-stroke gradient (or predetermined gradient threshold) may be comprised within the range of 70-90% of the detected gradient. If the end-of-stroke part of the curve (in the present example, from point 23 to Pmax) contains several gradients 25a, 25b, as a consequence of the pump shifting stage, each one of these gradients 25a, 25b may be detected and saved as a reference end-of-stroke gradient (or predetermined gradient threshold) associated with the respective pump stage.

[0167] According to embodiments, other reference values may (additionally or alternatively) be determined based on the one or more learning cycles L1-L3. For example, a predetermined retract time threshold may be set based on (such as equal to, or slightly lower than) a retract time for a complete stroke of the piston measured during the one or more learning cycles L1-L3. According to a further example, a predetermined advance time threshold may be set based on (such as equal to, or slightly lower than) an advance time for a complete stroke of the piston measured during the one or more learning cycles L1-L3.

[0169] It will be appreciated that any one of the reference values referred to above may be determined based on average values detected during a plurality of learning cycles L1-L3 in order to enhance the accuracy of the system. Further, it may be envisaged that any one of the reference values referred to above may be updated during subsequent cycling of a tightening operation.

[0170] Auto-cycling of the system for performing a tightening according to embodiments will now be described with reference to Figures 3 to 9.

[0172] Figure 3 is a pressure-time curve of a complete tightening according to an example. The tightening operation comprises a number of cycles TC0-TC4. In the present example, the target pressure is set to 300 bar and the pump unit will operate in two stages to get there.

[0174] According to an embodiment, what may be referred to as a trial cycle TCO may be performed as an initial cycle of the tightening. The purpose of the trial cycle may not just be to start the tightening operation, but also to estimate in which state the joint is in so as to be able to act accordingly, such as to estimate an exit pressure of the initial cycle or to terminate the tightening in case it is estimated as a rehit.

[0176] First, it may be awaited that any play in the mechanical connection between the piston and the output interface (such as in the pawl and ratchet mechanism) has started to close. As already explained with reference to the learning cycles in Figure 2, this event may be preceded by a small peak/bump in the pressure. This peak/bump may be detected by the control device and awaited to pass.

[0178] Then, the control device may monitor the pressure signal of the trial cycle in order to detect if any one of a number of conditions a)-c) occurs. These conditions a)-c) will in the following be described with reference to Figures 4 to 6.

[0180] Condition a) may occur when the joint is in a rundown phase, i.e. before the joint reaches snug, or at least first in a rundown phase and then, during the trial cycle, the joint may reach snug. An example of a pressure/time curve resulting from a rundown trial cycle is illustrated in Figure 4.

[0182] Once the initial peak/bump 61 has passed, a period 62 of steady pressure may follow as the play in the mechanical connection between the piston and the output interface closes and as the piston then advances and turns the fastener of the joint with very little resistance as the fastener runs freely. As the piston reaches the end of stroke, the gradient 67 of the pressure may get significantly steeper. In the example illustrated in Figure 4, at point 68, the pump shifts stage, whereupon the gradient 69 may get slightly reduced. Pressure may continue to be applied until the target pressure is reached.

[0184] Now, this condition may be detected by the control device by detecting that the gradient of the signal exceeds and stays over the predetermined gradient threshold (alternatively referred to as the reference end-of-stroke gradient) until the target pressure is reached. The time it takes to then retract the piston (in the trial cycle) back to its' start position (or at least to a given point close to the start position) for the next cycle may then be measured and compared with the predetermined retract time threshold. If that retract time is longer than the predetermined retract time threshold, the requisites for condition a) are fulfilled. That retract time being longer than the predetermined retract time threshold is an indication of that the piston has travelled a full advance stroke during the trial cycle.

[0186] In the example in Figure 4, gradient 67 may be detected as exceeding the predetermined gradient threshold associated with the first pump stage and gradient 69 may be detected as exceeding the predetermined gradient threshold associated with the second pump stage.

[0188] As condition a) is detected, the control device may estimate an exit pressure of the trial cycle (i.e. the pressure installed in the joint at the end of the advance stroke of the trial cycle). This may be done in different ways.

[0190] For example, the exit pressure may by default be estimated and set to zero, assuming that the joint is still in rundown phase at the end of the trial cycle. This may result in that, in embodiments wherein the exit pressure of the previous cycle is used as a bar for enabling end-of-stroke detection, end-of-stroke detection is enabled from start (i.e. from zero pressure), or at least from when the initial bump /peak has just passed, in the following cycle. However, if the fastener reached snug during the trial cycle, the actual exit pressure will not be zero, but some pressure will have been installed in the joint. The following cycle will then start with (after any play in the mechanical connection is closed) a steep pressure gradient, exceeding the reference end-of-stroke gradient, as the pressure builds to the level of the actual exit pressure of the trial cycle, at which the piston will start to move. As the gradient exceeds the reference end-of-stroke gradient, the control device may erroneously detect end-of-stroke and command retraction of the piston before it has even started to move. In order to detect this situation, the control device may measure the time it takes to reach end-ofstroke in that following cycle and compare that with the predetermined advance time threshold. If it turns out shorter than the predetermined advance time threshold, that may be an indication of that the piston has not moved (or just moved a little bit) before end-of-stroke was detected and, consequently, that the exit pressure of the trial cycle was in fact not zero. In order to remedy that, the control device may command the system to perform another (second) trial cycle to again detect if any one of conditions a)-c) occurs.

[0192] Alternatively, the exit pressure of the trial cycle may be estimated by actually analysing the pressure signal to find a point (or small range) where the pressure starts to rise with an end-of-stroke gradient. This may be done in the same manner as the exit pressure is estimated for the intermediate cycles of the tightening operation as will be described in more detail with reference to Figure 8 further below.

[0194] The tightening operation may then be continued by further auto-cycling.

[0196] Condition b) may occur when the joint is a rehit. That means that the target torque (or more) is already installed in the joint. An example of a pressure/time curve resulting from a rehit trial cycle is illustrated in Figure 5.

[0198] After the initial bump /peak 71 and closing of the play in the mechanical connection between the piston and the output interface (see flat section 72), the pressure may directly rise with a steep gradient 77 as the piston will not move any further. In the example illustrated in figure 5, the pump shifts stage at point 78, whereby the gradient 79 may become slightly lower. Pressure may be applied up to the target pressure, whereupon retraction may be commanded.

[0199] Now, this condition may be detected by the control device by detecting that the gradient of the signal exceeds and stays over the predetermined gradient threshold (alternatively referred to as the reference end-of-stroke gradient) until the target pressure is reached. The time it takes to then retract the piston (in the trial cycle) back to its' start position for the next cycle may then be measured and compared with the predetermined retract time threshold. If that time is shorter than the predetermined retract time threshold the requisites for condition b) are fulfilled. That time being shorter than the predetermined retract time threshold is an indication of that the piston has not travelled a full stroke, such as maybe barely moved at all.

[0201] In the example in Figure 5, gradient 77 may be detected as exceeding the predetermined gradient threshold associated with the first pump stage and gradient 79 may be detected as exceeding the predetermined gradient threshold associated with the second pump stage.

[0203] As condition b) is detected, the control device may command termination of the tightening as (at least) the target torque is already installed in the joint.

[0205] Condition c) may occur when the joint is a partial rehit. That means that some torque (but less than the target torque) is already installed in the joint. An example of a pressure/time curve resulting from a partial rehit trial cycle is illustrated in Figure 6.

[0207] After the initial bump /peak 81 and closing of the play in the mechanical connection between the piston and the output interface (see flat section 82), the pressure may first rise with a steep gradient until it reaches a level (see point 88) able to start to move the piston (and turn the fastener). As the piston (and fastener) starts to move, the pressure gradient 89 will be reduced to some extent.

[0209] Now, this condition may be detected by the control device by detecting that the gradient of the signal first exceeds the predetermined gradient threshold (see gradient 87) (alternatively referred to as the reference end-of-stroke gradient), and then falls below the predetermined gradient threshold (see gradient 89) for a predetermined amount of time (for the purpose of hysteresis in order to avoid small irregularities in the pressure signal to trigger condition c) detection).

[0210] As soon as condition c) is detected (e.g. at point 85), the control device may command termination of the trial cycle. Hence, the trial cycle may be terminated before the piston hits the end-of-stroke and the target pressure is reached. The control device may then estimate the exit pressure of the trial cycle. That is, the pressure that occurred just before the termination of the trial cycle may be estimated. This may e.g. be made by analysing the signal so as to identify the peak pressure of the trial cycle. Alternatively, level of the pressure read at time of detection of condition c) or read at the time of commanding retraction may be estimated and set as the exit pressure of the trial cycle.

[0212] The tightening operation may then be continued by further auto-cycling.

[0214] Optionally, instead of analysing the pressure signal directly, a derivative thereof may be analysed in order to detect any one of conditions a)-c).

[0216] The exit pressure estimated for the trial cycle may be used for further control of the tightening and/or for estimating one or more parameters of the tightening as will be described in more detail below.

[0218] Further auto-cycling according to an embodiment will now be describe with reference to Figure 7 showing an enlargement of two of the cycles TC2, TC3 of the curve of Figure 3.

[0220] Firstly, with reference to Figure 7, it will be explained how the pressure typically may change during one or more intermediate cycles of the tightening.

[0222] In similarity with the learning cycles L1-L3 and the trial cycle TCO, the start of a first one TC2 of the cycles, the pressure may first build while the piston stands still, up to a point 31, where a bump/small peak in the pressure may occur as the static friction between the piston and cylinder as well as any retract pressure are overcome, whereupon the piston starts to advance. The pressure may then level out for a short period 32 after the initial pressure peak 31 as the piston moves (but not the ratchet and output interface/fastener) to close the play between the ratchet and the pawl.

[0224] When the pressure is high enough to turn the output interface/fastener, the piston advances and applies torque to the joint (with start from point 33 in the illustrated example). This may be reflected in the pressure signal as a steady increase in the pressure. Hence, the pressure during this part of the curve (that is, after any gap in the pawl ratchet mechanism is closed and then as long as the piston advances) corresponds to the torque installed in the joint. As a full advance stroke is completed, the piston reaches it's dead end and stops moving, whereupon the pressure starts to build more rapidly. In the illustrated example, this occurs at point 34a. This may eventually be detected by the control device, (e.g. at point 35a), which in response may initiate retraction of the piston whereupon the pressure decreases down to the base line again.

[0226] The pressure at the end of the advancing stroke (that is, at point 34a) may be referred to as an exit pressure P<ex1 >of that cycle TC2. In other words, the exit pressure P<ex1 >may be the pressure occurring just when the piston reaches the end of its stroke. The exit pressure P<ex1 >is the pressure corresponding to the torque installed in the joint during the just performed cycle TC2.

[0228] Once the piston has been fully retracted, another cycle TC3 begins. In similarity with the previous cycle of the present example, the pressure may first build rapidly until the piston starts to move, then level out for a short period of time as the piston moves to close the gap between the pawl and the ratchet. Once the pawl and ratchet are engaged, the piston will again stop (at point 36) as the pressure is still too low to turn the fastener. The pressure now builds rapidly, see slope 37 of cycle TC3, until (approximately) the exit pressure P<ex1 >of the previous cycle is overcome (at point 38), whereupon installation of torque in the joint starts again and the gradient of the curve decreases to some extent. In the illustrated example, a small notch 40 occurs in the pressure during the advancing of the piston as a consequence of the pump unit shifting stage.

[0229] As the piston again reaches the end of its stroke (at point 34b), the pressure will drastically increase until retraction of the piston is again initiated, whereupon the pressure drops down to the base line again.

[0231] The system may continue to cycle until the pressure target Pmax is reached during advancing of the piston (with no end-of-stroke detected).

[0233] Embodiments of the present disclosure aims at making an accurate detection of when the piston reaches the end of its stroke, so the control device know when to command retraction of the piston. As can be seen in Figures 3 and 7, the pressure gradient may, at several occasions, be similar to the end-of-stroke gradient. For example, the very first steep pressure increase up to point 31 as well as the pressure increase 37 just before the fastener starts to turn (between points 36 and 38) are very similar (or even equal) to the pressure increase caused by the piston having reached the end of its stroke. To avoid detection of any of these slopes as an indication of the end of the stroke, embodiments of the present disclosure have been developed.

[0235] The control device of the system may be configured to analyse the pressure signal in order to estimate/detect the exit pressure P<ex1 >of a just performed cycle (which may be referred to as a first cycle, such as e.g. cycle TC2 in Figure 7). This analysis may e.g. be performed after the completion of the advance stroke of the piston, or even after completion of the whole cycle.

[0237] In order to identify the exit pressure, data points of the pressure signal may e.g. be analysed as will be described in the following with reference to Figure 8. Figure 8 shows an enlarged portion of the pressure signal at the transition from advancement of the piston to the end of the stroke (such as at point 34a of the first cycle TC2). The pressure signal may be analysed with respect to a plurality of its' data points 50, 51 (for the sake of simplicity and clarity, only three data points are marked with reference numbers 50 and 51 in Figure 8).

[0239] For each data point 50, a first interval h of data points 51 reaching forward from the data point 50 and a second interval h of data points 51 reaching backward from the data point 50 may be determined/selected. Then, an average gradient dPi, dP2 of each interval Ii, h may be estimated/calculated. For example, linear regression may be performed on the data points in each interval h, I2, wherein the gradient (e.g. k-value) of the resulting linear function may be determined as the average gradient dPi, dP2. Other methods of estimating average gradients dPi, dP2 of the intervals h, I2, may also be envisaged.

[0241] Further, the estimation of the value of the exit pressure may be based on the determined first and second average gradients dPi, dP2. For example, a score may be set for each data point 50 based on the difference between the first and second average gradients dPi, dP2. For example, the score may be e.g. the actual calculated difference, or a ratio between the average gradients dPi, dP2. Then, the data point 50 having the score indicating the largest difference between the first and second average gradients dPi, dP2 may be selected as the data point representing the exit pressure.

[0243] Preferably, data points 50 having one of the two determined average gradients being negative may be discarded and not used as basis for estimating the exit pressure. Thereby, the risk of small notches/irregularities, which are not uncommon in pressure signals, being erroneously detected as exit pressure is reduced.

[0245] Preferably, the analysis of the data points may start from behind, such as from a data point after advancement of the piston is completed in the first cycle TC2, such as after the pressure peak of the first cycle TC2 or even after the complete first cycle TC2. The analysis may then be made, for data point by data point, backwards in the signal. In other words, the estimation of the exit pressure may be made retrospectively, in contrast to during the cycle.

[0247] According to an embodiment, the exit pressure of the trial cycle TC0 may be estimated in a similar manner as described above, in particular in case condition c) is detected.

[0249] Optionally, instead of analysing the pressure signal directly, a derivative thereof may be analysed in order to estimate/detect the exit pressure. Then, for example, the average gradients will correspond to the y-intercept values of the linear functions derived by the linear regression instead of its k-value. For example, the derivative of the signal may be derived by means of appropriate filtering of the signal. Other types of filtering of the signal may also be envisaged.

[0250] The value of the estimated exit pressure P<ex1 >of the just performed cycle TC2 may be saved by the control device.

[0252] In the next cycle TC3, which may be referred to as a second cycle, monitoring of the pressure signal in order to detect the end of the advance stroke may not be started until the estimated exit pressure P<ex1 >of the previous cycle is exceeded. That is, not until point 38 in the graph in Figure 7 is reached, the monitoring may start. Before that, end-of-stroke detection is disabled. This means that the risk of the first steep slopes 37, 39 of the curve of the cycle TC3, where the pressure builds against a standing still piston, being mistakenly detected as the end-ofstroke gradient is reduced.

[0254] Preferably, the end-of-stroke detection may be performed, as from the exceeding of the exit pressure P<ex1 >of the previous cycle, continuously (or intermittently) during the cycle TC3. In this way, the end of stroke can be detected as soon as it occurs. For example, the gradient (derivative) of the pressure signal may be continuously (or intermittently) compared to the reference end-of-stroke gradient. When these matches, the end of the stroke may be detected. In response to the end-of-stroke detection, the control device may command retraction of the piston. In the example illustrated in Figure 7, the control device detects the end-of-stroke at point 35b. It may take a few milliseconds for the control device to identify the end-of-stroke and then to realize initiation of retraction of the stroke, as can be seen in Figure 7.

[0256] As long as the pressure has not reached the set pressure Pmax during advancement of the piston of the second cycle TC3 (indicating that the tightening operation is completed), the exit pressure P<ex2 >for the second cycle TC3 may be estimated/detected in the same manner as the exit pressure P<ex1 >for the first cycle TC2 and used as a bar for enabling end-of-stoke detection in the next (third) cycle (such as in cycle TC4 in Figure 3). In other words, the control device may continue to estimate the exit pressure of each cycle and use it as a threshold for enabling end-of-stroke detection in the next cycle, until the tightening operation is terminated.

[0258] According to embodiments of the present disclosure, an angle of the tightening may be estimated. These embodiments will be described with reference to figure 9 showing an enlargement of three of the cycles TC2-TC4 of the curve of Figure 3.

[0260] It may be set in the control device that a total angle Atot of the tightening should be counted as from when the pressure installed in the joint passes/exceeds a predetermined value P<startAngel>. In the present example, this predetermined pressure P<startAngel >is 50 bars.

[0262] The control device may be configured to detect, by analysing the pressure signal, the cycle in which this predetermined pressure P<startAngel >is passed. In the present example, that is the first cycle TC2.

[0264] The control device of the system may be further configured to analyse the pressure signal in order to estimate an entry pressure P<ent1 >and an exit pressure P<ex1 >of the first cycle TC2. The entry pressure P<ent1 >vs the pressure at which torque is started to be installed in the joint, while the exit pressure P<ex1 >is the pressure at which installation of torque stops.

[0266] The exit pressure P<ex1 >may be estimated/determined by finding the bending of the curve resulting from the piston hitting the end of the stroke, e.g. in the same manner as described above with reference to Figure 8.

[0268] The entry pressure P<ent1 >may be estimated/determined in a similar way by finding the bending of the curve resulting from the piston starting to move, such as based on the same principles as those described with reference to Figure 8.

[0269] In some cases, the entry pressure of a particular cycle may be slightly lower than the exit pressure of the previous cycle due to losses in the system. Hence, estimating the actual entry pressure of a cycle (such as in any of the ways described above) may result in a more accurate estimation of the angle than if the entry pressure of a cycle is just automatically set to equal the exit pressure of the previous cycle. Indeed, the exit pressure of the previous cycle may be utilized in the determination of the entry pressure, e.g. by using it (or any predetermined percentage thereof) as a starting point for analysis of the signal. For example, stepping through the signal, data point by data point, for determining the above describe (with reference to Figure 8) score, may be started at a predetermined pressure above the exit pressure of the previous cycle.

[0271] Further, an angle A<startcycle >of the first cycle TC2, as counted from the passing of the predetermined pressure P<startAngle>, may be estimated. This estimation may be made based on a (determined) relation/ratio between the pressure range: P<ent1 >to P<ex1 >and the pressure range: P<startAngle >to P<ex1 >and a predetermined full stroke angle Α<full>. For example, the following equation (or any of its' equivalents) may be used:

[0274]

[0277] Alternatively (or as a complement), the estimation of the angle A<startcycle >of the first cycle TC2 may be based on the relation/ratio between the time range: T<ent1 >to T<ex1 >and the time range: T<startAngle >to T<ex1 >and a predetermined full stroke angle Α<full>. Here, the times T<ent1>, T<startAngle >, and T<ex1 >are the points in time when the pressures P<ent1>, P<startAngle >, and P<ex1 >are passed.

[0279] For example, the following equation (or any of its' equivalents) may be used:

[0282]

[0285] Further, a total angle of the tightening may be estimated by adding to the angle A<startcycle >of the first cycle TC2, one predetermined angle Α<full >for each subsequent full stroke cycle performed. In the present example, one full stroke cycle TC3 is performed after the first cycle TCI, so one predetermined angle A<full >is added to the angle A<startcycle >of the first cycle TC2.

[0287] Any cycle occurring between the first and the last cycles may be detected as a full stroke cycle.

[0289] Further, an angle A<endCycle >of the last cycle TC4 of the tightening may be estimated. The last cycle TC4 is the cycle in which the maximum pressure P<max >is reached during advance of the piston (i.e. before end of stroke is detected).

[0290] The angle A<endCycle >of the last cycle TC4 may be estimated in different ways. Preferably, it may be estimated by first estimating the entry and exit pressures P<entEnd>, P<exEnd >of the last cycle TC4 as well as the entry and exit pressures P<entFull>, P<exFull >of at least one preceding full stroke cycle TC3. Then, the angle A<endCycle >of the last cycle TC4 may be estimated based on a (determined) relation/ratio between the pressure range: P<entEnd >to P<exEnd >and the pressure range: P<entFull >to P<exFull >and on the predetermined angle Α<full>.

[0292] For example, the following equation (or any of its' equivalents) may be used:

[0295]

[0298] Alternatively (or as a complement), the angle A<endCycle >of the last cycle TC4 may be estimated based on a (determined) relation/ratio between the time range: T<entEnd >to T<exEnd >and the time range: T<entFull >to T<exFull >and on the predetermined angle Α<full>. Here, the times T<entEnd>, T<exEnd>, T<exFull >and T<entFull >are the points in time when the pressures P<entEnd>, P<exEnd>, P<exFull >and P<entFull >are passed.

[0300] For example, the following equation (or any of its' equivalents) may be used:

[0303]

[0304] In case the last cycle is preceded by several full stroke cycles, average pressure/time ranges between the entry and exit pressures of two or more of these full cycles may be used as a basis for the angel estimation. Alternatively, only the time/pressure range between the entry and exit pressures of the just preceding full stroke cycle may be used.

[0306] In case the time range ratio approach is to be used, it may be beneficial to check whether there is an indication that the average pressure rate dPl between the entry and exit pressures P<entFull>, P<exFull >of the at least one preceding full stroke cycle TC3 deviates (more than to a predetermined extent) from the average pressure rate dP2 between the entry and exit pressures P<entEnd>, P<exEnd >of the last cycle TC4. If that is the case, the pump may have shifted stage during any one of these two cycles. In the present example, the pump has shifted stage during the full stroke cycle TC3 at point 40.

[0308] In case such a deviation is detected, a compensation (or correction) factor/function may be added to the estimation of the angle of the last cycle. The compensation factor/function may be adapted so as to compensate for the mismatch of the gradients in the calculation of the angle.

[0310] Alternatively, in case such deviation is detected, the ratio of the retract times of the last cycle TC4 and of at least one of the preceding full stroke cycles TC3 may be used as a basis for determining the angle A<endCycle >of the last cycle TC4.

[0312] The calculated/estimated total angle A<tot >(i.e. A<startcycle >+ ( Α<full >x number of intermediate full stroke cycles) A<endCycle>) may then be saved, e.g. together with other parameters of the tightening.

[0314] It will be appreciated that any of the above described embodiments may be performed based on a derivative of the pressure signal.

[0316] The person skilled in the art realizes that the present invention by no means is limited to the embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims.

[0318] Additionally, variations to the disclosed embodiments can be understood and effected by the skilled person in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

Claims

1. A method, performed by a control device (3), for controlling autocycling of a hydraulic tightening system (1), wherein the autocycling comprises regulating a hydraulic pressure in the system (1) to alternately reciprocate a piston (5) in the system (1) to drive an output interface (9) of the system (1) for tightening a joint, wherein the method comprises:

- setting a target pressure (Pmax) for a tightening; and

- receiving a signal indicating said hydraulic pressure; wherein the method is characterized in that it further comprises:

- detecting whether any of the following conditions occur during a test cycle (TCO), after any looseness in the mechanical connection between the piston (5) and the output interface has begun to close, the test cycle (TCO) being performed as an initial cycle of tightening:

a) the signal gradient exceeds and remains above a predetermined gradient threshold until the target pressure (P<max>) is reached and

the time it takes to then retract the piston (5) is above a predetermined threshold value for retraction time,

b) the gradient of the signal exceeds and remains above the predetermined gradient threshold until the target pressure (P<max>) is reached and

the time it takes to then retract the piston (5) is below the predetermined threshold value for retraction time, and

c) the gradient of the signal first exceeds the predetermined gradient threshold and then drops and remains below it for a predetermined period of time;

wherein the method comprises:

- in response to detecting conditions a) or c), estimating an exit pressure for the performed test cycle (TCO) and continuing the tightening by further autocycling; and/or

- in response to detection of condition b), command termination of the tightening.

2. The method of claim 1, wherein estimating the exit pressure in response to condition a) comprises, for each of a plurality of data points (50, 51) of the signal:

- for a first interval (I<1>) extending forward from the data point (50), estimate a first average gradient (dP<1>); and for a second interval (I<2>) extending backward from the data point (50), estimate a second average gradient (dP<2>);

wherein the test cycle exit pressure (TCO) is estimated based on the estimated first and second average gradients (dP<1>, dP<2>).

3. The method of claim 1, further comprising, in response to detecting condition a),

estimating the exit pressure for the test cycle (TCO) to zero; in the next cycle, estimating the time it takes until the end of the stroke is reached; and

in response to the estimated time being below a predetermined threshold for forward movement time, performing another test cycle and detecting whether any of conditions a) -c) occur.

4. A method according to any one of the preceding claims, further comprising:

- in response to detecting condition c), commanding retraction of the piston (5) in the test cycle (TCO) without waiting for the end of the stroke to be reached; and estimating an exit pressure for the performed test cycle (TCO) based on the value of the pressure just before the retraction of the piston (5) in the test cycle (TCO).

5. A method according to any one of the preceding claims, further comprising using the estimated test cycle exit pressure (TCO) as an input value for continued control of the tightening.

6. The method of any preceding claim, further comprising enabling detection of the end of the stroke in the next cycle only after the pressure in the next cycle has exceeded the estimated test cycle exit pressure (TCO).

7. The method of any preceding claim, further comprising using the estimated test cycle exit pressure (TCO) as an input value for estimating a parameter for the tightening.

8. The method of claim 7, wherein the estimated tightening parameter is an angle of tightening.

9. Method according to any one of the preceding claims, wherein a pump unit (4) in the system (1) arranged to provide said hydraulic pressure is arranged to operate in at least two different stages, and wherein each stage is associated with a respective predetermined gradient threshold value used to detect conditions a) -c).

10. A method according to any one of the preceding claims, wherein said predetermined gradient threshold value has been set based on a pressure gradient that occurs when forward pressure is applied and the piston (5) is not moving.

11. A method according to any one of the preceding claims, wherein said predetermined threshold value for retraction time has been set based on a retraction time for a complete stroke of the piston (5).

12. Method according to any one of the preceding claims, wherein the detection of conditions a) -c) is done based on a derivative of the signal.

13. Control device (3) configured to perform the method according to any one of the preceding claims.

14. Hydraulic tightening system (1) comprising a control device (3) according to claim 13.

15. A computer program comprising instructions which, when the program is executed by a computer, causes the computer to perform the method according to any one of claims 1-12.

16. A computer-readable storage medium comprising instructions that, when the program is executed by a computer, cause the computer to perform the method according to any one of claims 1-12.