Field of the invention

[0001] The present invention concerns a device and a system for cleaning a nozzle.

Prior and related art

[0002] We consider a nozzle for injecting an additive fluid into a flow of a main fluid, where one or both fluids are gases or liquids. The main fluid contains contaminants that deposits on the nozzle walls, in particular in a relatively narrow nozzle opening and especially when the additive fluid is supplied at low pressure.

[0003] A first example regards a mixer for adding inert gas, e.g. N ₂ or C0 ₂ , to a main flow of liquid in a bubble separator for use in the oil and gas industry. The main flow typically comprises a large fraction of seawater with contaminants that create so-called‘scaling’, which may clot the nozzle opening quickly, especially if the gas is supplied at low pressure. Due to safety regulations in the oil and gas industry, the time needed to flush the separator, clean or replace the nozzle and restart the separator may be in the order of a day or more. Stopping the separator for a day is expensive, for instance because a large storage tank is needed for the unseparated liquid.

[0004] A second example regards adding clean air to a main flow of sooty flue gas in a secondary combustion stage for use in an industrial boiler or gas-turbine application. As in the previous example, the nozzle opening may be relatively wide. However, soot may deposit in the nozzle opening, especially if the air is supplied at low flow rates. Here, deposits may affect combustion, and replacing or cleaning the nozzle may interrupt heat production.

[0005] A third example regards adding a liquid additive intermittently to a gaseous main flow through an atomising nozzle. In this example, the nozzle opening is particularly narrow so a small deposit may have a large effect. However, some applications, e.g. a direct injection system supplying fuel to an internal combustion engine, supply the liquid at high pressure and thereby prevent deposits from building up in the nozzle opening.

[0006] A fourth example regards adding a liquid additive to a liquid main flow in a chemical process plant. In this example, deposits in the nozzle opening may disturb the flow rate and thereby concentrations and the intended chemical reaction.

[0007] In all the previous examples, there is a need to clean a nozzle before deposits disturb operation, and certainly before the deposits clog the nozzle and interrupt operation. However, cleaning a nozzle too often may also disturb or interrupt a process, e.g. if process equipment must be shut down during cleaning. Further, excessive cleaning may cause excessive wear, even if the nozzle is cleaned during operation.

[0008] WO2017215787 Al describes a cleaning device for mechanically cleaning a gas nozzle of an inert-gas welding torch, having a motor-drivable cleaning head which is mounted so as to rotate, and which is mounted in a sprung manner in the axial direction, specifically between a basic position and a cleaning position. After cleaning, the gas nozzle is moved away from the cleaning head and the cleaning head is restored automatically from its cleaning position into its basic position in a restoring stroke under the action of a restoring force.

[0009] US5845357 A describes a nozzle cleaning device comprising a support structure capable of releasably receiving a hollow nozzle body having an inner surface and an open end or terminal edge portion, and a scraper apparatus coupled to the support structure and including at least one scraper finger adapted to move into and out of the nozzle body. The scraper finger has a distal end and a camming surface extending from the distal end to a scraping edge.

[0010] A general objective of the present invention is to solve or alleviate at least one of the problems above while retaining benefits from prior art.

SUMMARY OF THE INVENTION

[0011] This is achieved by a device for cleaning a nozzle according to claim 1 and a system comprising said device according to claim 6. Further features and benefits appear in the dependent claims. The claims' texts follow the usual convention that articles‘a’,‘an’,‘the’ mean‘(the) at least one’, whereas‘one’ means exactly one. Further,‘consists of implies an exclusive list, whereas‘comprises’,‘includes’,‘with’ ‘contains’, etc. imply a list to which items may be added.

[0012] In a first aspect, the invention concerns a device for cleaning a nozzle, the nozzle having a cylindrical nozzle opening. The device comprises a plunger axially movable within the nozzle. The plunger has a plunger head with a diameter less than the diameter of the nozzle opening and a length greater than the length of the nozzle opening. The device further comprises a linear actuator with a motor and an axially movable rod, a linkage between the plunger and the rod, and an adjustable stopper for stopping the plunger in a retracted position that permits fluid flow past the plunger head.

[0013] During normal operation, the stopper enables precise and repeatable positioning of the plunger with inexpensive actuators. During cleaning, the linear actuator and the linkage push the plunger head through the nozzle opening to remove deposits. [0014] In preferred embodiments of the device, the motor is selected from a group comprising a pneumatic cylinder with an associated fluidic circuit, a hydraulic cylinder with an associated fluidic circuit, an electric motor and a solenoid.

[0015] Further, the linkage is selected from a group comprising a lever, a piston rod, a teethed rod, a lead screw and a solenoid shaft.

[0016] Preferably, the actuator and linkage together constitute a force amplifier, such that a force applied to the plunger becomes greater than a force supplied by the motor.

[0017] Force amplifiers include a lever with a short and a long arm, a hydraulic cylinder and a reduction gear coupled to an electric motor.

[0018] The adjustable stopper may comprise a stopper screw, which permits precise and repeatable axial adjustment. Precision is important to adjust a flow rate through the nozzle. Repeatability is important for calibrating a flow rate after cleaning.

[0019] In a second aspect, the invention concerns a system for cleaning a nozzle, wherein the system includes the device just described. The system comprises a control unit with communication means for transmitting signals to and from a remote control room and an electrical connection to the device.

[0020] The control unit may be implemented with relays, power transistors etc. that supplies driving electrical power for an electric motor or solenoid driving the plunger (120) in response to a weaker control signal from the control room.

[0021] Preferably, the control unit further comprises an embedded processor. The embedded processor enables addressable actuators and offloads a processor in the remote control room.

[0022] More preferably, the embedded processor runs a secondary controller that fits a regression curve through a set of measurements acquired at different times in order to determine a time after which the device should clean the nozzle for deposits.

[0023] The secondary controller is a program in the sense‘a set of computer instructions’. The regression curve is a line unless there is available data supporting significant higher order effects such as exponential growth of deposits. A confidence level is convenient for deciding when to clean the nozzle. For example, 80% probability for clogging during the time interval to the next measurement may provide reasonable time to clean the nozzle. In this example, the probability for clogging increases from 80% as we approach the estimated time for clogging.

[0024] In some embodiments, the secondary controller issues a signal causing the plunger head to be pushed through the nozzle opening. In alternative embodiments, it is desirable that a human operator issues a similar signal. [0025] In embodiments with a secondary controller, the secondary controller preferably controls a pressure in a supply line leading to the nozzle as response to a desired flow rate provided by a primary controller. Thereby, a primary controller in a process control system only needs to specify a desired flow rate for the additive fluid, and let the secondary controller compensate for pressure drops caused by deposits and clean the nozzle when needed.

BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The invention will be described in greater detail below by means of exemplary embodiments with reference to the accompanying drawings, in which:

Fig. 1 illustrates a first embodiment of a device according to the invention,

Fig. 2 shows the embodiment in Fig. 1 in a different state,

Fig. 3 illustrates a second embodiment of the device,

Fig. 4 illustrates a generalised control system for the device according to the invention,

Fig. 5 illustrates an algorithm for determining when to clean the nozzle,

Fig. 6a illustrates a linear actuator with a rack-and-pinion arrangement,

Fig. 6b illustrates a linear actuator with a lead screw,

Fig. 6c illustrates a linear actuator with a solenoid and a solenoid shaft,

Fig. 7 illustrates a remotely controlled device according to the invention, and

Fig. 8 illustrates a control unit for remotely controlling one or more devices according to the invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

[0027] The drawings are schematic and not to scale. Several details known to the skilled person are omitted from the drawings for clarity of illustration.

[0028] Figs. 1 and 2 illustrate principles of the invention. More specifically, Fig. 1 shows a device 100 according to the invention with a nozzle 110 having a nozzle opening 111 located within a main fluid channel 190. Contaminants in the fluid flowing through the main fluid channel 190, e.g. process water, seawater or well fluid from an oil or gas well, have formed a solid deposit 200 within the nozzle 110 and the nozzle opening 111.‘Scaling’ known from the oil and gas industry is an example of such deposits 200.

[0029] The main fluid channel 190 may be part of the device 100 and is not necessarily a pipe. For example, the device 100 may be considered a self-cleaning mixer in which the main fluid channel 190 contains vanes to set up a rotation or turbulence in the main fluid (increase the Reynolds number locally). In such a mixer, the nozzle 110 might supply any additive fluid to any main flow, e.g. gas for a bubble separator or a liquid additive for a chemical process.

[0030] The device 100 comprises a carrier 101 for mechanical connection of the nozzle 110, a plunger 120, an actuator 130 and an adjustable stopper 140. A typical carrier 101 would be a housing or a bracket.

[0031] The plunger 120 is an element that is axially movable within the nozzle 110, with a fluid tight connection between an inner wall of the nozzle 110 and an outer cylindrical wall of the plunger 120. A plunger head 121 having a diameter slightly smaller than the nozzle opening 111 extends axially from the plunger 120. The plunger 120 may have a short piston like cylindrical part and a long rod-like plunger head 121. As best seen in Fig. 2, the axial length of the plunger head 121 enables the plunger 120 to push deposits out of the nozzle opening 111.

[0032] One or more fluid channels 122 distributed along the circumference of the plunger 120 fluidly connects an axial bore 123 within the plunger 120 to the interior of the nozzle 110. Preferably, the downstream opening(s) of the fluid channel(s) 122 is/are located near the inner wall of the nozzle 110. This may help preventing deposits 200 from building up on the inner walls of the nozzle 110. The effect of this feature is limited if the additive fluid is a gas supplied at low flow rates.

[0033] The actuator 130 represents a linear actuator of any known kind, for example a pneumatic or hydraulic cylinder with an associated fluidic circuit driving an axially movable rod 131. For clarity of illustration, Figs. 1 and 2 neither show a fluidic circuit with a pump and valves for driving the actuator 130 nor a fluid supply line leading to the nozzle 110. These elements are shown in Fig. 3 and will be explained later. Figs. 6 a-c with accompanying brief descriptions illustrate alternative linear actuators.

[0034] We expect price to be an important design criterion in some applications, so the device 100 preferably comprises simple and effective features that replace or supplement features built into relatively expensive actuators.

[0035] The adjustable stopper 140 provides an example. During normal operation, the plunger 120 should be retained in a fixed and repeatable axial position to permit a desired fluid flow past the plunger head 121. However, limited actuator precision, vibration, axial shocks and various other factors may cause undesired deviations from the desired axial position. Such deviations cause undesirable fluctuations in a fluid flow past the plunger head.

[0036] As shown in Fig. 1, the stopper 140 comprises a stopper screw 141 rotatable in a nut 142 affixed to the carrier 101. The nut 142 illustrates threads fixed with respect to the carrier 101. A hand wheel 143 illustrates means to rotate the stopper screw. Any screw head or rotational actuator may replace the hand wheel 143.

[0037] One arm of a lever 150 abuts the tip of the stopper screw, and the actuator 130 presses the arm toward the tip with a suitable (small) force. Forces needed to rotate the stopper screw 141 are perpendicular to axial forces exerted on the tip. Preventing the screw 141 from rotating between adjustments requires small forces and makes the arrangement 140 resistant to vibration and axial shocks. Thus, an inexpensive actuator 130, e.g. an inexpensive pneumatic piston, can be used in the design without sacrificing precision. A repeatable position is important for calibrating a controlled device, e.g. for establishing a flow rate after cleaning.

[0038] In Fig. 1, the lever 150 acts as a force amplifier that enables a small and inexpensive actuator 130 to exert a larger force on the plunger 120 in order to expel, and possibly crush, the deposit 200. Specifically, a pivot 151 divides the lever arm into a shorter arm with length a, and a longer arm with length b. The shorter arm is attached to the plunger 120 through a first joint 152, and the longer arm is attached to the rod 131 through a second joint 153.

[0039] If a is much longer than the axial motion, the ratio bla equals the ratio of forces at the shorter and longer ends, and the inverse ratio a/b equals the ratio of motion at the shorter and longer ends. Fig. 1 illustrates a lever 150 with b/a = 2, such that the axial force applied to the plunger 120 is twice the force exerted by the rod 131 and the plunger 120 moves half the axial distance as the rod 131 moves. Other ratios b/a are obviously possible. All ratios b/a > 1 reduce a need for a precise and powerful actuator 130.

[0040] Fig. 2 shows the embodiment in Fig. 1 in a state wherein the plunger head 121 has expelled the deposit 200 from the nozzle 110. Reaching this state may require a combination of crushing and flushing through the fluid channel(s) 122 and possibly a few up and down movements. This depends, of course, on mechanical properties of the deposit 200.

[0041] Fig. 3 illustrates an alternative embodiment without the lever 150. This enables a more compact device 100 suitable for applications where space is unavailable or expensive.

[0042] Reference numerals 100 -200 are mostly the same as in Fig. 1 and need no repeated explanation. A threaded connection 132 illustrates that the rod 131 is firmly attached to the plunger 120. A hydraulic piston 330 is an embodiment of the linear actuator 130, and an adjustable stopper 340 fixes the upper position of the plunger 120 similarly to the stopper 140 in Figs. 1 and 2. A movable block illustrates the adjustable stopper 340. However, the stopper 340 may be realised with a stopper screw similar to the stopper screw 141. [0043] During normal operation, a hydraulic circuit 301 -305 maintains a small positive pressure difference p2 - pi over a hydraulic piston 333 with circular area A to maintain a net upward force that presses the plunger 120 against the adjustable stopper 340.

[0044] The hydraulic circuit 301 - 305 comprises a hydraulic pump 301, a first hydraulic line 302 with an actuated valve 304, and a second hydraulic line 303 with an actuated valve 304. We use the explicit term‘hydraulic pump’ to avoid confusion with a pump in a process control system 400 described below. Arrows pointing to the hydraulic pump 301 from both lines 302, 303 illustrate that a hydraulic liquid may flow in either direction to extend and retract the plunger 120, respectively. The actuated valves 304, 305 are preferably of a type that close at failure, thereby locking the plunger 120 rather than inadvertently closing the nozzle opening 111 or putting excessive strain on the adjustable stopper 340. We leave implementation details to those skilled in the art.

[0045] The net force on piston 333 is F = i- {A -A _r ) -prA, where A _r is the area of a cross section of the rod 131. is directed upwards when p2 A -A _r ) -pi A is positive and downwards when pi (A -A _r ) -piA is negative.

[0046] In order to push the plunger 120 downwards to expel the deposit 200 from the nozzle opening 111, the hydraulic pump 301 makes pi > p2 { 1 -AJA) such that F becomes negative. Reference numeral 332 illustrates a circular area a corresponding to the inlet from line 302. Since the same pressure pi acts on both areas a and A, the force applied through line 302 is amplified by a factor A/a on the hydraulic piston 333.

[0047] The hydraulic force amplification A/a may be compared to the amplification b/a of the lever 150. Further, a longer time needed to move liquid from one side of the hydraulic piston 333 to the other may be compared to a longer movement of a longer lever arm. Still further, axial forces applied to the piston 333 is reduced by a factor aJA in line or pipe 302 where a relatively inexpensive valve 304 may handle the smaller forces. Similar

considerations apply to the underside of the piston 333, which has hydraulic area (A - A/).

[0048] For a numerical example, we first rewrite the hydraulic amplification A/a = {D/d) ² where D and d are the inner diameters of the hydraulic piston 330 and the hydraulic line 302, respectively. Next, we set D = 40 mm for the piston 333 and d = 4 mm for line 302 to achieve a nice and round hydraulic amplification A/a = (40/4) ² = 100. Thus, in the present example, the force acting on piston 333 is 100 times the pump force. Conversely, only 1 % of an axial force acting on the piston 333 is applied to the liquid in line 302. A major cost of larger hydraulic amplification is slower operation, because a larger volume of liquid must move from one side of piston 333 to the other. In the present example, a 50 cm long liquid column must move through the lines 302, 303 in order to move the hydraulic piston 5 mm. A 50 cm liquid column is quite long for a small hydraulic pump 301.

[0049] In the claims, we generalise the hydraulic circuit 301 - 305 to a‘fluidic circuit’ that includes a similar pneumatic circuit suitable for driving a pneumatic cylinder, e.g. for the embodiment in Figs. 1 and 2. However, the examples just described rely on incompressibility of liquid and do not apply to a pneumatic cylinder. Thus, the lever 150 in Figs. 1 and 2 is a useful linkage for a pneumatic cylinder, whereas a hydraulic cylinder amplifies the force from the circuit 301 - 305 because liquids are incompressible for all applications of interest herein.

[0050] For the following, we assume there is an external process control system 400 aiming to supply a certain flow rate u through the nozzle opening 111. In the illustrated system 400, solid arrows represent process data and dashed arrows represent physical phenomena.

[0051] The flow rate u depends on the application. For example, u may represent an amount of gas supplied to a bubble separator or boiler measured in kg/s or m ³ /h, or the amount of a liquid substance for a chemical process measured in mol/s. In terms of control theory, the flow rate u is a primary process variable, and the desired value is a set point 401. A feedback loop 402 illustrates a comparison of a measured response with the set point 401.

[0052] The existing process control system 400 comprises a primary controller (master) 410 and an actuator 420. The actuator 420 applies a pressure / _¾ in order to maintain the desired flow rate u. For this, the actuator 420 presumably comprises a pump, which has a different purpose and probably a different design than the hydraulic pump 301 described above. The pressure / _¾ is called a secondary process variable. We want to provide a nozzle cleaning system that takes care of‘everything’ related to the nozzle. For this, we will replace the actuator 420 with an inner feedback loop responsible for supplying a desired flow rate u. This is called cascade control and will be explained with reference to Fig. 4.

[0053] The existing process control system 400 supplies the additive fluid at pressure / _¾ to the axial bore 123 through a fluid supply line 429. The additive fluid proceeds to a primary process 430, e.g. bubble separation or a chemical reaction such as combustion. Reference numeral 440 represents primary disturbances, e.g. un-modelled features and random noise.

[0054] Notice that the letter u next to the feedback 402 is for illustration and should not be construed literally. In a real system 400, the quantity measured in response to a new set point for u is often not a flow rate. For instance, the response may be content of oil in water measured on an output line from a bubble separator or the outcome of a chemical process. Either way, the primary controller 410 is responsible for computing a new flow rate u from measurements to approach a desired output. [0055] Fig. 4 ¹ is a block diagram of a cascade control system with an outer loop having feedback 402 relating to the primary process variable u and an inner feedback loop 421 - 425 relating to the secondary process variable / _¾ . As mentioned, this corresponds to the process control system 400 in Fig 3 with the inner loop 421 - 425 replacing the actuator 420.

Reference numerals related to the outer loop are explained with reference to Fig. 3.

[0056] A basic principle of cascade control is to move responsibility for a secondary process variable from the primary controller (master) to a secondary controller (slave). In Fig. 4, the primary controller 410 specifies a value for the flow rate u to a secondary controller 421. The secondary controller 421 decides how to achieve the specified flow rate u, that is, which pressure / _¾ to apply and when to clean the nozzle opening 111. The secondary controller 421 may be embedded in a control unit 820 illustrated in Fig. 8 and described below.

[0057] For this, an actuator 423 represents the above-mentioned pump to adjust the pressure P3 in the supply line 429 as well as the actuators 130; 330, linkage 150, actuated valves 304, 305 etc. described above.

[0058] A secondary process 425 represents a mathematical model of how the pressure p ₃ affects the flowrate u. This depends on the amount of deposits 200 in the nozzle opening 111. In an example algorithm illustrated by Fig. 5, we assume that the deposits 200 grow approximately proportional to time since the latest cleaning.

[0059] Inner loop disturbances 424 represent physical phenomena, for example a valve or pump not operating as expected, non-linear growth of deposit and/or additive fluid pushing some deposit 200 out from the nozzle opening 111 between cleanings.

[0060] The conditions for a cascade control system to work properly are known from control theory. For reference, these are:

• The inner loop must influence the outer loop. Specifically, the actions of the secondary controller 421 must affect the primary process variable u in a predictable and repeatable way such that the primary controller 410 can influence its own process.

• The inner loop must be faster than the outer loop. The secondary process 425 must react to the secondary controller’s efforts at least three or four times faster than the primary process 430 reacts to the primary controller 410. This enables the secondary controller to compensate for inner loop disturbances before they can affect the primary process.

¹ Fig. 4 and related text are adapted from www.controleng.com/single-article/fundamentals-of-cascade-co ntrol/ • The inner loop disturbances 424 must be less severe than the outer loop disturbances 440. Otherwise, the secondary controller will be unable to apply consistent corrections to the primary process, e.g. because it constantly corrects disturbances to the secondary process.

[0061] The pressure / _¾ clearly influences the flow rate u and can be adjusted in a predictable and repeatable way. Further, adjusting p is faster than measuring the effects of a new set point 401 on a bubble separator or chemical process. Disturbances related to a pump circuit for increasing / _¾ are probably less severe than disturbances related to bubble separation, combustion or a process plant. Thus, / _¾ is a secondary process variable suitable for a cascade control system. Given that the original control system 400 shown in Fig. 3 controls a pump and/or valves in a pump circuit for this purpose, moving functionality to the secondary controller 421 should be relatively easy.

[0062] Deposits 200 may grow slowly compared to changes of flow rate u. For example, a bubble separator may treat several batches of contaminated water, each with an optimal flow rate u, during the time needed for deposits 200 to grow significantly. Hence, we need an internal process variable that accounts for different values of u.

[0063] As a first approximation, we assume that / _¾ = R U, where R is a‘resistance’ that grows as deposits 200 settle in the nozzle opening 111, cf. Ohms law. Dividing both sides by u we obtain R = pJu. Further, we multiply by a suitable constant to obtain a dimensionless parameter y = C-pJu, where C is proportional to R and scaled such that 1.0 corresponds to no deposits in the nozzle opening 111. The constant C depends on specific units, for example Pa or psi for / _¾ and kg/s, m ³ /h or mol/s for u. However, a dimensionless y exists for all consistent measurement systems. Moreover, the parameter y may be constructed from an infinite number of alternative measurements, cf. Buckingham’s pi-theorem.

[0064] Fig. 5 illustrates an algorithm 500 for deciding when to clean the nozzle 110. In particular, the algorithm is illustrated by a diagram showing y = C-pilu as a function of time t. As deposits 200 settle in the nozzle opening 111 over time, the pressure p3 required to achieve a flow rate u through the supply line 429 increases, and so does y. For illustration, we arbitrarily assume that a cleaning is required whenever y approaches 1.5, that is, when the pressure p3 approaches 1.5 times the pressure needed to supply u through a clean nozzle opening 111. A dashed line 501 indicates the threshold value.

[0065] We measure y at regular intervals D T. Each measurement 502 has a mean value indicated by a short horizontal bar and a variance (uncertainty) indicated by a vertical bar. According to common practice, a sample mean and a sample variance from a series of tests obtained during a short period would constitute each measurement 502. The test period may be hours or days as long as it is short compared to the growth rate of deposits 200.

[0066] The mean values show an increasing trend over time as illustrated by a linear regression line 503. The mean values are located above and below the line 503, for example because deposits 200 may have been pushed out of the nozzle opening 111 by the supplied fluid, grow irregularly depending on concentration of contaminants in the main flow, or because other factors such as noise affect the measurements differently at different times.

[0067] It is possible that other regression curves provide better estimates than a line.

However, a line is the best approximation until available data indicate a different regression curve, e.g. due to second order effects or exponential growth.

[0068] The regression line 503 provides a better estimate of the average growth than any single measurement (test series) 502. In statistical terms, we reduce the number of‘false positives’, here measurements that may lead to unnecessary cleaning and wear, and‘false negatives’, here measurements that do not detect real clogging which may disturb or interrupt operation. However, one measurement or test series 502 over the threshold 501 should cause immediate cleaning.

[0069] At time T _\ = NAT, that is, after N intervals, it is estimated with some predetermined probability, or confidence level, e.g. 80%, that the regression line 503 will cross the threshold 501 during the next time interval AT. The estimated time for crossing is G ₂ in this example.

[0070] During the time between 7i and G ₂ , the confidence level grows. That is, the probability for clogging increases from 80% in the present example. A fully automatic system might clean the nozzle immediately when a probability or confidence level exceeds a certain threshold, here 80%, here at 7). Other systems may raise an alarm and leave the decision for cleaning to a human operator.

[0071] An arrow 504 illustrates that y returns to the original value 1.0 as the plunger 120 expels the deposits 200 from the nozzle opening 111 at time 7). The flow rate through a clean opening 111 depends on the position of the plunger head 121. This is why repeatability is important for calibration of the device 100, cf. the description of Fig. 1.

[0072] Relevant statistical theory and methods can be found in textbooks and online.

[0073] Figs 6a and 6b illustrate alternative linear actuators 630 driven by an electric motor 601, e.g. a stepping motor. A reduction gear 602 amplifies the force supplied by the electric motor 601.

[0074] Fig. 6a illustrates a rod 631 with a teethed section 633 engaging the gear 602 in a so- called rack-and-pinion arrangement. [0075] Fig. 6b illustrates a lead screw 632 driven by a small gearwheel 602 rotatably mounted in a nut 634. The nut 634 represents threads fixed to the carrier 101 in Fig. 1.

[0076] Fig. 6c illustrates a solenoid 603 with a solenoid shaft 635. The direction and magnitude of a force exerted on the shaft 635 depends on the current or voltage applied to the solenoid. Control circuits for current or voltage- activated solenoids are commercially available and come with their respectively known advantages and disadvantages. Solenoid valves operating on this principle are probably the most widely used actuated valves in all branches of industry, cf. the actuated valves (fail -> close) 304 and 305 in Fig. 3.

[0077] The shaft 635 may be coupled to the long arm of the lever 150 in order to limit a voltage supplied to the solenoid 603, and thereby a potential explosion hazard. Alternatively, the solenoid 603 may be driven by a‘high’ current and low voltage to avoid the potential explosion hazard.

[0078] Fig. 7 illustrates a remotely controlled device 100 according to the invention. In particular, a production platform 710 on a sea surface 10 contains a control room 711, and the controlled device 100 is located in a subsea installation 720 residing on a seafloor 20. A production line 715 connects the production platform 710 with the subsea installation 720.

[0079] The control room 711 contains a computer system 712 and a console 713. The primary controller 410 shown in Figs. 3 and 4 is a computer program running on one or more, usually several, processors in the computer system 712, for example on a Windows or Unix- like operating system. During production, a human operator uses the console 713 to monitor various processes and intervene when required or desirable. In the present example, the process control system 400 in Figs. 3 and 4 relates to the subsea installation 720, which as mentioned contains a device 100 according to the present invention.

[0080] The production line 715 represents a riser 716 for conveying fluid to the production platform 710 and an umbilical with a communication channel 717 for transmitting signals between the production platform 710 and the subsea installation 720.

[0081] A dashed circle 721 contains an enlarged view of a detail in the subsea installation 720 and shows the device 100 as an integral unit within the carrier 101. As explained with reference to Fig. 1, a housing is one form of the carrier 101. The exact design depends on the application. In the present example, a housing 101 able to withstand the pressure at the seafloor 20 is obviously useful. Making the surface of the housing 101, and hence the device 100, small, is advantageous in subsea applications - force equals area times pressure. Thus, a solenoid based linear actuator may have a better cost/benefit ratio than a pneumatic cylinder in a subsea application. Conversely, an inexpensive pneumatic cylinder 130 with a lever 150 may have a better cost/benefit ratio than a solenoid based linear actuator in a process plant on dry land.

[0082] In Fig. 7, the carrier 101 is mounted on a main fluid channel 190 and receives additive fluid through a supply line 429. As mentioned, the housing 101 may contain a mixer.

[0083] Fig. 8 illustrates a system 800 for remote control of one or more devices 100. As in Fig. 7, a closed housing 101 represents each device 100. Reference numeral 801 represents electrical lines. Fluid lines 190, 429 are omitted from Fig. 8 for clarity of illustration.

[0084] A control room 811 is similar to the room 711 in Fig. 7, and may be located in any suitable place, for example in a process plant on dry land. A two-way communication line 817 transmits signals and is a generalisation of the communication channel 717 in Fig. 7.

[0085] In a basic embodiment, communication electronics 822 represents relays, transistors, programmable logical devices etc. that are able to supply a driving electrical power to the linear actuators within the housings 101 in response to a weak electrical signal from the remote control room 811. For example, a small electrical signal applied to the base of a power transistor opens for a substantially larger electrical power through the emitter of the transistor. The larger current and/or voltage can drive an electrical motor or solenoid with power from a local source. Thus, the larger power may, but need not, be supplied through the line 817, or equivalently, without power supplied through the umbilical in Fig. 7.

[0086] It follows that the device 100 can be controlled remotely from the control room 811 through a control unit 820 without an embedded processor, regardless of whether power is supplied from a local source or through a power line. This corresponds to the more abstract actuator 420 in Fig. 3.

[0087] A more sophisticated control unit 820 comprises an embedded processor 821 with associated storage (not shown) and more advanced communication electronics 822, e.g. for an Ethernet, a WiFi or fibre optical communication channel 817. Keep in mind that we are not confined to a subsea application. The embedded processor 821 would typically have low power consumption and might be similar to the kind used in smartphones. Similar to the computer system 712 in Fig. 7, the embedded processor 821 may run a Windows or Unix-like operating system, however with substantially less computing power than the system 712.

[0088] The more sophisticated control unit 820 just described would supply driving power to electric motors, solenoids etc. through relays, power transistors etc. similar to the basic embodiment.

[0089] So far, nothing is said about the dimensions of the electrical line(s) 801. Indeed, some lines 801 may be dimensioned for supplying electrical power (voltage times current) suitable for driving an electrical motor or solenoid directly. Other lines 801 may convey a weaker electrical signal to equipment within a housing 101.

[0090] An embedded processor 821 does not necessarily run a computer program

implementing the secondary controller 421 in Fig. 4. In Fig. 8, reference numeral 421 is in square brackets to indicate that running the secondary controller 421 is optional, and in bold typeface to indicate that the secondary controller 421 is highly recommended.

[0091] The secondary controller 421 preferably compensates for pressure drops caused by deposits 200 and cleans the nozzle 110 automatically or alternatively alerts a human operator that the nozzle 110 needs cleaning. See the description of Figs. 4 and 5 for details.

[0092] For the sake of order, we consider it well within the capabilities of the skilled person to select a suitable motor, linear actuator, linkage and other equipment for a given application, especially in view of the explanations herein.

[0093] While the present invention has been illustrated by examples, the scope of the present invention is set forth in the accompanying claims.