AU2018102092A4 - Self-powered device monitoring system - Google Patents

Abstract

The invention provides a monitoring system, the system comprising: a plurality of monitor units (10) adapted to communicate over a wireless link (16) with a data receiving computer and attachable to a corresponding plurality of devices to be monitored, at least one of which is a reciprocating or rotating device. Each monitor unit (10) comprises: an energy generator (47) adapted to generate electrical energy from vibration of the monitor unit (10) when attached to the corresponding device to be monitored during normal operation thereof; a battery or capacitor (31) adapted to store the generated electrical energy; a one or more condition measurement sensors (42,45) adapted to perform one or more condition measurements of the corresponding device to be monitored; a non-volatile memory (43) adapted to store data from the measurement operation; and one or more processors (41,44,48) programmed to direct the performance of the one or more condition measurements, the data storage in the non-volatile memory (43), data retrieval from the non-volatile memory (43) and data transmission to the data receiving computer of data retrieved from the non-volatile memory (43) or derived therefrom. u -b uu 0 a' V) br_ Lta q~00

Description

TITLE [0001] SELF-POWERED DEVICE MONITORING SYSTEM

FIELD [0002] The present invention relates to vibration and other monitoring of pumps and air-cooled heat exchangers which vibrate during normal operation, in particular but not limited to an apparatus enabling cost-effective installation in a large-scale industrial environment.

BACKGROUND [0003] A table of abbreviations is provided at the end of this description. The inlet pressure of a typical LNG plant is around 60-70 bar(g) and approximately atmospheric temperature. Once liquefied however, the temperature is -161°C and atmospheric pressure. Whilst some heat is lost through the expansion of the gas to atmospheric pressure (JT effect), this contribution is negligible.

[0004] The bulk of heat rejection in all natural gas liquefaction processes - whether they be C3-MR, DMR, Cascadetype etc. is via a refrigeration loop. The heat transfer can be enacted by either a heat transfer fluid (e.g. sea water), or by air using an air-cooled heat exchanger (ACHE), the latter being by far the most popular approach.

[0005] The biggest challenge posed by ACHEs in a liquefaction environment is managing the sheer number of units; a typical LNG plant - the Karratha Gas Plant in Western Australia for instance

2018102092 19 Dec 2018 has in excess of 1000 ACHE units. Whilst constant vibration, acoustic and/or ultrasonic monitoring of the ACHE units would likely enhance machine availability (and hence enhance production), it is fundamentally impractical for two reasons. Firstly, monitoring every single ACHE unit in the typical (cabled) fashion would require a level of additional infrastructure that would be exceedingly intrusive and most likely cost prohibitive. Secondly, managing battery change-outs of battery powered units would be inordinately onerous.

[0006] There is therefore a need to provide an improved vibration monitoring system for multiple device units which overcomes these problems .

SUMMARY OF THE INVENTION [0007] The inventor has conceived that by virtue of the fact that heat exchanger and other rotating or reciprocating device monitoring necessarily involves vibration, the coupling of vibration energy harvesting with ultra-low-power electronics enables provision of a self-powered system to obtain the vibration or other condition measurement data and wirelessly transmit the data. The monitoring devices can be simple and entirely self-powered, mitigating the previous impracticalities and making per-device vibration monitoring feasible and economically justified.

[0008] In accordance with a broad aspect of the invention there is provided a monitoring system, the system comprising:

a plurality of monitor units adapted to communicate over a wireless network with a data receiving computer and attachable to a corresponding plurality of devices to be monitored;

2018102092 19 Dec 2018 wherein each monitor unit comprises:

an energy generator adapted to generate electrical energy from vibration of the monitor unit when attached to the corresponding device to be monitored during normal operation thereof;

a battery or capacitor adapted to store the generated electrical energy;

one or more condition measurement sensors adapted to perform one or more condition measurements of the corresponding device to be monitored;

a non-volatile memory adapted to store data from the measurement operation; and one or more processors programmed to direct the performance of the one or more condition measurements, the data storage in the non-volatile memory, data retrieval from the non-volatile memory and data transmission to the data receiving computer of data retrieved from the non-volatile memory or derived therefrom; and wherein the reciprocating or rotating device is a pump or air-cooled heat exchanger.

[0009] In one embodiment, the one or more condition measurement sensors of at least one of the monitor units includes a vibration sensor.

[0010] In one embodiment, the one or more condition measurement sensors of at least one of the monitor units includes an acoustic and/or ultrasonic sensor.

[0011] In one embodiment, the one or more processors are programmed to vary the direction of any one or more of: the one or more condition measurements, the data storage, the data

2018102092 19 Dec 2018 retrieval, and the data transmission; the variation being based on a measurement indicative of an amount of available energy in the battery or capacitor. The one or more processors may vary the direction by incorporating or varying a length of one or more sleep periods during which power consumption is lowered to below an energy accumulation rate to allow re-accumulation of energy in the battery or capacitor. The sleep periods may comprise sleep periods interposed between one or more of the measurement, the data storage, the data retrieval and the data transmission. The sleep periods comprise sleep periods interposed during the data transmission. The sleep periods may comprise a sleep period between repeat performance of complete cycles of the measurement, the data storage, the data retrieval and the data transmission. The one or more processors may vary the direction by varying an amount of data collected during the measurement. In one embodiment, the one or more processors are adapted to determine the measurement indicative of the amount of available energy in the battery or capacitor at least in part by performing a calibration process whereby the one or more processors: operate the system in a quantified high-power mode (QHPM); and measure an elapsed time of operation in QHPM before an energy depletion state is reached, the elapsed time thereby providing the measurement indicative of the amount of available energy. The one or more processors may be adapted to determine the energy depletion state by monitoring for loss of a power good signal originating from an energy harvesting power supply associated with the energy generator and the battery or capacitor.

[0012] In one embodiment, at least one of the monitor units is attached to a fan bearing cartridge of an air-cooled heat exchanger.

2018102092 19 Dec 2018 [0013] In one embodiment, the energy generator comprises a piezoelectric element with weights attached to a free end thereof.

BRIEF DESCRIPTION OF DRAWINGS [0014] Figure 1 shows an air cooler arrangement of an induceddraft fin-fan air-cooled heat exchanger.

[0015] Figure 2 shows external appearance of a prototype design for an embodiment of the vibration monitoring device of the invention.

[0016] Figure 3 shows is an exploded view of the embodiment of Figure 2.

[0017] Figure 4 shows architecture and identification of the electronic components of the embodiment.

[0018] Figure 5 shows an overall high-level state machine for the operation of the system of the embodiment.

[0019] Figure 6 shows a state-level machine for the sampling process of the system of the embodiment.

[0020] Figure 7 shows a state-level machine for the transmission process of the system of embodiment.

[0021] Figure 8 shows a flowchart of the power calibration process of the system of the embodiment.

2018102092 19 Dec 2018

DETAILED DESCRIPTION OF EMBODIMENTS [0022] An embodiment of the current invention will now be described, directed towards air cooled heat exchangers.

[0023] The mounting of vibration sensors to rotating or reciprocating equipment is a well-established practice. A typical transducer will couple to the vibration source either by directly threading into a tapped hole in a suitable location, or threading a magnetic base into the tapped hole and coupling by magnetic force. As the sensor unit is expected to be eccentric in design as described below in this embodiment, the magnetic base option is preferred.

[0024] Referring now to Figure 1 there is shown an air cooler arrangement of an induced-draft fin-fan air cooled heat exchanger. This is the type typically used in an LNG environment. A motor 1 turns a drive pulley 2 which in turn connects to fan belt 3 which turns fan pulley 4. The most frequent failure modes are upper and/or lower drive bearing failure. An appropriate location for vibration sensing is therefore the upper bearing cartridge 5 or the lower bearing cartridge 6. The cartridges 5, 6 themselves are often pre-tapped with a suitably sized hole; for example the boss 7 for the upper bearing cartridge is visible in Figure 1.

[0025] Whilst mounting directly to one of the bearing cartridges 5, 6 provides the most accurate depiction of the system's vibration characteristic, the sensor could be mounted to any component that is (rigidly) physically coupled to the vibration source, such as structural steel work, provided that there is sufficient physical space for the unit. Since the further from the vibration source, the lesser the harvestable

2018102092 19 Dec 2018 vibration energy, direct coupling to the vibration source is preferable.

[0026] The upper bearing cartridge 5 can be difficult to access in a brownfields situation, (it typically requires a machine shutdown, guard removal and possibly pulley removal), and for this reason the lower bearing cartridge 6 is the a more convenient installation location than the upper bearing cartridge 5. Fortunately, the lower bearing also accommodates both the axial and thrust loads and furthermore receives the belt pull load due to fan belt 3 tension, and for these reasons the lower bearing cartridge 6 is also the better measurement location for vibration monitoring.

[0027] The prototype design for the monitor unit of the invention is illustrated in Figure 2 and 3. In the system there are a plurality of similar monitor units each attachable to a corresponding plurality of devices to be monitored, being this case air cooled heat exchangers, and each being adapted to communicate over a wireless link with a data receiving computer. The data receiving computer does not form part of the invention and as is known in the prior art could come in a number of forms, could be located locally or remotely and could affect the communication of the wireless link by a combination of means, including the Internet or a wired network to a wireless router.

[0028] Referring to Figure 2, the monitor unit incorporates a housing 10 comprising rectangular container 11 of dimensions of length 79 mm width 42 mm at height 27 mm and having a lid 12. Housing 10 houses the electronic and mechanical componentry to be described below. Extending from a base 13 of housing 10 is a wireless communications antenna 16 and a cylindrical magnetic

2018102092 19 Dec 2018 coupling element 14 of height approximately 20 mm in diameter approximately 25 mm. Cylindrical magnetic coupling element 14 is adapted to be held in place by magnetic attraction from magnetic coupling surface 15 to the magnetic base screwed into the measurement point described above. The magnetic coupling element 14 or the magnetic base itself may contain a permanent magnet.

[0029] Figure 3 is an exploded view of housing 10 and its contents. Rectangular container 11 houses printed circuit board 20, to which are attached an energy management system and processor, communications, accelerometer, memory and energy storage device in the form of supercapacitor 21. Printed circuit board 20 is secured by securing screws 22 to screw posts 23. Wireless communications antenna 16 connects to printed circuit board 25 via an SMA connection (not shown). Printed circuit board 20 comprises a slot 27 which cooperates with a lower clamp component 28 projecting through slot 27 when printed circuit board 20 is installed. Lower clamp component 28 and upper clamp component 34 clamp piezoelectric element 31 from below and above to provide a fixed point for oscillation of piezoelectric element 31. Securing bolts and terminals 24, 25, 30 secure and provide electrical connection for piezoelectric element 31. Piezoelectric element 31 (Mide, www.mide.com, model PPA-1011) is a layered flexible element comprising piezoelectric materials and circuit elements which feed a generated voltage to terminals 24, 25 which are in turn connected to printed circuit board 20. Attachable to a free end of piezoelectric element 31 by securing screws 33 are weights 32 which are chosen to provide a frequency of oscillation of the free end resonance matched to a frequency at which maximum power is produced from vibration during operation of the ACHE. Typically, in an installation where each ACHE is identical, the weights 32 will be the same for each ACHE

2018102092 19 Dec 2018 but may be varied if needed.

[0030] Lid 12 is secured by securing screws 29.

[0031] In Figure 4, architecture and identification of the electronic components contained within printed circuit board 20 is depicted. Energy harvesting power supply 40 (Linear, www.linear.com, model LTC3588-1) harvests energy from piezoelectric element 31 and stores the energy in 6.8 mF super capacitor 47 (AVX BZ013A144ZSB). If the energy harvesting power supply 40 is capable of harvesting from more than one piezoelectric element, then ideally two piezoelectric elements are incorporated, tuned by a selection of weight selection to each of two major frequencies that have the capacity to deliver the highest power. Energy harvesting power supply 40 is configured to supply 3.6 VDC to voltage supply rail Vcc.

[0032] The broad aspect of the invention provides one or more processors programmed to direct the measurement by the one or more condition measurement sensors, the data storage in the nonvolatile memory, data retrieval from the non-volatile memory and data transmission to the data receiving computer of data retrieved from the non-volatile memory or derived therefrom. In this embodiment, the one or more processors is shared between host processor and communications unit 41(Linear, www.linear.com, model LTC5800-IPM - SmartMesh IP Node 2.4GHz 802.15.4e Wireless Mote-on-Chip), power controller 48 (Texas Instruments TPS22860DBVR) and field programmable gate array (FPGA) 44 (Microsemi AGLN060V2-VQ100). Host processor and communications unit 41 boots when rail voltage on supply rail Vcc exceeds a minimum requirement. Accelerometer 42 (Analog Devices, www.analog.com, model ADXL345) provides acceleration

2018102092 19 Dec 2018 measurements as a measure of vibration. Acoustic-ultrasonic sensor 45 (Knowles SPH0641LU4H-1) provides measurements of acoustic or ultrasonic output of the heat exchanger, adapted measure in either the acoustic or ultrasonic range under selectable control by FPGA 44. Acoustic-ultrasonic sensor 45 may be positioned such that it faces the main fan drive shaft. Data memory 1 Mbit FRAM module 43 (Fujitsu, www.fujitsu.com, model MB85RS1MT) is used to store data collected by accelerometer 42 and acoustic-ultrasonic sensor 45. Host processor and communications unit 41, FPGA 44, accelerometer 42 and data memory 43 communicate by an SPI or I²C bus, with pull-up resistors 46 required in the case of an I²C bus. As the I²C bus is an open drain configuration, this will result in current draw when the bus state is SDA=0 and/or SCL=0. The idle state is SDA=Vdd, SCL=Vdd, so there should be no parasitic draw when idle.

[0033] The FPGA 44 performs the coordination of signals between the sensors 42,45 and the data memory FRAM 43; the role of the host processor 41 is simply to command the FPGA 44 to perform the measurement. At measurement time:

1. The host processor 41 powers up the devices required to perform the sampling by signalling power controller 48 to energise switchable voltage rail Vcc_SW

2. The host processor 41 signals the FPGA 44 to perform a measurement (either vibration or acoustic)

3. The FPGA 44 performs any relevant initialisation and gates the sensor and FRAM clocks such that the data from the sensor (whether direct-connected or as an I2C / SPI slave) is clocked directly into the FRAM (also an SPI slave). This frees the host processor 41 from requesting, and caching data from sensors 42, 45 and caching that data in FRAM 43. Furthermore, the host processor power consumption can be minimised as sampling frequency is entirely independent of the host processor

2018102092 19 Dec 2018 clock speed.

At transmission time, the FPGA 44 acts as a pure relay for the SPI control signals to allow the host processor (SPI master) to talk directly to the FRAM 43(SPI slave) to grab a data chunk for transmission.

[0034] The above selection of components provides a system capable of sampling from accelerometer 42 and storing in data memory 43 according to the following conditions: IEC Zone 1 Hazardous Areas (IEC-Ex Zone lcertification pending), acceleration up to +- 8g, frequency resolution 0 to 40 kHz (i.e. sampling 0 to 80kHz), sampling duration at least 5 seconds.

[0035] Figure 5 depicts a simplified high-level state machine for the operation of the system, which is mainly selfexplanatory. Briefly, if rail voltage 44 exceeds a minimum requirement, host processor and communications unit 41 commences a booting routine and if successful starts a sample timer and enters an idle loop waiting until expiry of the sample timer. At expiry of the sample timer, a sampling process from the accelerometer 42 begins and once completed is transmitted over a communications unit 41. Control then passes back to the idle loop with the sample timer reset to repeat the process. In different embodiments, at any time during the sampling and transmitting potentially the system may power down if required due to power constraints for a period sufficient to allow reaccumulation of energy from the energy generator, as described below.

[0036] Figure 6 depicts a state-level machine for an example of the sampling process. Briefly, after expiry of the abovementioned sample timer, host processor and communications unit

2018102092 19 Dec 2018 attempts to wake up the FPGA 44, the relevant sensor for next measurement (accelerometer 42 or Acoustic-ultrasonic sensor 45) and data memory FRAM 43. Upon success, host processor 41 instructs the relevant sensor to sample and store data in data memory 43 until completion of a sampling cycle, after which all unnecessary peripherals such as accelerometer 42, data memory 43 and at least a communications subunit of host processor and communications unit 41, are placed in the lowest possible power mode for an idle period to accumulate more energy, pending eventual transmission of the data.

[0037] Figure 7 depicts a state machine for an example of the transmission process. As is evident from state machine, a strategy of powering up and down the FRAM data memory 43 is utilised in between transmissions of chunks of memory, in order to allow re-accumulation of energy from the energy generator, as described above.

[0038] For the purposes of calculating power budgeting, an example mode of operation (in this case, sampling vibration) is:

l.Sit idle (i.e. mote and peripherals in sleep mode) until the charge store is adequately charged.

2. When charged, sample and store samples in FRAM:

a. 16-bit samples are collected in X, Y and Z planes at the Nyquist frequency of 3 kHz to support 1.5 kHz bandwidth.

b. Sample period is 5 s.

Therefore, total storage required is 3000

2018102092 19 Dec 2018 samples/sec x 48 bits/sample x 5 sec = 720 kbit.

d. Nominally 1 Mbit of NV storage (i.e. FRAM) is required.

3. Place accelerometer in sleep mode.

4. At an arbitrary interval that allows the full sample to be transmitted within the sample capture period (accounting for net link reliability), transmit 90 bytes (the maximum SmartMesh packet size) until all data is transmitted.

5. The energy storage device must be capable of delivering sufficient charge to enact the most onerous single operation (e.g. sampling, transmit a packet).

[0039] To calculate power availability, based on a typical example with the following operational data: Motor power: 37 kW; Speed: 1406.2 RPM; Load: 100%; an FFT applied to a vibration sample revealed a primary peak of 0.25 Gs at 23.6 Hz. A device such as the Mide PPA-1011 will develop approximately 2.5 mW of average power at 23 Hz, with a pure sinusoid.

[0040] Power consumption data may be obtained from manufacturer data of the components listed above and a power estimate obtained through simulation of the system and its various operating states.

[0041] Such an analysis is performed on the basis of a predicted assumed power availability from the piezoelectric element. Despite matching the weights 32 to the peak resonance characteristic of the system, there will be some variation in performance of the piezoelectric element, the microelectronics,

2018102092 19 Dec 2018 and even the characteristic of the machine to which the sensor is attached.

[0042] Should the material circumstance yield a significantly wide variation in performance, there is the risk that that assumed safety margin will not provide sufficient time to accumulate charge between operations, leading to brown-out prior to completion of a particular operation.

[0043] To address the issue of power availability variation, the device of this embodiment includes a calibration step. Core to this calibration step is a Quantified High-Power Mode (QHPM).

[0044] QHPM in its simplest form is a fully deterministic power consumption mode that significantly exceeds the maximum possible power generation capacity of the piezoelectric element. Depending on the actual implementation it may or may not include enviro-corrective features such as temperature compensation. It could be realised in any number of ways including, but not limited to placing peripheral components in a high-power mode or having a dedicated bleed resistor coupled to the GPIO via a FET.

[0045] A flowchart of the power calibration process is detailed in Figure 8. The calibration process expends a burst of a standardised amount of power in QHPM for a predetermined period, checks whether the power is exhausted from the energy storage device 21, and if not expends more bursts until the stored energy is exhausted. As soon as the stored energy is exhausted, the value of the total elapsed time in QHPM mode is stored in the non-volatile memory 43. Detection of exhaustion of the stored energy is in this embodiment ascertained by monitoring of a power good Digital output from the energy harvesting power

2018102092 19 Dec 2018 supply 40, which indicates exhaustion or impending exhaustion by setting the digital output to 0. The stored value of the elapsed time establishes a total amount of time t_elapsed corresponding to the accumulated bursts which elapses at the QHPM for the available stored energy to be exhausted, which directly correlates with the actual power output and therefore a measure of available energy from the piezoelectric element 31. The system then powers down and awaits a rebooting action to occur when sufficient power is available. After reboot, the value of t_elapsed is then read from the non-volatile memory 43, and using a predetermined look up table of t_elapsed versus delay, the system can then determine the appropriate delay between data acquisition periods or other multiple delay parameters or other variation. In this way, the host processor is programmed to vary the direction of how the various processes proceed thereby ensuring that sufficient energy is accumulated between the periods to allow successful operation of the system. Because the system of this embodiment automatically adjusts as a result of a determination of how much energy is available, reliable operation in the face of unpredictable power is enabled.

[0046] A simulator or a combination of a simulator and experimental data can be used to develop the look up table in the design phase.

[0047] Persons skilled in the art will appreciate that many variations may be made to the invention without departing from the scope of the invention, which is determined from the broadest scope and claims.

[0048] In the claims which follow and in the preceding description of the invention, except where the context requires

2018102092 19 Dec 2018 otherwise due to express language or necessary implication, the word comprise or variations such as comprises or comprising is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention. Further, any method steps recited in the claims are not necessarily intended to be performed temporally in the sequence written, or to be performed without pause once started, unless the context requires it.

[0049]

It is to be understood that, if any prior art

publication is referred to herein, such reference does not constitute an admission that the publication forms a part of the common general knowledge in the art, in Australia or any other country.

[0050]

The following abbreviations are used in the description

above.

ACHE	Air Cooled Heat Exchanger
C3-MR	Propane - Mixed Refrigerant
DC	Direct Current
DMR	Dual Mixed Refrigerant
FFT	Fast Fourier Transform
FET	Field Effect Transistor
FPGA	Field programmable Gate array
FRAM	Ferro-magnetic Random Access Memory
GPIO	General Purpose Input / Output
IEC	International Electrotechnical Commission
JT	Joule-Thomson
LNG	Liquefied Natural Gas
MR	Mixed Refrigerant

2018102092 19 Dec 2018

NV	Non-Volatile
PCB	Printed Circuit Board
RAM	Random Access Memory
RPM	Revolutions Per Minute
SMA	SubMiniature version A
UNF	Uniform National Fine

Claims (5)

1. A monitoring system, the system comprising:

a plurality of monitor units adapted to communicate over a wireless link with a data receiving computer and attachable to a corresponding plurality of devices to be monitored;

wherein each monitor unit comprises:

an energy generator adapted to generate electrical energy from vibration of the monitor unit when attached to the corresponding device to be monitored during normal operation thereof;

a battery or capacitor adapted to store the generated electrical energy;

a one or more condition measurement sensors adapted to perform one or more condition measurements of the corresponding device to be monitored;

a non-volatile memory adapted to store data from the measurement operation; and one or more processors programmed to direct the performance of the one or more condition measurements, the data storage in the non-volatile memory, data retrieval from the non-volatile memory and data transmission to the data receiving computer of data retrieved from the non-volatile memory or derived therefrom; and wherein at least one of the devices to be monitored is a reciprocating or rotating device.

2. The monitoring system of claim 1, wherein:

the one or more condition measurement sensors of at least one of the monitor units includes a vibration sensor, an acoustic sensor and/or an ultrasonic sensor; and the one or more processors are programmed to vary the direction of any one or more of: the one or more condition measurements, the data storage, the data retrieval, and the data transmission; the variation being based on a measurement indicative of an amount of available energy in the battery or capacitor.

2018102092 19 Dec 2018

3. The monitoring system of claim 2 wherein the one or more processors vary the direction by varying an amount of data collected during the measurement.

4. The monitoring system of claim 2 or claim 3, wherein the one or more processors are adapted to determine the measurement indicative of the amount of available energy in the battery or capacitor at least in part by performing a calibration process whereby the host processor:

operates the system in a quantified high-power mode (QHPM); and measures an elapsed time of operation in QHPM before an energy depletion state is reached, the elapsed time thereby providing the measurement indicative of the amount of available energy.

5. The monitoring system of any one of claims 1 to 4, wherein the reciprocating or rotating device is a pump or air-cooled heat exchanger.