Computing device, pressure control station, system and method for controlling fluid pressure in fluid distribution network

Abstract

A method performed by a computing device is provided for controlling fluid pressure in a fluid distribution network (FDN) above a threshold pressure by communicating with a plurality of pressure control stations distributed throughout the FDN to independently control fluid pressure at each of the plurality of pressure control stations. The method includes training a machine learning algorithm to establish one or more correspondences between changes in fluid demand measured in the FDN over a first period of time and environmental conditions measured over the first period of time. The method includes predicting a change in fluid demand in the FDN for a later second time period based on the predicted environmental conditions for the second time period and one or more correspondence relationships established between the measured change in fluid demand in the FDN for the first time period and the measured environmental conditions of the FDN for the first time period. The method includes determining a change in fluid pressure to be applied at the plurality of pressure control stations for a second time period based on a change in fluid demand predicted for the second time period to meet a fluid pressure condition at one or more predetermined points in the FDN downstream of the plurality of pressure control stations. The method includes sending to the plurality of pressure control stations an indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations for the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN.

Description

Background Art

[0001] The present technology relates to computing devices, pressure control stations, systems and methods for controlling fluid pressure in a fluid distribution network.

[0002] This application claims the Paris Convention priority of UK Patent Application No. 2206933.0, the contents of which are incorporated herein by reference in their entirety.

[0003] The "background" description provided herein is for the purpose of generally presenting the context of the present disclosure. The work of the presently named inventors, to the extent it is described in this background section, and insofar as the description may not have qualified as prior art at the time of filing, is neither explicitly nor implicitly admitted to be prior art against the present invention.

[0004] Fluids such as liquids or gases are sometimes distributed by a network of pipes that transport the fluid under pressure from a source to one or more consumer points. An example of such a network is a gas distribution network, such as one used to provide combustible gas to consumers for consumption by both industrial and domestic consumers.

[0005] Gas distribution network (GDN) distributes gas from gas source to gas consumers, such as family residences, commercial or industrial sites. GDN is usually formed by a pipeline network, through which gas reaches consumers from sources under pressure. The gas pressure in the network is set by one or more pressure control stations, which are called regional regulators (governor) stations (or "regulator stations"), and the pressure control stations receive gas from sources. It is usually desirable to set the pressure in the network to achieve a balance between safety, maintaining consumer service levels and gas leakage in GDN. Too low gas supply pressure may cause danger to consumers. For example, low gas supply pressure may cause incomplete combustion and subsequent formation of carbon monoxide, toxic gases, or consumer equipment does not work at all. Therefore, the statutory (statutory) requires that the gas supply pressure in GDN should not be less than the minimum value. On the contrary, GDN may be prone to leakage, which is a concern for the environment and finance of gas suppliers. Generally, leakage increases with gas supply pressure. Therefore, gas suppliers may wish to impose a maximum gas supply pressure in GDN to reduce the economic losses and environmental impact of gas leakage. Therefore, it is desirable to control the gas supply pressure in the GDN to be high enough to comply with the statutory minimum requirements, but low enough to reduce the economic losses and environmental impacts due to gas leakage.

[0006] The pressure of the gas supply at a point in the GDN is determined by a number of factors, including: the pressure at the regulator station or stations supplying the point in the GDN, the distance of the point from the regulator station or stations, and the demand for the gas supply. Typically, the pressure of the gas is measured at one or more pressure low points in the GDN using a digital gas pressure data logger (which may alternatively be referred to as a "low point logger"). A low point is a point in the GDN that has a low (possibly minimum) pressure. A GDN may have multiple low points.

[0007] A higher demand for gas results in a decrease in pressure at low points in the GDN as consumers of the gas supply consume the gas. Increased gas supply demand decreases the gas supply pressure in the GDN and vice versa. Therefore, the gas supplier needs to supply gas at a sufficiently high pressure that takes into account the potential pressure drop due to the increased demand.

[0008] In some GDNs, the gas supply pressure is manually set at the regulator station. The gas supply pressure is often set to a high value to prepare for the worst case scenario. For example, the pressure may be set at a sufficiently high pressure so that even if the gas demand is expected to be the highest gas demand on any given day, the pressure at the low point is expected to remain above the legal minimum requirement. In some GDNs, the regulator station is configured to automatically change the gas supply pressure based on a "clock". For example, the regulator station may supply gas at one pressure during the day and another pressure at night. In some GDNs, the regulator station is configured to change the gas supply pressure based on a predetermined pressure profile.

[0009] For some GDNs, such as those supplying biomethane gas, it is desirable to prevent the GDN pressure from exceeding a predetermined threshold. For example, if the pressure in the GDN exceeds a predetermined threshold, it becomes difficult to feed the biomethane gas into the GDN. Similarly, if the pressure ratio between the GDN pressure and the biomethane planetary outlet pressure exceeds a predetermined threshold, it becomes difficult to feed the biomethane gas into the GDN. Current methods address this problem by burning excess biomethane gas.

[0010] The above technical challenges faced in controlling gas pressure in a GDN are more generally representative examples of the difficulties encountered in accurately controlling fluid pressure in a fluid distribution network FDN. Therefore, it is desirable to more accurately control fluid pressure at a pressure control station in a fluid distribution network in response to changing environmental conditions. Summary of the invention

[0011] Embodiments may provide a method for controlling fluid pressure in a fluid distribution network FDN by communicating with one or more pressure control stations, performed by a computing device. The method includes training a machine learning algorithm to establish one or more correspondences between changes in measured fluid demand in the FDN during a first time period and environmental conditions measured during the first time period. The method includes predicting changes in fluid demand in the FDN during a later second time period based on environmental conditions predicted during the second time period and one or more correspondences established between changes in measured fluid demand in the FDN during the first time period and environmental conditions measured during the first time period of the FDN. The method includes determining changes in fluid pressure to be applied at one or more pressure control stations during the second time period based on the changes in fluid demand predicted during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the one or more pressure control stations. The method includes sending an indication of the determined changes in fluid pressure to be applied at one or more pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN to one or more pressure control stations.

[0012] According to example embodiments, by predicting a change in fluid demand over a time period for predicted environmental conditions (e.g., based on a weather forecast), a corresponding predicted change in fluid pressure to be applied at a pressure control station may be made (e.g., by changing a pressure control station setting) to match the demand of the predicted environmental conditions. An indication of the change in fluid pressure to be applied at the pressure control station (e.g., the pressure control station setting) is sent to the pressure control station. The indication may be sent prior to a later time period to control fluid pressure in the fluid distribution network during that time period. Thus, by taking demand into account, fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station may be met.

[0013] In some embodiments, the fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station are conditions that minimize overpressure in the FDN. For example, the overpressure may be the difference between the fluid pressure at a low pressure point at the end of the FDN and the minimum allowable fluid pressure. The minimum fluid pressure may be a statutory minimum fluid pressure.

[0014] In some embodiments, the fluid pressure condition at one or more predetermined points in the FDN downstream of the pressure control station is a condition that minimizes the difference between the fluid pressure at the one or more predetermined points in the FDN and the maximum allowable fluid pressure. In some embodiments, the fluid pressure condition at one or more predetermined points in the FDN downstream of the pressure control station is a condition that the fluid pressure at the one or more predetermined points falls within predetermined maximum and minimum fluid pressures.

[0015] The one or more predetermined points downstream of the one or more pressure control stations may be any point along the FDN that receives fluid from the one or more pressure control stations. For example, the one or more predetermined points may be a low pressure point at the end of the FDN.

[0016] In an exemplary embodiment, the FDN is a GDN and the one or more pressure control stations are regulator stations.

[0017] In an exemplary embodiment, a computing device is used to control the fluid pressure in the FDN above a threshold pressure by communicating with a plurality of pressure control stations distributed throughout the FDN to independently control the fluid pressure at each of the plurality of pressure control stations. Such an embodiment provides increased flexibility in the control of fluid pressure in the FDN.

[0018] Various aspects and features of the present disclosure are defined in the following claims.

[0019] It should be understood that the foregoing general description and the following detailed description are exemplary rather than restrictive of the present technology. The described embodiments and further advantages will be best understood by reference to the following detailed description taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] A more complete understanding of the present disclosure and its many attendant advantages will be readily obtained as the same becomes better understood by reference to the following detailed description when considered in conjunction with the accompanying drawings in which like reference numerals designate identical or corresponding parts throughout the several views, and in which:

[0021] Figure 1 A simplified gas distribution network (GDN) is shown;

[0022] Figure 2 is a schematic diagram showing a portion of a GDN that provides gas to a plurality of consumers along a route;

[0023] Figure 3 is a graph of gas pressure versus time over a 24 hour period for a cross section of the GDN for a constant regulator pressure;

[0024] Figure 4 is a diagram of a method for controlling a gas according to an exemplary embodiment;

[0025] Figure 5A is a graph of average low pressure versus time for cold and warm winters;

[0026] Figure 5B is a graph of daily average low pressure against nighttime temperature;

[0027] Figure 6 is an example of a gas demand proxy measured over a week with corresponding air and night thermostat average and low pressures over a 24 hour period;

[0028] Figure 7 is a flow chart illustrating training a machine learning algorithm to predict a gas demand agent according to an exemplary embodiment;

[0029] Figure 8 is a graph of a parameterized gas demand proxy versus time according to an exemplary embodiment;

[0030] Fig. 9 is a flow chart illustrating how the low point pressure may be used to obtain an estimate of the optimum regulator pressure according to an exemplary embodiment.

[0031] Fig.10 shows simulation results of a low point pressure model according to an exemplary embodiment;

[0032] Fig.11 is a graph of pressure versus time for the variable regulator pressure.

[0033] Fig.12 shows a comparison between using a constant regulator pressure and a variable regulator pressure in a GDN according to an exemplary embodiment;

[0034] Fig.13 is a schematic diagram illustrating an actuator for adjusting a pilot valve that provides a remote component for adjusting gas pressure at a regulator station according to an exemplary embodiment;

[0035] Fig.14 shows an overview of a fluid control system according to an exemplary embodiment;

[0036] Fig.15 is a flow chart illustrating a method performed by a computing device according to an example embodiment. DETAILED DESCRIPTION

[0037] As described above, embodiments of the present technology can improve the accuracy of controlling the pressure in a fluid distribution network relative to the demand for the fluid. A better understanding of the example embodiments can be understood with respect to the example of a gas distribution network (such as a regional combustible gas distribution network). Gas consumers can be industrial, commercial, or household consumers, etc., which receive a supply of gas from a gas distribution network (GDN). The term "gas distribution network" is used herein to refer to a pipeline network and one or more regulator stations used to distribute gas to one or more consumers. It should be understood that a GDN is an example of a fluid distribution network (FDN). Examples of GDNs are Figure 1 Shown in. Figure 1is a simplified representation of a GDN 2, in which a gas source 14 supplies gas to a regulator station 12 via a distribution pipeline 10. The regulator station 12 controls the pressure of the gas from the gas distribution pipeline 10 to the gas distribution network 2 at a lower pressure than the pressure in the gas distribution pipeline 10. The GDN 2 supplies gas from the source 14 to a plurality of consumers 16 via the regulator station 12. A plurality of low point recorders 6 are used to measure the gas pressure at low points in the GDN 2.

[0038] The gas source 14 generally represents a gas source, which may be an independent gas container or one or more other gas networks. For example, the National Grid system is a network that supplies high pressure gas that is delivered to the GDN throughout the United Kingdom. The gas source may also be a source of biogas (such as biomethane) produced by a source such as a farm or a dedicated plant. It should be understood that although Figure 1 A single gas source 14 is shown in FIG, but the GDN 2 may have multiple gas sources.

[0039] Gas pressure is typically highest at the source entering the GDN and lowest at the end of the GDN due to gas leaks and gas use by consumers. For example, a national grid may supply high pressure gas at one of the sources of the GDN. The gas moves through the pipeline driven by pressure, and as consumers use the gas and leaks cause the pressure to drop. The regulator station (or regulator) in the GDN typically receives gas from a higher pressure gas source and contains a pressure control component to reduce the pressure of the gas received from the gas source. Therefore, the pressure of the gas arriving at the regulator is higher than the pressure of the gas leaving the regulator. In one example, the gas arrives at the regulator station at a pressure of about 1 to 2 bar and leaves the regulator station at a pressure of up to about 50 millibars. Due to consumer use of the gas and due to gas leaks, the gas pressure received by one or more consumers will typically be lower than the pressure leaving the regulator. Therefore, there are one or more points in the GDN where the gas pressure may be low or minimum. Low point recorders 6 are typically placed at some or all of these locations to monitor the gas pressure there, such as Figure 1 For example, a model calibrated by pressure measurements across the GDN may be used to locate low points, and a low point recorder 6 placed at the low point.

[0040] Figure 2 It is shown Figure 1 Schematic diagram of a portion of a GDN 2 is shown, which provides gas to a plurality of consumers. Figure 2As shown, a regulator station 12 receives medium pressure gas via a gas distribution supply pipeline 10. As described above, the medium pressure gas supply source is not limited and may be, for example, a connection to another GDN. The regulator station 12 changes the medium pressure gas supply to a pressure Pd. The regulator station 12 may change the gas pressure by using pressure control components such as pilot valves and actuators. The gas at pressure Pd leaves the regulator station and enters the pipeline network of the GDN 2 to supply the gas to a plurality of consumers 16. The low point recorder 6 is typically located at a point near the edge of the GDN 2 that may have a low or minimum gas pressure. At Figure 2 In the embodiment, the pressure P1 is measured at the low point 6. The regulator 12 is manually configured with a setting to achieve the pressure Pd.

[0041] The gas pressure Pd is usually set manually. As will be appreciated, gas consumption usually varies throughout the year in proportion to environmental conditions. For example, in the Nordic winter, the weather is usually colder, so gas consumption will increase. Therefore, the gas pressure Pd at the regulator station 12 is manually set with different pressures between summer and winter. The pressure Pd set by the regulator station 12 is required to ensure that the gas pressure Pl at the low point 6 is higher than the minimum value required for the consumer 16 to operate the gas combustion device. However, the pressure in the GDN 2 will vary depending on the demand for gas from the consumer 16 connected to the GDN 2. This requires that the pressure Pd at the regulator station 12 be set to a value that delivers the minimum pressure at the low point 6 when the consumer demand is highest. As a result, when the demand for gas is low, the pressure set by the regulator station 12 is higher than the pressure it needs, which may increase the amount of gas leakage from the GDN 2. Figure 3 An example illustration is provided of the need to set the gas pressure at the regulator station 12 to a maximum value when demand is greatest, which may result in the pressure in the GDN 2 being too high at other times.

[0042] Figure 3 Provided Figure 2 Illustration of a graph of gas pressure versus time over a 24 hour period for GDN 2. Figure 3 A graph of network pressure versus time of day is provided, showing an example of the relationship between the pressure measured at the regulator station 12 (hereinafter referred to as "regulator pressure 30, Pd") and the pressure measured at the low point 18 of the GDN (hereinafter referred to as "low point pressure 32, Pl"). The regulator pressure 30 may also be referred to herein as the "gas supply pressure". Figure 3 It can be seen that the regulator pressure 30 is constant over time over the 24 hours recorded. This is because the regulator pressure 30 in this example is manually set at the regulator station 12. However, the low point pressure 32 varies over time over the 24 hours. For example, the low point pressure 32 may drop due to increased consumer demand. Figure 3, the low point pressure 32 drops during the early morning hours. This may be a result of the cold temperatures typical of the early morning hours and the increased consumer demand as the home heating system is turned on. An overpressure 22 representing the difference between the minimum consumer pressure 20 and the low point pressure 32 is shown. For example, the minimum consumer pressure 20 may be a minimum legal pressure or a pressure required to meet a customer service level. A high overpressure is undesirable because a higher pressure may increase the likelihood of a higher gas leak than the pressure required to meet the minimum consumer pressure 20.

[0043] Exemplary embodiments may provide a system and method that may predict the likely demand for gas of a selected GDN and automatically control one or more gas regulators of the GDN based on the predicted demand over a predetermined period of time (such as a day), predicting environmental conditions based on the day's weather forecast to reduce over-pressurization of the GDN.

[0044] Attempts have previously been made to change the pressure distribution in response to consumer demand. For example, GB2252848B discloses a gas supply pressure control device for controlling the gas pressure in the gas main according to one of a plurality of pressure distributions stored in an electrical controller to provide an appropriate pressure for the time of day, day of the week, season of the year, etc.

[0045] The pressure distribution required in GB22252848B is a graph of gas pressure in a gas distribution network against time. The pressure distribution is predetermined based on historical data. For example, the pressure distribution of pressure against time for the upcoming winter season can be based on the pressure distribution recorded from the previous winter season. The pressure distribution can then be used to control the gas supply pressure. The network can only use one pressure distribution at a given point in time, and the switching of the pressure distribution is triggered by meeting predetermined criteria. For example, the summer distribution can be triggered based on a measurement of ambient temperature changes.

[0046] However, there is a need for autonomous gas demand forecasting and increased flexibility in controlling gas supply pressure in response to the forecast in order to minimise financial losses and environmental impacts due to gas leaks whilst ensuring that minimum statutory requirements are met at low points in the GDN.

[0047] Example embodiments of the present technology may control the gas pressure in a selected GDN according to predicted demand by:

[0048] ● For specific days of the year and for general environmental conditions (such as air temperature, etc.)

[0049] measuring gas pressure at one or more low points in a selected GDN;

[0050] - calculating a measured gas demand allocation based on the measured gas pressure at one or more low points and the regulator pressure in the selected GDN;

[0051] = creating a parameterized model based on the measured gas demand distribution in the selected GDN, the parameterized model being usable to generate a predicted gas demand distribution by generating a plurality of parameters to parameterize the measured gas demand with respect to time, a reduced number of parameters being used to characterize daily variations in gas demand;

[0052] ● using a machine learning process to generate selected values ​​for said reduced number of parameters, the selected values ​​being required to generate a parameterized model for predicting demand distribution for each day of the year based on prevailing environmental conditions, applying the predictions and repeatedly training the machine learning algorithms for environmental and predicted demand;

[0053] • generating a predicted gas demand allocation from a parameterized model of the selected GDN for a particular day of the year by predicting at least some parameters of the measured allocation forecast needs to be predicted to represent the predicted demand allocation;

[0054] ● Based on predicted environmental conditions, e.g., from a weather forecast, setting parameters to corresponding values ​​to generate a predicted demand allocation for a particular day of the year;

[0055] ● For the forecasted gas demand allocations for the selected GDN, simulate regulator pressure settings and resulting low points for the forecasted demand allocations, score the achievements, and iterate to estimate optimized or improved pressure settings for one or more regulators for the GDN;

[0056] ● Sending regulator pressure settings during a predetermined period, such as one or more days in advance, to match predicted demand allocations;

[0057] =Measure the low point pressure of the selected GDN for the applied regulator settings over one or more days and refine one or more of the pressure settings and parameters used to generate the predicted gas demand allocation. In other words, the measured low point pressure is used to recalibrate the parameterized model over time. Thus, the input parameters of the parameterized model can be modified over time in response to changes in demand for a particular GDN.

[0058] It will be appreciated that a "predicted gas demand allocation" is the predicted change over time of any measure of gas demand over a predetermined upcoming time period of the GDN. It will be appreciated that a "measured demand allocation" is the measured change over time of any measure of gas demand over a measured period of the GDN.

[0059] The above steps according to example embodiments will be explained in more detail below.

[0060] exist Figure 4A diagram of the effects of the example embodiments is shown graphically in FIG. Figure 4 As shown, the example embodiment can be divided into three main stages: a demand prediction stage 60, a regulator pressure calculation stage 62, and an application stage 64. As described above, the demand prediction stage 60 involves predicting the measurement of gas demand in the GDN for a predetermined time period from a parameterized model based on predicted environmental conditions. The regulator pressure calculation stage 62 involves using a parameterized model for predicting gas demand to determine the effect of regulator pressure on pressure at low points in the GDN, and estimating an optimal change in regulator pressure over time that minimizes overpressure in the GDN for a predetermined time period. The application stage 64 involves transmitting instructions to the regulator station to achieve the estimated optimal change in regulator pressure over time for a predetermined time period. The "optimal change in regulator pressure over time for a predetermined time period" is a change in regulator pressure over time for a predetermined time period that can meet a pressure condition at one or more predetermined points (such as a pressure low point) in the GDN. For example, the optimal change in regulator pressure over time for a predetermined time period can be a change in regulator pressure that can at least reduce gas leakage in the GDN, but preferably minimize overpressure in the GDN for a predetermined time period.

[0061] As part of the demand forecasting phase, a weather service (such as a weather bureau) provides the demand forecaster 104 with predicted environmental conditions 102 for a predetermined time period. The predicted environmental conditions for the predetermined time period may be predicted environmental conditions for the area where the GDN is located. For example, the predicted environmental conditions may be a weather forecast for an upcoming day or an upcoming week. In an exemplary embodiment, the predicted environmental conditions are predicted for an upcoming day. The predicted environmental conditions may include, but are not limited to, one or more of temperature, wind speed, humidity, etc.

[0062] The demand forecaster 104 uses the predicted environmental conditions (as will be explained below) in a trained machine learning algorithm to predict a gas demand proxy (a measure of gas demand) for the GDN over a predetermined time period. The gas demand proxy can be any quantity that represents a measurement of the gas demand in the GDN over time. The demand forecaster 104 provides the predicted gas demand proxy to the regulator scheduler 106. As will be explained below, a predicted gas demand proxy can be generated for a particular GDN by measuring the low point pressure of the selected GDN throughout the measurement period in order to model the gas demand of the GDN.

[0063] As part of the regulator pressure calculation phase, the regulator scheduler 106 uses the predicted gas demand proxy to obtain a low point pressure model for the network. The low point pressure model is configured to use the predicted gas demand proxy to simulate the effects of changing one or more regulator pressures in the GDN on one or more low points in the GDN.

[0064] Regulator scheduler 106 estimates an optimal variation of regulator pressure for one or more of regulator stations 108 over time for a predetermined period of time, such as an upcoming day or days, that minimizes or at least reduces excess pressure while ensuring, as much as possible, that all low point pressures remain above a desired minimum pressure value. In an exemplary embodiment, regulator scheduler 106 estimates an optimal variation of regulator pressure over time for an upcoming day.

[0065] It should be understood that the demand forecaster 104 and the regulator scheduler 106 are logical entities defined by the functions they perform and may be implemented in the same or different devices. Fig.14 In the exemplary embodiment explained in detail, both the demand forecaster 104 and the regulator scheduler 106 are implemented in a remote computing device.

[0066] As part of the application phase, the regulator scheduler 106 commands one or more regulator stations 108 in the GDN to set corresponding pressures in order to obtain the estimated optimal change in regulator pressure over time determined by the regulator scheduler 106. The one or more regulator stations 108 change the gas pressure of the gas input into the one or more regulator stations 108 according to the pressure received during the predetermined time period. Gas flows out of the one or more regulator stations 108 according to the regulator pressure setting for the predetermined time period. One or more low point recorders record one or more pressure low points of the gas pressure. One or more of the regulator stations 108 and / or one or more of the low point recorders 112 are configured to store the regulator pressure and / or the low point pressure, respectively, in the database 114.

[0067] like Figure 4As shown, one or more of the regulator stations 108 and / or one or more of the low point recorders 112 can provide feedback to the demand forecast feedback agent 120 and the regulator scheduling feedback agent 116. The demand forecast feedback agent 120 and the regulator scheduling feedback agent 116 can be logical entities that exist in the cloud. The demand forecast feedback agent 120 can determine the actual measured gas demand based on the feedback received from one or more regulator stations 108 and one or more low point recorders 112. The demand forecast feedback agent 120 can receive the predicted gas demand agent from the demand forecaster 104. The demand forecast feedback agent 120 can score the degree of execution of the machine learning algorithm based on the degree of match between the measured demand and the predicted demand agent. The demand forecast agent 120 can determine the parameters in the refined parameterized model based on the score. The refinement of the parameters can include retraining the machine learning algorithm. The demand forecast agent 120 can provide the refined predicted gas demand agent to the demand forecaster 104. The regulator scheduling feedback agent 116 can determine the measured overpressure based on the feedback received from one or more regulator stations 108 and one or more low point recorders 112. Based on the measured overpressure, the regulator of the regulator scheduling feedback agent 116 can score the degree to which the overpressure is reduced. The regulator scheduling feedback agent 116 can determine to refine the optimized regulator pressure based on the score. For example, if the overpressure is not significantly reduced, the regulator scheduling feedback agent 116 can determine to reduce the regulator pressure. The regulator scheduling feedback agent 116 can send the refined optimized regulator pressure to the regulator scheduler 106. In some embodiments, the regulator scheduling feedback agent 116 refines the optimized regulator pressure based on the feedback and the consumer-defined goals 118. The consumer-defined goals can include one or more of the following:

[0068] Minimum service pressure;

[0069] The minimum and / or maximum regulator pressure range for one or more regulator stations; and

[0070] The maximum pressure variance (alternatively referred to as "set point variance") of the pressure to which the regulator stations of one or more regulator stations are set.

[0071] In some embodiments, meeting more of the consumer-defined goals 118 results in a higher score, while meeting fewer of the consumer-defined goals 116 results in a lower score.

[0072] As will be appreciated, if a larger number of learning data sets are used, the accuracy of the gas demand forecast and low point pressure model will be improved. In this way, the gas demand forecast and low point pressure model can be continuously modified and improved. The customer defined goal 118 may include an efficiency indicating the amount of gas leakage prevented by using embodiments of the present disclosure.

[0073] Forecasting demand using a gas demand proxy (demand forecasting phase)

[0074] As described above, as a first step, a measured gas demand allocation is generated based on measurements taken from a selected GDN to which the example embodiment will be applied. In an exemplary embodiment, a gas demand proxy is used as the gas demand allocation. As will be explained below, the gas demand proxy is a measurement of gas demand in the GDN. The measured gas demand proxy is parameterized to reduce the number of parameters that are subsequently used to generate a predicted gas demand proxy for the GDN for a predetermined time period. The predicted gas demand proxy for the predetermined time period is predicted based on environmental conditions and is used to generate a data set for downloading to a regulator station. Since the data set is reduced based on the parameterized prediction of the gas demand proxy, the regulator can receive the data set in advance via a low bandwidth network (such as, for example, a mobile communication network).

[0075] Measurements for selected networks to generate demand proxies with respect to time

[0076] According to an example embodiment, a gas demand proxy is used as a measure of gas demand in the GDN. In an example embodiment, the gas demand proxy is given by Equation 1. It should be understood that the gas demand proxy is a measure of gas demand, and any other measure of gas demand may be used, as will be understood by those skilled in the art.

[0077] Equation 1. Demand Proxy ≈ Governor Pressure - Lowpoint Pressure

[0078] The demand proxy in Equation 1 has pressure units (e.g., bar, Pa, etc.). In Equation 1, the regulator pressure and the low point pressure are measured values ​​that vary with respect to time for the selected GDN to be controlled. If there are multiple regulator stations 12 in the GDN 2, the regulator pressure can correspond to an average regulator pressure obtained by summing the regulator pressures of each of the multiple regulator stations in the GDN and dividing by the number of regulator stations. If there are multiple low point recorders 6 in the GDN 2, the low point pressure can correspond to an average low point pressure obtained by summing the low point pressures of each of the multiple low points in the GDN and dividing by the number of low points. Each quantity in Equation 1 is measured as a function of time.

[0079] Figure 5A and Figure 5B is a graph showing the relationship between low point pressure and nighttime temperature. Figure 5A is a graph of average low pressure versus time for a cold winter and warm winter 24 hour period for a selected GDN and a constant regulator pressure. In particular, curve 50 represents the change in average low pressure for a warm winter (with a nighttime temperature of approximately 7.0°C) over a 24 hour period, and curve 52 represents the change in average low pressure for a cold winter (with a nighttime temperature of approximately 0.4°C) over two different 24 hour periods for the same GDN. As will be seen from Figure 5A As will be appreciated, the cold winter curve 52 and the warm winter curve 50 have substantially the same shape. For example, the lowest average low pressure 56 of the cold winter curve 52 occurs at approximately the same time as the lowest average low pressure 54 of the warm winter curve 50 (approximately 07:30). The shape of the curves 50, 52 across a 24 hour period may be determined by the daily variation in consumer demand for gas. For example, around 07:30, consumer demand may be at its highest value over a 24 hour period. Thus, the average low pressure of the GDN is at its lowest point around 07:30. Gas demand around 07:30 may be particularly high because a large number of household heating systems are being turned on when consumers are up during the day. As will be seen from Figure 5A As will be appreciated, the average low pressure of the cold winter curve 52 is generally lower over the entire 24 hour period than the average low pressure of the warm winter curve 50. In other words, there is a negative correlation between nighttime temperature and the average low pressure of the GDN. The decreasing average low pressure as nighttime temperatures decrease may occur because consumer demand for gas increases as nighttime temperatures decrease. For example, when nighttime temperatures are lower, a home heating system may need to use more gas to achieve a desired indoor temperature.

[0080] Figure 5B is a graph showing the daily average low pressure drop relative to nighttime temperature. The daily average low pressure drop is the drop in low pressure from a reference value over a 24-hour period. Figure 5B As can be seen, the daily average pressure drop decreases as nighttime temperatures increase. Figure 5B As shown, the daily average pressure drop for the cold winter curve 52 is higher than the daily average pressure drop for the warm winter curve 50. The increased daily average low pressure drop at lower nighttime temperatures can be explained by the increased consumer demand for gas at lower nighttime temperatures as described above.

[0081] As referenced above Figure 5A and Figure 5BAs explained, consumer demand may depend on nighttime temperature and time of day. However, it should be understood that these are examples and consumer demand may depend on other factors, such as wind speed, humidity, etc.

[0082] Figure 6 is a diagram with three graphs showing the evolution of temperature, network pressure and gas demand in the GDN over a week. Each data point is derived from a six-minute average. Figure 6 In the example above, there are 240 data points per day because there are six average data points every six minutes. The curve above shows how the temperature changes over time during a week. Figure 6 As can be appreciated, the nighttime average temperature 82 appears to be substantially constant over the week. However, the air temperature 80 varies over the week. The middle graph shows the variation in network pressure with respect to time of the week. The regulator pressure average 84 is substantially constant because, in this example, the regulator pressure is manually set at the regulator station. However, the low point average pressure 86 varies with time. The lower graph shows the variation in measured gas demand proxy 88 over the week. The measured gas demand proxy 88 is measured using the average low point pressure 86 and the average regulator pressure 84, as shown in Equation 1. This graph clearly shows the variation in gas demand commensurate with ambient conditions. For example, during Figure 6 A high demand period 81 is marked on the graph. The high demand period 81 occurs approximately in the early hours of Monday morning. The air temperature 80 at this time is at its lowest point of the week. The mean low pressure 86 at the same time is also at its lowest point of the week, and the measured gas demand proxy 88 is at its highest point of the week. In addition, Figure 6A low demand period 83 is marked on the top. The low demand period 83 occurs approximately in the early hours of Friday morning. The air temperature 80 during the low demand period 83 is higher than the air temperature 80 during the high demand period 81. In addition, the gas demand proxy 88 measured during the low demand period 83 is lower than the gas demand proxy 88 measured during the high demand period 81, so it will be understood that there is a negative correlation between air temperature and gas demand. In other words, gas demand increases as air temperature decreases, and vice versa. As described above, this may be because lower air temperatures mean that more consumers turn on their home heating systems or generally stay indoors and use gas appliances. Conversely, since consumers tend to spend more time outdoors and avoid turning on their home heating systems, gas usage decreases for higher air temperatures. As will be understood, many home gas heating systems operate on a thermostat and do not require consumers to manually turn on the thermostat in response to cold weather. Thermostats don't usually trigger the heating system at night, so cooler nights mean a colder house, and when the thermostat triggers the heating system to turn on in the morning, the heating system has to work harder to heat it back to the set temperature.

[0083] Requirement Agent for Parameterized / Modeled Measurements

[0084] In order to reduce the number of proxies that must be used to characterize gas demand (such as Figure 6 According to an example embodiment, a machine learning algorithm is used to predict a gas demand proxy for a selected GDN based on the number of parameters of the measured gas demand proxy (as shown in FIG. 1 ). To generate a predicted gas demand proxy for the selected GDN within a predetermined time period, the machine learning algorithm may be trained using measurements obtained from the selected GDN at a previous time. In particular, the machine learning algorithm may be trained using a parameterized model of the measured gas demand proxy and the corresponding environmental conditions. For example, the machine learning algorithm may be trained using Figure 6 The measured demand proxy 88 and the corresponding air temperature 80 are used to train the machine learning algorithm.

[0085] Figure 7 An example method for using machine learning to generate a parameterized model of a measured gas demand proxy is shown. The parameterized model of the measured gas demand proxy may also be referred to herein as a "parameterized measured gas demand proxy." Then, according to an exemplary embodiment, the parameterized model may be used to predict the gas demand proxy for a selected network. As a first step, as described above with reference to Figure 6 The average low point pressure, average regulator pressure and corresponding environmental conditions of the selected network during the measurement period are measured S72. "Corresponding environmental conditions" include any environmental conditions of the area of ​​the GDN during the measurement period mentioned herein.

[0086] In step S74, the regulator pressure and the low point pressure are used to calculate a measured gas demand proxy for the measurement period using Equation 1. Figure 6 An example of a measured gas demand proxy with corresponding measured ambient conditions is given in . Figure 6 In the example, the measurement period is one week and the “corresponding measured ambient conditions” is an air temperature of 80°C.

[0087] In an exemplary embodiment, a number of gas demand proxies are measured for different measurement periods. For example, if average regulator pressure, average low point pressure, and corresponding measured ambient conditions are available for a number of days, a measured gas demand proxy can be calculated for each day.

[0088] In step S76, the measured gas demand proxy is parameterized / modeled. Parameterizing the measured gas demand proxy may result in a reduction in computer processing power (e.g., a large amount of computing processing power will be required to predict Figure 6 Parameterization may include fitting a parameterized curve to the measured demand proxy. Parameterized curve 1002 may be composed of three highly scaled skewed normal distributions. For example, Figure 8 As shown, parameterized curve 1002 is composed of distribution A 1004, distribution B 1006, and distribution C 1008, each of which is a highly scaled skewed normal distribution. In some embodiments, each distribution 1004, 1006, 1008 can be represented by four parameters:

[0089] the location of the peak of the distribution (e.g., the x-axis position of the peak),

[0090] the height of the distribution peak (e.g., the y-axis position of the peak),

[0091] the scale of the peak of the distribution (alternatively called the “spread” of the distribution), and

[0092] The shape of the distribution (alternatively called the "skewness" of the distribution).

[0093] In some embodiments, the parameters defining the height of distribution C 1006 are determined by the height of distribution A 1004 and the height of distribution B 1006. In other words, the height of distribution C 1006 is represented by two parameters. For example, the height of distribution C 1006 can be defined as: (n*height of distribution A 1004 height) + (m*height of distribution B 1008), where n and m<1. Therefore, in some embodiments, the parameterized curve is represented by 13 parameters. Nine of the parameters correspond to network parameters that are constant for the GDN. For example, for a particular GDN, the nine parameters may be fixed on a daily basis and may be different for different GDNs. Four parameters of the GDN are variable depending on the prevailing environmental conditions. In one example, the four variable parameters are the height and positioning of distribution A 1004 and distribution C 1006. Therefore, by parameterization, only four parameters will need to be predicted to obtain a predicted gas demand proxy for the GDN over a predetermined time period.

[0094] In step S78, the machine learning algorithm uses the parameterized measured gas demand proxy for each day together with the corresponding measured environmental conditions for each day to establish one or more corresponding relationships between the measured environmental conditions and the four variable parameters. In other words, the machine learning algorithm finds the relationship between the four variable parameters and the measured environmental conditions. For example, if 13 parameters are used to parameterize Figure 6 , the machine learning algorithm can determine that there is a negative correlation between the air temperature 80 and the measured gas demand proxy. In other words, the machine learning algorithm can determine a relationship between four parameters that vary with the GDN as a function of the prevailing ambient conditions, such that if the outside temperature is low, the four parameters result in an increased gas demand prediction. It should be understood that a machine learning algorithm that establishes a correspondence between measured ambient conditions and measured gas demand proxies is an example of a "training" machine learning algorithm. It should be understood that any suitable machine learning algorithm can be used. For example, a gradient boosted decision tree (XGBoost) algorithm can be used. In one example, the ambient conditions are input into a gradient boosted decision tree algorithm, which outputs four variable parameters (e.g., the height and positioning of distribution A 1004 and distribution C 1006). The four variable parameters are then fed into a parameterized model to generate the following: Figure 8 Parameterized curve 1002 is shown. The gradient boosted decision tree algorithm and parameterized curve 1002 can then be calibrated / trained with weeks / months of data from the GDN.

[0095] In step S80, the trained machine learning algorithm uses the established correspondence between the four variable parameters and the environmental conditions to predict the values ​​of the four parameters for the predetermined time period based on the predicted environmental conditions for the predetermined time period. The predicted environmental conditions may be received from a weather forecast. In other words, the trained machine learning algorithm generates a predicted gas demand proxy based on the predicted environmental conditions. As will be appreciated by those skilled in the art, the predicted environmental conditions may include one or more of temperature, wind speed, air pressure, humidity, time of year, etc.

[0096] In step S82, according to an example embodiment, feedback is provided to the machine learning algorithm 82. The feedback may include measured regulator pressure and measured low point pressure, which may be used to generate a measured gas demand proxy in the GDN within a predetermined time period and corresponding measured environmental conditions. The measured gas demand proxy and the corresponding measured environmental conditions may be used to retrain the machine learning algorithm. In other words, the measured gas demand proxy and the corresponding measured environmental conditions for the predetermined time period may be used in combination with the measured gas demand proxy and the corresponding environmental conditions for the measured time period to establish one or more correspondences between the four variable parameters and the environmental conditions. For example, the measured demand proxy may be calculated from Equation 1 using the regulator pressure and the low point pressure received in the feedback. The measured demand proxy and the measured environmental conditions received from the weather bureau 102 within the predetermined time period may be fed into the machine learning algorithm. A parameterized curve may be fitted to the measured gas demand proxy for the predetermined time period with 13 parameters. The machine algorithm may be retrained to obtain an improved correspondence between the four parameters and the environmental conditions.

[0097] As will be appreciated, inputting a greater number of measured gas demand proxies with corresponding measured ambient conditions results in a higher accuracy prediction.

[0098] Predict low point pressure using low point pressure model (regulator pressure calculation stage)

[0099] Once the measured gas demand proxy for a selected GDN with respect to time during the measurement period has been determined, a predicted demand proxy may be generated based on variable parameters derived from the weather forecast, as described above in Figure 7 Next, the low point pressure model uses the predicted demand proxy to generate an estimate of the low point pressure in the selected GDN with respect to time, which is generated by a set of given regulator pressures over a predetermined time period. The low point pressure model can then be used to establish an estimate for the optimized regulator pressure setting.

[0100] Typically, the low point pressure in a GDN is a function of multiple factors, including, but not limited to: the number of regulators serving the low point, the number of consumers adjacent to the low point. Therefore, in some GDN networks, the low point pressure depends on more than one regulator. In an embodiment of a GDN with two pressure low points and two regulator stations, the following equation 2 can be used as a low point pressure model to predict the pressure at two different low points in the GDN:

[0101] Equation 2.

[0102]

[0103] LP 1 =G 1 M a +G 2 M b -dD 1

[0104] LP 2 =G 1 M c +G 2 M d -dD 2

[0105] LP 1 = Low pressure at the first pressure low point at a given time

[0106] LP 2 = Low pressure at the second low pressure point at a given time

[0107] GP 1 = regulator station pressure at the first regulator station at a given point in time

[0108] GP 2 = regulator station pressure at the second regulator station at a given point in time

[0109]

[0110]

[0111] d = proxy for gas demand at a given point in time

[0112] Although Equation 2 relates to a GDN including two regulator stations and two low pressure points, those skilled in the art will appreciate that Equation 2 may be applicable to a GDN having one regulator station and one low pressure point, or to a GDN having more than two regulator stations and more than two low pressure points.

[0113] The regulator coefficient M and demand coefficient d are constants for the selected GDN and satisfy 0<M, d<1. The regulator coefficient represents the responsiveness of the low point to changes in regulator pressure. The demand coefficient represents the responsiveness of the low point to changes in demand. For example, if D 1 Greater than D 2 , then pressure LP 1 The change of the ratio pressure LP by d 2 This could be, for example, because LP 1 Proximity Ratio LP 2 There are larger numbers of consumers nearby.

[0114] The quantity "d" is a proxy for gas demand and is a function of time. When using Equation 2 to determine the coefficients M and D (as explained below), "d" is a proxy for the measured gas demand, and LP 1 LP 2 , GP 1 and GP 2 are the measured low point pressure and the regulator pressure. Once M and D are known, equation 2 is used to estimate the low point pressure resulting from a given set of regulator pressures. The low point pressure model can then be used to estimate the optimal regulator pressure (as explained below). In this case, "d" is a proxy for the predicted gas demand, and LP 1 LP 2 is set by the given regulator pressure GP 1 , GP 2 The estimated low point pressure generated.

[0115] Fig. 9 is a flow chart showing how to use a low point pressure model (such as Equation 2) to obtain an estimate of the optimum regulator pressure. After the starting point, constants M and D of the GDN are determined in step S92. M and D can be determined by numerically solving Equation 2. As described above, M and D are constants of the GDN, and LP 1 LP 2 , GP 1 , GP 2 and d are functions of time. Therefore, we can 1 LP 2 , GP 1 , GP 2The values ​​of and d are used to find numerical solutions for the constants M and D, thereby generating a large number of simultaneous equations. For M and D, the simultaneous equations can be solved by any numerical solution technique known to those skilled in the art. For example, a sequential least squares programming algorithm or any other constrained numerical optimization technique can be used. The numerical solution technique for M and D can be run as part of a machine learning algorithm that is used to establish a correspondence between environmental conditions and gas demand proxies. LP 1 LP 2 , GP 1 , GP 2 The values ​​of and d are obtained by measuring the gas pressures at the first and second low points and the first and second regulator stations, respectively.

[0116] After M and D have been determined for the selected GDN, the determined M and D are combined with the predicted demand proxy for d (e.g., Figure 7 The predicted demand proxy generated in step S80) is input into Equation 2.

[0117] In step S94, by inserting M, D, and d into Equation 2, the equation can be used to simulate the low point pressure generated by a given set of regulator pressures. In other words, Equation 2 can be used as a low point pressure model. For example, GP 1 and GP 2 The change with respect to time can be inserted into Equation 2 along with the change with respect to time of the predicted demand proxy for d. Equation 2 then simulates the LP that will occur in the GDN 1 and LP 2 Changes relative to time.

[0118] Fig.10 An example of a simulation of the low point pressure model according to Equation 2 is shown for five low points 1200, 1202, 1204, 1206, 1208 in the GDN. Fig.10 It is understood that for a given set of regulator pressures, the simulated low point pressures substantially match the observed low point pressures. This simulation can be used to verify that the machine learning algorithm used to determine d performs adequately and that M and D have been determined to sufficient accuracy.

[0119] Return to reference Fig. 9 The low point pressure model is used to simulate the low point pressure generated by a given set of regulator pressures until the regulator pressure is approximately optimized in step S96. As an example of estimating the optimal regulator pressure, an embodiment can be used to minimize overpressure to ensure that the minimum legal gas pressure requirements are met while minimizing financial losses and environmental impacts caused by gas leaks. Fig.11A graph of a GDN with one regulator station and one low point pressure point is shown, showing the variation of regulator pressure at the regulator station that minimizes overpressure at the low point, according to an exemplary embodiment.

[0120] from Fig.11 It will be appreciated that if the regulator pressure 30 varies as shown in the graph, the low point pressure 32 generated at the low point in the GDN is substantially constant and close to the minimum consumer pressure 20. In other words, if the regulator pressure 30 varies as shown in the graph, the low point pressure 32 generated at the low point in the GDN is substantially constant and close to the minimum consumer pressure 20. Fig.11 The overpressure 22 is minimized if the change shown is obtained. Thus, the low point pressure model (Equation 2) can be used to repeatedly input multiple distributions of regulator pressure 30 with respect to time, determine the resulting low point pressure distributions with respect to time, and determine the change in regulator pressure 30 with respect to time that results in the minimum overpressure 22 with respect to time. As an example, the area 24 between the low point pressure distribution 32 and the minimum consumer pressure 20 within a predetermined time period can be minimized.

[0121] Currently, if Figure 2 As shown, the regulator pressure is manually varied at the regulator station 12. Therefore, it would be impractical to manually and continuously vary the regulator pressure 30 to achieve the minimum overpressure 22.

[0122] Thus, embodiments may be used to compensate for a decrease in low point pressure 32 caused by increased demand by forecasting a proxy for gas demand in the GDN over a predetermined time period, simulating the low point pressure that would result from a given set of regulator pressures, determining an estimate of the optimal regulator pressure, and implementing the estimate of the optimal regulator pressure at the regulator station. For example, if the low point pressure 32 is expected to decrease due to an expected increase in consumer demand, the regulator pressure 30 should be increased in advance of this event to ensure that overpressure is minimized 22.

[0123] Fig.12 A comparison between using five different constant regulator pressures 1300 and five different variable regulator pressures 1310 in a GDN is shown according to some embodiments. Fig.12 Five different low point pressure curves 1302 produced by five different constant regulator pressures 1300 and five different low point pressure curves 1312 produced by five variable regulator pressures 1310 are shown. In both graphs, a critical safety threshold 1304 is shown. An overpressure 1306 of one of the low point pressures 1302 produced by the constant regulator pressure 1300 is also shown as the difference between one of the low point pressures 1302 and the safety threshold 1304. An overpressure 1314 of one of the low point pressures 1312 produced by the variable regulator pressure 1310 is also shown as the difference between one of the low point pressures 1312 and the safety threshold 1304.

[0124] As described above, the low point pressure model can be used to predict the low point that will result from a given set of regulator pressures.Exemplary embodiments can minimize or at least reduce overpressure of a selected GDN.

[0125] As will be appreciated by those skilled in the art, whether the estimated regulator pressure is "optimal" may depend on a number of factors. For example, the optimal regulator pressure may be the regulator pressure that results in the least overpressure of the GDN. Fig.12 , a measure of the overpressure of each low point curve 1302 is given by the area area between the corresponding low point curve and the minimum threshold 1304. For example, area 1308 represents a measure of the overpressure of one of the low point pressure curves 1302 produced by one of the constant regulator pressures 1300, and area 1316 represents a measure of the overpressure of one of the low point pressure curves 1312 produced by one of the variable regulator pressures 1310. One or more optimization techniques may attempt to minimize the sum of the areas between each low point pressure curve of the GDN and the minimum threshold 1304. The one or more optimization techniques may be any constrained numerical optimization algorithm, such as a sequential least squares programming algorithm.

[0126] In other examples, one or more optimization techniques may attempt to minimize the cost value. In other words, the optimal regulator pressure is the regulator pressure that produces the lowest cost value. The cost value may be consumer-dependent. A high cost may be assigned to a regulator pressure that results in a low point pressure below a minimum threshold. A low cost may be assigned to a regulator pressure that results in a low point pressure above a minimum threshold but close to the minimum threshold. An indication of the amount of gas leakage saved by employing an embodiment of the present disclosure may be sent to one or more consumers. For example, the indication may include an efficiency as shown in Equation 3.

[0127] Equation 3.

[0128] In an example embodiment, the constants M and D may be recalculated. For example, the machine algorithm may be retrained based on feedback including measured regulator pressure and low point pressure, and M and D may be recalculated using the retrained machine learning algorithm.

[0129] Application Phase

[0130] Once an estimate of the optimal regulator pressure has been determined for the selected GDN, the estimated optimal regulator pressure over time is achieved at the corresponding regulator station in the selected GDN for a predetermined period of time.Exemplary embodiments provide a remote assembly that adjusts pressure at a regulator station in a GDN.

[0131] Typically, the delivery pressure or set point of a regulator station is usually adjusted by changing the position of a fixed stop within a pilot valve, against which a spring forming part of the pilot valve reacts. The accuracy of a fixed stop controlled pilot valve is generally good. However, if any adjustment to the delivery pressure is required, it is necessary to access the regulator to make a manual adjustment to the pilot fixed stop position. However, since the estimated optimum regulator pressure is variable over a predetermined period of time, changing the regulator pressure in this way is impractical both in terms of cost and logistics. Therefore, a remote assembly for regulator pressure adjustment is provided.

[0132] To provide a remote component for adjustment and to make adjustments easier, Fig.13 An actuator having a controller configured to receive adjustment instructions is provided.

[0133] exist Fig.13 3, a motor and gear head assembly 306 controlled by a controller 328 and an external power source 302 rotates a shaft 308. The shaft passes through but is not connected to a bracket 320. The shaft 308 is connected to a sun gear 312, which forms part of the differential gear system 3.

[0134] The differential gear arrangement 314 is formed by a sun gear 312, two planetary gears 316 and 318 and a further sun gear 322. The planetary gears 316 and 318 are mounted on a carrier 320 which is arranged to rotate about the motor shaft 308 via bearings 338 mounted between the carrier and the shaft. The carrier 320 has a ring gear 310 cut into its circumference and is normally held in a fixed position by engagement with a spur gear 336 connected to an emergency motor gear head assembly 334.

[0135] In normal operation of the gear system, the sun gear 312 then rotates the planetary gears 316. The planetary gears 316 then rotate the planetary gears 318, which in turn rotate the sun gear 322. The sun gear 322 is attached to the actuator shaft 340, which is connected to the rotary to linear device (RLD) RLD346 via a coupling 324. The role of the coupling 324 is to allow linear motion of the rotary to linear device 346 without affecting the linear position of the actuator shaft 340. Therefore, the rotation of the motor 306 is converted into the rotation of the rotary to linear device. The rotation of the rotary to linear device 346 converts its rotational motion into linear motion, and the linear motion is transmitted to the regulator stop 348, which in turn affects the outlet pressure of the controlled valve gas regulator as required. The position of the rotary to linear device 346 is sensed by an encoder 342 that monitors the angular position or linear position of the encoder wheel 344.

[0136] The emergency motor gear head 334 has a power supply independent of the main power supply, formed by an independent backup battery 332 and a normally closed relay 330. The relay 330 is normally maintained in the open position by the action of the controller 328.

[0137] It should be understood that Fig.13 The described device represents one example of a pressure control assembly that can be used to control the pressure at the regulator station. However, the pressure control assembly is not limited, and those skilled in the art will understand that other pressure control assemblies can be used.

[0138] According to an exemplary embodiment, the controller 328 can be configured to control the fluid pressure based on the received information (such as the regulator station settings). The regulator station may also include communication circuitry including transceiver circuitry configured to send and / or receive signals. The communication circuitry may include controller circuitry configured to control the transceiver circuitry. As will be described below with reference to Fig.14 Explained in more detail, the communication circuit can be configured to communicate with a computing device for generating an estimate of an optimal regulator pressure using a mobile communication network. The communication circuit can receive the estimated optimal regulator pressure from the computing device and forward the received optimal regulator pressure to the pressure control component. In response, the pressure control component implements the estimated optimal regulator pressure over time within a predetermined time period.

[0139] Exemplary Fluid Control Systems

[0140] Fig.14 is a schematic diagram showing an exemplary embodiment of a fluid control system according to an embodiment of the present disclosure. Fig.13 In the present invention, a single regulator station 218 and a single low point recorder 208 are considered in the GDN, but it should be understood that one or more regulator stations and / or one or more low point recorders may be used. Fig.13 As will be appreciated, the predicted environmental conditions 216 are provided to the computing device 202. The predicted environmental conditions 216 may be sent to the computing device via any wireless communication technology known in the art. As described above, the predicted environmental conditions 216 may be provided to the computing device 202 by the weather bureau. The computing device 202 may perform the following operations: Figure 4 The computing device 202 may include some or all of the functionality of the demand forecaster 104, the regulator scheduler 106, and / or the database 114 shown. In other words, the computing device 202 may determine an estimated optimal change in regulator pressure over time to be achieved at the regulator station 218 over a predetermined time period. The computing device 202 includes a receiver circuit configured to receive a signal, a transmitter circuit configured to transmit a signal, and a controller circuit configured to control the receiver and the transmitter. The transmitter circuit and the receiver circuit may alternatively be referred to as transceiver circuits that are configured to transmit and / or receive signals. Figure 1 As shown, the computing device 202 is configured to communicate with the communication circuit 206 of the regulator station 218. The computing device 202 sends a signal 210 to the communication circuit 206, which includes instructions to the regulator station 218 to achieve the estimated optimal change in regulator pressure over time within a predetermined time period. For example, the signal 210 may include regulator settings to be implemented at the regulator station 218 to achieve the estimated optimal change in regulator pressure over time within the predetermined time period. The signal 210 is received by the communication circuit 206 in the regulator station.

[0141] In response to receiving signal 210, communication circuit 206 in regulator station 218 forwards an indication of signal 210 to pressure control component 204 in regulator station 218. For example, regulator station settings to be implemented may be forwarded to pressure control component 204. In response, pressure control component 204 implements the received settings. Pressure control component 204 may broadly correspond to Fig.13 The device shown in .

[0142] The pressure control assembly 204 implements the received settings to vary the pressure of the intermediate gas pressure supply input to the regulator station 218 to achieve the estimated optimal change in regulator pressure over time over a predetermined period of time.

[0143] The regulator station 218 may include a measurement component to measure the pressure of the gas leaving the regulator station. The measurement component provides the pressure of the gas leaving the station to the communication circuit 206. The communication circuit 206 in the regulator station 218 may then send a signal 212 including an indication of the pressure of the gas leaving the regulator station over time to the computing device 202. The gas leaving the regulator station 218 then flows through the rest of the GDN, and the pressure of the gas is measured at the low point by the low point recorder 208. The low point recorder may send a signal 214 containing an indication of the measured low point pressure to the computing device 202. It should be understood that the low point pressure data may not be available until a certain time after the low point has been measured. In this case, the process of sending / receiving the low point pressure occurs substantially when the data becomes available. The signals 212, 214 sent to the computing device 202 from the communication circuit 206 in the regulator station 218 and the low point recorder 208, respectively, contain feedback including the measured regulator pressure and the measured low point pressure that can be stored by the computing device 202.

[0144] It should be understood that the words "providing", "sending" and "receiving" used herein refer to the communication process. Any wired or wireless communication components can be used for the described communication. For example, communication can be through Wi-Fi, WLAN, Ethernet, etc.

[0145] Fail-safe

[0146] In an example embodiment, if the transmission of the signal 210 including the regulator setting from the computing device 202 to the communication circuit 206 does not occur at a predetermined point in time, the fail-safe mode is initiated. In other words, if the regulator station 218 does not receive instructions from the computing device 202, the pressure control component 204 operates according to the fail-safe instructions stored by the data storage component at the regulator station 218. The fail-safe instructions may include predetermined instructions. For example, if the communication circuit 206 determines that it has not received instructions from the computing device 202 at a predetermined point in time, it may send this indication to the pressure control component 204. In response, the pressure control component 204 may control the regulator pressure to match the predetermined seasonal distribution.

[0147] Fig.15 is a flow chart illustrating a method performed by a computing device for communicating with one or more pressure control stations to control fluid pressure in a fluid distribution network FDN.

[0148] In step S1502, the computing device trains a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN during a first time period and environmental conditions measured during the first time period. For example, machine learning can be used to create a predictive model of demand for the FDN using historical pressure and weather data.

[0149] In step S1540, based on one or more correspondences established between the predicted environmental conditions of the FDN in the second time period and the measured changes in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, the computing device uses a trained machine learning algorithm to predict changes in fluid demand in the FDN in a later second time period.

[0150] In step S1506, based on the predicted change in fluid demand over the second time period, the computing device determines a change in fluid pressure to be applied at one or more pressure control stations over the second time period to satisfy the fluid pressure condition at one or more predetermined points in the FDN downstream of the one or more pressure control stations. For example, machine learning can be used to create a limited / dedicated digital twin to model the relationship between pressure at the pressure control station, pressure at the predetermined point in the FDN, and demand. Future predictions of demand can then be used to "query" the digital twin to find an allocation of pressure control station settings that satisfy the fluid pressure condition.

[0151] In some embodiments, the fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station are conditions that minimize overpressure in the FDN. For example, the overpressure may be the difference between the fluid pressure at a pressure low point at the end of the FDN and the minimum allowable fluid pressure. The minimum fluid pressure may be a statutory minimum fluid pressure. In such embodiments, fluid leakage may be reduced while maintaining consumer safety.

[0152] In some embodiments, the fluid pressure condition at one or more predetermined points in the FDN downstream of the pressure control station is a condition that minimizes the difference between the fluid pressure at the one or more predetermined points in the FDN and the maximum allowable fluid pressure. For example, where the fluid is biomethane gas, such an embodiment may maintain the fluid pressure at the one or more predetermined points at the maximum allowable fluid pressure. As previously described, if the pressure of the biomethane gas exceeds the maximum allowable fluid pressure, it becomes difficult to feed the biomethane into the FDN. Thus, exemplary embodiments may improve the ease of feeding biomethane into the FDN without having to burn excess gas through flaring.

[0153] In some embodiments, the fluid pressure condition at one or more predetermined points in the FDN downstream of the pressure control station is a condition where the fluid pressure at the one or more predetermined points falls within predetermined maximum and minimum fluid pressures. As will be appreciated by those skilled in the art, some countries require that the fluid pressure at predetermined points in the FDN be between the maximum and minimum allowable fluid pressures. Thus, exemplary embodiments can accurately control the fluid pressure between the maximum and minimum allowable fluid pressures.

[0154] The one or more predetermined points downstream of the one or more pressure control stations may be any point along the FDN that receives fluid from the one or more pressure control stations. For example, the one or more predetermined points may be a low pressure point at the end of the FDN.

[0155] In step S1508, the computing device sends to one or more pressure control stations an indication of the determined change in fluid pressure to be applied at the one or more pressure control stations during the second time period to satisfy the fluid pressure condition at one or more predetermined points in the FDN. For example, the indication may be an indication of the pressure control settings to be used by the one or more pressure control stations during the second time period. The settings for the pressure control stations may include changes in the set points to be applied by the pressure control stations during the second time period.

[0156] In some embodiments, feedback can be used to improve the machine learning algorithms. For example, the performance of each machine learning algorithm can be monitored and the model can be retrained with new data. The retrained model can then be tested against historical data. The pressure control station settings can then be output to adapt the accuracy of the measurement.

[0157] Those skilled in the art will understand that Fig.15 The method shown can be adjusted according to embodiments of the present technology. For example, other intermediate steps can be included in the method, or the steps can be performed in any logical order.

[0158] In some embodiments, the steps performed by the computing device may be performed by the pressure control station. For example, the pressure control station may perform steps S1502, S1504, and S1506. The pressure control station may then use a pressure control component to adjust the pressure of the fluid at the pressure control station based on the determined change in fluid pressure to be applied at the pressure control station during the second time period to meet the fluid pressure conditions at one or more predetermined points in the FDN. In such an embodiment, the pressure control station may include a pressure control component for adjusting the fluid pressure at the pressure control station, a transceiver circuit configured to send and / or receive signals, and a controller circuit configured to be combined with the transceiver circuit. In such an embodiment, the computing device may be considered to be included in the pressure control station.

[0159] The following numbered paragraphs provide further example aspects and features of the present technology:

[0160] Paragraph 1. A method performed by a computing device for communicating with one or more pressure control stations to control fluid pressure in a fluid distribution network FDN, the method comprising:

[0161] training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0162] using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0163] determining a change in fluid pressure to be applied at the one or more pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the one or more pressure control stations based on the change in fluid demand predicted during the second time period, and

[0164] An indication of the determined change in fluid pressure to be applied at the one or more pressure control stations during the second time period is sent to the one or more pressure control stations to satisfy the fluid pressure conditions at the one or more predetermined points in the FDN.

[0165] Paragraph 2. A method according to paragraph 1, wherein the instructions sent by the computing device include pressure control settings for the one or more pressure control stations to control the fluid pressure based on the changes in fluid pressure determined to be applied at the multiple pressure control stations during the second time period.

[0166] Paragraph 3. A method according to any one of paragraphs 1 to 2, wherein training the machine learning algorithm to establish one or more corresponding relationships between changes in the measured fluid demand during the first time period and the measured environmental conditions during the first time period comprises:

[0167] modeling the variation of the measured fluid demand during the first time period using a set of parameters including network parameters that are constant for the FDN and environmental parameters that depend on prevailing environmental conditions of the FDN,

[0168] The machine learning algorithm is trained to establish one or more corresponding relationships between the value of each of the environmental parameters obtained from the modeling and the measured environmental conditions during the first time period, wherein

[0169] Using a trained machine learning algorithm to predict the change in fluid demand during the second time period includes using a trained machine learning algorithm to predict the change in fluid demand during the second time period based on the predicted environmental conditions and one or more correspondences established between the values ​​of each of the environmental parameters obtained from the modeling and the measured environmental conditions of the FDN during the first time period.

[0170] Paragraph 4. The method of any one of paragraphs 1 to 3, wherein said determining said change in fluid pressure to be applied at said one or more pressure control stations during said second time period comprises:

[0171] creating a model for simulating fluid pressures at the one or more predetermined points in the FDN during a second time period for a given set of fluid pressures at the one or more pressure control stations during the second time period based on the predicted change in fluid demand during the second time period, and

[0172] Numerical techniques are used to estimate a change in fluid pressure to be applied at the one or more pressure control stations during the second period that will satisfy the predetermined pressure condition at the one or more predetermined points in the FDN.

[0173] Paragraph 5. The method of any of paragraphs 1 to 4, wherein the pressure condition is to minimize the sum of one or more areas between the curve of each simulated fluid pressure at the one or more predetermined points in the FDN and the minimum allowable fluid pressure in the FDN. Area

[0174] Paragraph 6. A method according to any one of paragraphs 1 to 5, comprising:

[0175] The change in the measured fluid demand during the first time period is determined based on a change in measured fluid pressure at the one or more pressure control stations during the first time period and a change in measured fluid pressure at the one or more predetermined points in the FDN during the first time period.

[0176] Paragraph 7. The method of paragraph 6, wherein determining a change in the measured fluid demand over the first time period based on a change in the measured fluid pressure at the one or more pressure control stations and a change in the measured fluid pressure at the one or more predetermined points in the FDN over the first time period comprises:

[0177] receiving from the plurality of pressure control stations measured fluid pressures at the plurality of pressure control stations during the first time period; and

[0178] Measured fluid pressures at the one or more predetermined points during the first time period are received from one or more devices located at the one or more predetermined points in the FDN.

[0179] Paragraph 8. A method according to any one of paragraphs 1 to 7,

[0180] Wherein said training the machine learning algorithm to establish one or more corresponding relationships between the changes in the measured fluid demand during the first time period and the measured environmental conditions during the first time period comprises:

[0181] Environmental conditions measured during the first time period are received.

[0182] Paragraph 9. A method according to any one of paragraphs 1 to 8, comprising:

[0183] A predicted environmental condition of the FDN during the second time period is received.

[0184] Paragraph 10. The method of any of paragraphs 1 to 9, wherein receiving the predicted environmental condition of the FDN during the second time period comprises:

[0185] The predicted environmental conditions are received as a weather forecast for the second time period.

[0186] Paragraph 11. A method according to any one of paragraphs 1 to 10, comprising:

[0187] receiving, after the second time period, from the one or more pressure control stations, a measured fluid pressure at the one or more pressure control stations during the second time period,

[0188] receiving, after the second time period, from one or more devices located at the one or more predetermined points in the FDN, fluid pressures measured at the one or more predetermined points in the FDN during the second time period,

[0189] receiving, after the second time period, the measured environmental conditions during the second time period,

[0190] determining a change in measured fluid demand during the second time period based on the measured fluid pressures at the one or more pressure control stations during the second time period and the measured fluid pressures at the one or more predetermined points in the FDN during the second time period,

[0191] The machine learning algorithm is retrained to establish one or more corresponding relationships between the measured environmental conditions during the second time period and changes in the measured fluid demand during the second time period.

[0192] Paragraph 12. The method according to paragraph 11, comprising:

[0193] The measured environmental conditions received during the second time period, the fluid pressures received at the plurality of pressure control stations during the second time period, and the fluid pressures received at the one or more predetermined pressure control stations during the second time period are stored.

[0194] Paragraph 13. The method of any of paragraphs 1 to 12, wherein one or more predetermined points in the FDN are fluid pressure low points.

[0195] Paragraph 14. The method of any one of paragraphs 1 to 13, wherein the second time period is a 24 hour period.

[0196] Paragraph 15. A method according to any one of paragraphs 1 to 14, wherein the fluid is a gas.

[0197] Paragraph 16. A computing device for communicating with one or more pressure control stations to control fluid pressure in a fluid distribution network FDN, the computing device comprising:

[0198] a transceiver circuit configured to transmit and / or receive signals, and

[0199] A controller circuit, the controller circuit in combination with the transceiver circuit being configured to:

[0200] training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0201] using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on the predicted environmental conditions in a second time period and one or more established corresponding relationships between the change in the measured fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0202] determining a change in fluid pressure to be applied at the one or more pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the one or more pressure control stations based on the change in fluid demand predicted during the second time period, and

[0203] An indication of the determined change in fluid pressure to be applied at the one or more pressure control stations during the second time period is sent to the one or more pressure control stations to satisfy the fluid pressure conditions at the one or more predetermined points in the FDN.

[0204] Paragraph 17. The computing device of paragraph 16, the instructions sent by the computing device comprising pressure control settings for the one or more pressure control stations to control fluid pressure based on the determined change in fluid pressure to be applied at the one or more pressure control stations during the second time period.

[0205] Paragraph 18. A computing device according to any one of paragraphs 16 to 17, wherein:

[0206] The controller circuit in combination with the transceiver circuit is configured to:

[0207] modeling the variation of the measured fluid demand during the first time period using a set of parameters, the set of parameters comprising network parameters that are constant for the FDN and environmental parameters that depend on prevailing environmental conditions of the FDN,

[0208] training the machine learning algorithm to establish one or more corresponding relationships between the value of each of the environmental parameters obtained from the modeling and the measured environmental conditions during the first time period,

[0209] A trained machine learning algorithm is used to predict changes in fluid demand during the second time period based on the predicted environmental conditions and one or more correspondences established between the values ​​of each of the environmental parameters obtained from the modeling and the measured environmental conditions of the FDN during the first time period.

[0210] Paragraph 19. A computing device according to any of paragraphs 16 to 18, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0211] creating a model for simulating fluid pressures at the one or more predetermined points in the FDN during a second time period for a given set of fluid pressures at the one or more pressure control stations during the second time period based on the predicted change in fluid demand during the second time period, and

[0212] Numerical techniques are used to estimate a change in fluid pressure to be applied at the one or more pressure control stations during the second period that will satisfy the predetermined pressure condition at the one or more predetermined points in the FDN.

[0213] Paragraph 20. A computing device according to any of paragraphs 16 to 20, wherein the pressure condition is to minimize the sum of one or more areas between the curves of each simulated fluid pressure at the one or more predetermined points in the FDN and the minimum allowable fluid pressure in the FDN.

[0214] Paragraph 21. A computing device according to any of paragraphs 16 to 20, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0215] The change in the measured fluid demand during the first time period is determined based on a change in measured fluid pressure at the one or more pressure control stations during the first time period and a change in measured fluid pressure at the one or more predetermined points in the FDN during the first time period.

[0216] Paragraph 22. The computing device of paragraph 21, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0217] receiving from the one or more pressure control stations a measured fluid pressure at the one or more pressure control stations during the first time period, and

[0218] Measured fluid pressures at the one or more predetermined points during the first time period are received from one or more devices located at the one or more predetermined points in the FDN.

[0219] Paragraph 23. A computing device according to any of paragraphs 16 to 22, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0220] The measured environmental conditions during the first time period are received.

[0221] Paragraph 24. A computing device according to any of paragraphs 16 to 23, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0222] A predicted environmental condition of the FDN during the second time period is received.

[0223] Paragraph 25. A computing device as described in any of paragraphs 16 to 24, wherein the controller circuitry in combination with the transceiver circuitry is configured to receive the predicted environmental condition as a weather forecast for the second time period.

[0224] Paragraph 26. A computing device according to any of paragraphs 16 to 25, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0225] receiving, after the second time period, from the one or more pressure control stations, a measured fluid pressure at the one or more pressure control stations during the second time period,

[0226] receiving, after the second time period, from one or more devices located at the one or more predetermined points in the FDN, fluid pressures measured at the one or more predetermined points in the FDN during the second time period,

[0227] receiving, after the second time period, the measured environmental conditions during the second time period,

[0228] determining a change in measured fluid demand during the second time period based on the measured fluid pressures at the one or more pressure control stations during the second time period and the measured fluid pressures at the one or more predetermined points in the FDN during the second time period,

[0229] The machine learning algorithm is retrained to establish one or more corresponding relationships between the measured environmental conditions during the second time period and changes in the measured fluid demand during the second time period.

[0230] Paragraph 27. The computing device of paragraph 26, comprising a storage component configured to

[0231] The measured environmental conditions received during the second time period, the fluid pressures received at the plurality of pressure control stations during the second time period, and the fluid pressures received at the one or more predetermined points during the second time period are stored.

[0232] Paragraph 28. A computing device according to any of paragraphs 16 to 27, wherein the one or more predetermined points in the FDN are fluid pressure low points.

[0233] Paragraph 29. A computing device according to any of paragraphs 16 to 28, wherein the second period is a 24-hour period.

[0234] Paragraph 30. A computing device according to any of paragraphs 16 to 29, wherein the fluid is a gas.

[0235] Paragraph 31. A method for controlling fluid pressure in a fluid distribution network FDN performed by a pressure control station, the method comprising:

[0236] receiving from a computing device an indication of a determined change in fluid pressure to be applied at the pressure control station within a future time period to satisfy fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station; and

[0237] Fluid pressure at the pressure control station is adjusted based on the indication received from the computing device.

[0238] Paragraph 32. A method according to paragraph 31, wherein the instructions received from the computing device include pressure control settings for the pressure control station to control fluid pressure based on the determined change in fluid pressure to be applied at the pressure control station over a future time period.

[0239] Paragraph 33. A method according to any of paragraphs 31 to 32, wherein the pressure condition is to minimize the sum of one or more areas between the curves of each simulated fluid pressure at the one or more predetermined points in the FDN and the minimum allowable fluid pressure in the FDN.

[0240] Paragraph 34. A method according to any one of paragraphs 31 to 33, comprising:

[0241] measuring the fluid pressure at the pressure control station during a previous time period before the future time period,

[0242] In advance of the future time period, an indication of the measured fluid pressure at the pressure control station during the prior time period is sent to the communication device.

[0243] Paragraph 35. A method according to any one of paragraphs 31 to 34, comprising:

[0244] measuring the fluid pressure at the pressure control station during the future time period,

[0245] An indication of the measured fluid pressure at the pressure control station during the future time period is sent to the communication device.

[0246] Paragraph 36. A method according to any of paragraphs 31 to 35, wherein the future period is a 24-hour period.

[0247] Paragraph 37. A method according to any one of paragraphs 31 to 36, wherein the fluid is a gas.

[0248] Paragraph 38. A pressure control station for controlling fluid pressure in a fluid distribution network FDN, the pressure control station comprising:

[0249] a pressure control assembly for adjusting the fluid pressure at the pressure control station,

[0250] a transceiver circuit configured to transmit and / or receive signals, and

[0251] A controller circuit, the controller circuit in combination with the transceiver circuit being configured to:

[0252] receiving from a computing device an indication of a determined change in fluid pressure to be applied at the pressure control station within a future time period to satisfy fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station, wherein the pressure control assembly is configured to:

[0253] Fluid pressure at the pressure control station is adjusted based on the indication received from the computing device.

[0254] Paragraph 39. A pressure control station as described in paragraph 38, wherein the instructions received from the computing device include pressure control settings for the pressure control station to control fluid pressure based on the changes in fluid pressure determined to be applied at the multiple pressure control stations during the future time period.

[0255] Paragraph 40. A pressure control station according to any of paragraphs 38 to 39, wherein the pressure condition is to minimize the sum of one or more areas between the curves of each simulated fluid pressure at the one or more predetermined points in the FDN and the minimum allowable fluid pressure in the FDN.

[0256] Paragraph 41. A pressure control station according to any one of paragraphs 38 to 40, comprising a measurement assembly configured to

[0257] measuring a fluid pressure at the pressure control station during a previous time period prior to the future time period, wherein the controller circuitry in combination with the transceiver circuitry is configured to:

[0258] In advance of the future time period, an indication of the measured fluid pressure at the pressure control station during the prior time period is sent to the communication device.

[0259] Paragraph 42. A pressure control station according to any one of paragraphs 38 to 41, comprising a measurement assembly configured to

[0260] measuring a fluid pressure at the pressure control station during the future time period, wherein the controller circuit in combination with the transceiver circuit is configured to:

[0261] An indication of the measured fluid pressure at the pressure control station during the future time period is sent to the communication device.

[0262]

[0066] Paragraph 43. A pressure control station according to any of paragraphs 38 to 42, wherein the future time period is a 24 hour period.

[0263] Paragraph 44. A pressure control station according to any of paragraphs 38 to 43, wherein the fluid is a gas.

[0264] Paragraph 45. A method for controlling fluid pressure in a fluid distribution network FDN, the method comprising:

[0265] training a machine learning algorithm, via a computing device, to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0266] using the trained machine learning algorithm by the computing device to predict a change in the fluid demand in the FDN in a later second time period based on the predicted environmental conditions in a second time period and one or more corresponding relationships established between the change in the measured fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0267] determining, by the computing device, a change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to satisfy fluid pressure conditions at one or more predetermined points in the FDN downstream of the plurality of pressure control stations based on the change in fluid demand predicted during the second time period,

[0268] transmitting, by the computing device, to the plurality of pressure control stations an indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations for the second period of time to satisfy the fluid pressure conditions at the one or more predetermined points in the FDN,

[0269] receiving, via the plurality of pressure control stations, an indication of a determined change in fluid pressure to be applied at the plurality of pressure control stations over a second time period to satisfy the fluid pressure condition at the one or more predetermined points in the FDN, and

[0270] The fluid pressure at the respective pressure control stations is adjusted, by the plurality of pressure control stations, based on the indication received from the computing device.

[0271] Paragraph 46. A system for controlling fluid pressure in a fluid distribution network FDN, the system comprising

[0272] One or more pressure control stations, and

[0273] A computing device, the computing device being configured to:

[0274] training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0275] using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0276] determining a change in fluid pressure to be applied at the one or more pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the one or more pressure control stations based on the predicted change in fluid demand during the second time period, and

[0277] sending an indication of the determined change in fluid pressure to be applied at the one or more pressure control stations during the second time period to the one or more pressure control stations, wherein the one or more pressure control stations are configured to:

[0278] receiving an indication of a change in the fluid pressure determined to be applied at the one or more pressure control stations during the second time period to satisfy the fluid pressure condition at the one or more predetermined points in the FDN, and

[0279] Fluid pressure at the corresponding pressure control station is adjusted based on the indication received from the computing device.

[0280] Paragraph 47. A method for controlling fluid pressure in a fluid distribution network FDN performed by a pressure control station, the method comprising:

[0281] training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0282] using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0283] determining a change in fluid pressure to be applied at the pressure control station during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station based on the predicted change in fluid demand during the second time period, and

[0284] The fluid pressure at the pressure control station is adjusted based on the determined change in fluid pressure to be applied at the pressure control station during the second time period to meet the fluid pressure conditions at the one or more predetermined points in the FDN.

[0285] Paragraph 48. A pressure control station for controlling fluid pressure in a fluid distribution network FDN, the pressure control station comprising:

[0286] a pressure control assembly for regulating the fluid pressure at the pressure control station,

[0287] a transceiver circuit configured to transmit and / or receive signals, and

[0288] a controller circuit, the controller circuit in combination with the transceiver circuit being configured to: train a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0289] using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0290] determining a change in fluid pressure to be applied at the pressure control station during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station based on the change in fluid demand predicted during the second time period, wherein the pressure control assembly is configured to:

[0291] The fluid pressure at the pressure control station is adjusted based on the determined change in fluid pressure to be applied at the pressure control station during the second time period to meet the fluid pressure conditions at the one or more predetermined points in the FDN.

[0292] Paragraph 49. A computer program comprising instructions which, when executed by a computer, cause the computer to perform the method of any one of paragraphs 1 to 15, 31 to 37, 45 and 47.

[0293] The following numbered paragraphs provide further example aspects and features of the present technology:

[0294] Paragraph 1. A method performed by a computing device for controlling a fluid pressure in a fluid distribution network (FDN) above a threshold pressure by communicating with a plurality of pressure control stations distributed throughout the FDN to independently control a fluid pressure at each of the plurality of pressure control stations, the method comprising:

[0295] training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period,

[0296] using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period,

[0297] determining a change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the plurality of pressure control stations based on the change in fluid demand predicted during the second time period, and

[0298] An indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations during the second time period is sent to the plurality of pressure control stations to satisfy the fluid pressure condition at the one or more predetermined points in the FDN.

[0299] Paragraph 2. A method according to paragraph 1, wherein the instructions sent by the computing device include pressure control settings for the multiple pressure control stations to control the fluid pressure based on the changes in fluid pressure determined to be applied at the multiple pressure control stations during the second time period.

[0300] Paragraph 3. A method according to any one of paragraphs 1 to 2,

[0301] Wherein said training the machine learning algorithm to establish one or more corresponding relationships between the changes in the measured fluid demand during the first time period and the measured environmental conditions during the first time period comprises:

[0302] modeling the variation of the measured fluid demand during the first time period using a set of parameters including network parameters that are constant for the FDN and environmental parameters that depend on prevailing environmental conditions of the FDN,

[0303] The machine learning algorithm is trained to establish one or more corresponding relationships between the value of each of the environmental parameters obtained from the modeling and the measured environmental conditions during the first time period, wherein

[0304] Using a trained machine learning algorithm to predict the change in fluid demand during the second time period includes using a trained machine learning algorithm to predict the change in fluid demand during the second time period based on the predicted environmental conditions and one or more correspondences established between the values ​​of each of the environmental parameters obtained from the modeling and the measured environmental conditions of the FDN during the first time period.

[0305] Paragraph 4. The method of any one of paragraphs 1 to 3, wherein said determining said change in fluid pressure to be applied at said plurality of pressure control stations during said second time period comprises:

[0306] creating a model for simulating fluid pressures at the one or more predetermined points in the FDN during a second time period for a given set of fluid pressures at the plurality of pressure control stations during the second time period based on the predicted change in fluid demand during the second time period, and

[0307] Numerical techniques are used to estimate changes in fluid pressure to be applied at the plurality of pressure control stations during the second period of time that will satisfy the predetermined pressure conditions at the one or more predetermined points in the FDN.

[0308] Paragraph 5. A method according to any of paragraphs 1 to 4, wherein the pressure condition is to minimize the sum of one or more areas between the curves of each simulated fluid pressure at the one or more predetermined points in the FDN and the minimum allowable fluid pressure in the FDN.

[0309] Paragraph 6. A method according to any one of paragraphs 1 to 5, comprising:

[0310] The change in the measured fluid demand during the first time period is determined based on a change in measured fluid pressure at the plurality of pressure control stations during the first time period and a change in measured fluid pressure at the one or more predetermined points in the FDN during the first time period.

[0311] Paragraph 7. Wherein, determining a change in the measured fluid demand during the first time period based on a change in the measured fluid pressure at the plurality of pressure control stations and a change in the measured fluid pressure at the one or more predetermined points in the FDN during the first time period comprises:

[0312] receiving from the plurality of pressure control stations measured fluid pressures at the plurality of pressure control stations during the first time period; and

[0313] Measured fluid pressures at the one or more predetermined points during the first time period are received from one or more devices located at the one or more predetermined points in the FDN.

[0314] Paragraph 8. A method according to any one of paragraphs 1 to 7, wherein said training the machine learning algorithm to establish one or more corresponding relationships between the changes in the measured fluid demand during the first time period and the measured environmental conditions during the first time period comprises:

[0315] Environmental conditions measured during the first time period are received.

[0316] Paragraph 9. A method according to any one of paragraphs 1 to 8, comprising:

[0317] A predicted environmental condition of the FDN during the second time period is received.

[0318] Paragraph 10. The method of any of paragraphs 1 to 9, wherein receiving the predicted environmental condition of the FDN during the second time period comprises:

[0319] The predicted environmental conditions are received as a weather forecast for the second time period.

[0320] Paragraph 11. A method according to any one of paragraphs 1 to 10, comprising:

[0321] receiving, after the second time period, from the plurality of pressure control stations measured fluid pressures at the plurality of pressure control stations during the second time period,

[0322] receiving, after the second time period, from one or more devices located at the one or more predetermined points in the FDN, fluid pressures measured at the one or more predetermined points in the FDN during the second time period,

[0323] receiving, after the second time period, the measured environmental conditions during the second time period,

[0324] determining a change in measured fluid demand during the second time period based on the measured fluid pressures at the plurality of pressure control stations during the second time period and the measured fluid pressures at the one or more predetermined points in the FDN during the second time period,

[0325] The machine learning algorithm is retrained to establish one or more corresponding relationships between the measured environmental conditions during the second time period and changes in the measured fluid demand during the second time period.

[0326] Paragraph 12. The method according to paragraph 11, comprising:

[0327] The measured environmental conditions received during the second time period, the fluid pressures received at the plurality of pressure control stations during the second time period, and the fluid pressures received at the one or more predetermined points during the second time period are stored.

[0328] Paragraph 13. The method of any of paragraphs 1 to 12, wherein the one or more predetermined points in the FDN are fluid pressure low points.

[0329] Paragraph 14. The method of any one of paragraphs 1 to 13, wherein the second time period is a 24 hour period.

[0330] Paragraph 15. A method according to any one of paragraphs 1 to 14, wherein the fluid is a gas.

[0331] The described embodiments may be implemented in any suitable form, including hardware, software, firmware, or any combination of these. The described embodiments may optionally be implemented at least in part as computer software running on one or more data processors and / or digital signal processors. The elements and components of any embodiment may be implemented physically, functionally, and logically in any suitable manner. In fact, the functions may be implemented in a single unit, in multiple units, or as part of other functional units. In this way, the disclosed embodiments may be implemented in a single unit, or may be physically and functionally distributed between different units, circuits, and / or processors.

[0332] Although the present disclosure has been described in conjunction with some embodiments, it is not intended to be limited to the specific form set forth herein. In addition, although it may appear that features have been described in conjunction with specific embodiments, those skilled in the art will recognize that the various features of the described embodiments may be combined in any manner suitable for implementing the technology.

Claims

1. A method performed by a computing device for controlling a fluid pressure in a fluid distribution network (FDN) above a threshold pressure by communicating with a plurality of pressure control stations distributed throughout the FDN to independently control a fluid pressure at each of the plurality of pressure control stations, the method comprising: training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period, using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, determining a change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the plurality of pressure control stations based on the change in fluid demand predicted during the second time period, and An indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations during the second time period is sent to the plurality of pressure control stations to satisfy the fluid pressure condition at the one or more predetermined points in the FDN.

2. The method according to claim 1, wherein: The instructions sent by the computing device include pressure control settings for the plurality of pressure control stations to control fluid pressure based on the determined changes in fluid pressure to be applied at the plurality of pressure control stations during the second time period.

3. The method of claim 1 , wherein said training the machine learning algorithm to establish one or more corresponding relationships between the changes in the measured fluid demand during the first time period and the measured environmental conditions during the first time period comprises: modeling the variation of the measured fluid demand during the first time period using a set of parameters, the set of parameters comprising network parameters that are constant for the FDN and environmental parameters that depend on prevailing environmental conditions of the FDN, training the machine learning algorithm to establish one or more corresponding relationships between the value of each of the environmental parameters obtained from the modeling and the measured environmental conditions during the first time period, wherein Using a trained machine learning algorithm to predict the change in fluid demand during the second time period includes: using a trained machine learning algorithm to predict the change in fluid demand during the second time period based on the predicted environmental conditions and one or more corresponding relationships established between the values ​​of each of the environmental parameters obtained from the modeling and the measured environmental conditions of the FDN during the first time period.

4. The method according to claim 1, wherein: Determining the change in fluid pressure to be applied at the plurality of pressure control stations during the second time period includes: creating a model for simulating fluid pressures at the one or more predetermined points in the FDN during a second time period for a given set of fluid pressures at the plurality of pressure control stations during the second time period based on the predicted change in fluid demand during the second time period, and Numerical techniques are used to estimate changes in fluid pressure to be applied at the plurality of pressure control stations during the second period of time that will satisfy the predetermined pressure conditions at the one or more predetermined points in the FDN.

5. The method according to claim 1, wherein: The pressure condition is to minimize a sum of one or more areas between a curve of each simulated fluid pressure at the one or more predetermined points in the FDN and a minimum allowable fluid pressure in the FDN.

6. The method according to claim 1, comprising: A change in the measured fluid demand during the first time period is determined based on a change in measured fluid pressure at the plurality of pressure control stations during the first time period and a change in measured fluid pressure at the one or more predetermined points in the FDN during the first time period.

7. The method according to claim 6, wherein: Determining a change in the measured fluid demand over the first time period based on a change in the measured fluid pressures at the plurality of pressure control stations and a change in the measured fluid pressures at the one or more predetermined points in the FDN over the first time period includes: receiving from the plurality of pressure control stations measured fluid pressures at the plurality of pressure control stations during the first time period; and Measured fluid pressures at the one or more predetermined points during the first time period are received from one or more devices located at the one or more predetermined points in the FDN.

8. The method of claim 1, wherein said training said machine learning algorithm to establish one or more corresponding relationships between said measured changes in fluid demand during said first time period and said measured environmental conditions during said first time period comprises: Environmental conditions measured during the first time period are received.

9. The method according to claim 1, comprising: A predicted environmental condition of the FDN during the second time period is received.

10. The method of claim 1, wherein the receiving a predicted environmental condition of the FDN during the second time period comprises: The predicted environmental conditions are received as a weather forecast for the second time period.

11. The method according to claim 1, comprising: receiving, after the second time period, from the plurality of pressure control stations measured fluid pressures at the plurality of pressure control stations during the second time period, receiving, after the second time period, from one or more devices located at the one or more predetermined points in the FDN, fluid pressures measured at the one or more predetermined points in the FDN during the second time period, receiving, after the second time period, the measured environmental conditions during the second time period, determining a change in measured fluid demand during the second time period based on the measured fluid pressures at the plurality of pressure control stations during the second time period and the measured fluid pressures at the one or more predetermined points in the FDN during the second time period, The machine learning algorithm is retrained to establish one or more corresponding relationships between the measured environmental conditions during the second time period and changes in the measured fluid demand during the second time period.

12. The method according to claim 11, comprising: The measured environmental conditions received during the second time period, the fluid pressures received at the plurality of pressure control stations during the second time period, and the fluid pressures received at the one or more predetermined points during the second time period are stored.

13. The method according to claim 1, wherein: The one or more predetermined points in the FDN are points of low fluid pressure. The method of claim 1 , wherein the second time period is a 24-hour period.

15. The method according to claim 1, wherein: The fluid is a gas.

16. A computing device for controlling fluid pressure in a fluid distribution network (FDN) above a threshold pressure by communicating with a plurality of pressure control stations distributed throughout the FDN to independently control fluid pressure at each of the plurality of pressure control stations, the computing device comprising: a transceiver circuit configured to transmit and / or receive signals, and A controller circuit, the controller circuit in combination with the transceiver circuit being configured to: training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period, using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on the predicted environmental conditions in a second time period and one or more established corresponding relationships between the change in the measured fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, determining a change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the plurality of pressure control stations based on the change in fluid demand predicted during the second time period, and An indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations during the second time period is sent to the plurality of pressure control stations to satisfy the fluid pressure condition at the one or more predetermined points in the FDN.

17. The computing device of claim 16, wherein: The instructions sent by the computing device include pressure control settings for the plurality of pressure control stations to control fluid pressure based on the determined changes in fluid pressure to be applied at the plurality of pressure control stations during the second time period.

18. The computing device of claim 16, wherein: The controller circuit in combination with the transceiver circuit is configured to: modeling the variation of the measured fluid demand during the first time period using a set of parameters, the set of parameters comprising network parameters that are constant for the FDN and environmental parameters that depend on prevailing environmental conditions of the FDN, training the machine learning algorithm to establish one or more corresponding relationships between the value of each of the environmental parameters obtained from the modeling and the measured environmental conditions during the first time period, A trained machine learning algorithm is used to predict changes in fluid demand during the second time period based on the predicted environmental conditions and one or more correspondences established between the values ​​of each of the environmental parameters obtained from the modeling and the measured environmental conditions of the FDN during the first time period.

19. A method for controlling fluid pressure in a fluid distribution network (FDN) performed by a pressure control station, the method comprising: receiving from a computing device an indication of a determined change in fluid pressure to be applied at the pressure control station within a future time period to satisfy fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station; and Fluid pressure at the pressure control station is adjusted based on the indication received from the computing device.

20. A pressure control station for controlling fluid pressure in a fluid distribution network FDN, the pressure control station comprising: a pressure control assembly for regulating the fluid pressure at the pressure control station, a transceiver circuit configured to transmit and / or receive signals, and A controller circuit, the controller circuit in combination with the transceiver circuit being configured to: receiving from a computing device an indication of a determined change in fluid pressure to be applied at the pressure control station within a future time period to satisfy fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station, wherein the pressure control assembly is configured to: Fluid pressure at the pressure control station is adjusted based on the indication received from the computing device.

21. A method for controlling fluid pressure in a fluid distribution network (FDN) above a threshold pressure by independently controlling fluid pressure at each of a plurality of pressure control stations distributed throughout the FDN, the method comprising: training a machine learning algorithm, via a computing device, to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period, using the trained machine learning algorithm by the computing device to predict a change in the fluid demand in the FDN in a later second time period based on the predicted environmental conditions in a second time period and one or more corresponding relationships established between the change in the measured fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, determining, by the computing device, a change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to satisfy fluid pressure conditions at one or more predetermined points in the FDN downstream of the plurality of pressure control stations based on the change in fluid demand predicted during the second time period, transmitting, by the computing device, to the plurality of pressure control stations an indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to satisfy the fluid pressure conditions at the one or more predetermined points in the FDN, receiving, via the plurality of pressure control stations, an indication of a change in the fluid pressure determined to be applied at the plurality of pressure control stations over a second time period to satisfy the fluid pressure condition at the one or more predetermined points in the FDN, and The fluid pressure at the respective pressure control stations is adjusted, by the plurality of pressure control stations, based on the indication received from the computing device.

22. A system for controlling fluid pressure in a fluid distribution network (FDN) above a threshold pressure by independently controlling fluid pressure at each of a plurality of pressure control stations distributed throughout the FDN, the system comprising: the plurality of pressure control stations, and A computing device, the computing device being configured to: training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period, using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, determining a change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the plurality of pressure control stations based on the change in fluid demand predicted during the second time period, and sending an indication of the determined change in fluid pressure to be applied at the plurality of pressure control stations during the second time period to the plurality of pressure control stations, wherein the plurality of pressure control stations are configured to: receiving an indication of a change in the determined fluid pressure to be applied at the plurality of pressure control stations during the second time period to satisfy the fluid pressure condition at the one or more predetermined points in the FDN, and Fluid pressure at the corresponding pressure control station is adjusted based on the indication received from the computing device.

23. A method for controlling fluid pressure in a fluid distribution network (FDN) performed by a pressure control station, the method comprising: training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period, using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, determining a change in fluid pressure to be applied at the pressure control station during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station based on the predicted change in fluid demand during the second time period, and The fluid pressure at the pressure control station is adjusted based on the determined change in fluid pressure to be applied at the pressure control station during the second time period to meet the fluid pressure conditions at the one or more predetermined points in the FDN.

24. A pressure control station for controlling fluid pressure in a fluid distribution network (FDN), the pressure control station comprising: a pressure control assembly for adjusting the fluid pressure at the pressure control station, a transceiver circuit configured to transmit and / or receive signals, and A controller circuit, the controller circuit in combination with the transceiver circuit being configured to: training a machine learning algorithm to establish one or more corresponding relationships between changes in measured fluid demand in the FDN over a first time period and measured environmental conditions over the first time period, using a trained machine learning algorithm to predict a change in fluid demand in the FDN in a later second time period based on one or more established corresponding relationships between the predicted environmental conditions in a second time period and the measured change in fluid demand in the FDN in the first time period and the measured environmental conditions of the FDN in the first time period, determining a change in fluid pressure to be applied at the pressure control station during the second time period to meet fluid pressure conditions at one or more predetermined points in the FDN downstream of the pressure control station based on the change in fluid demand predicted during the second time period, wherein the pressure control assembly is configured to: The fluid pressure at the pressure control station is adjusted based on the determined change in fluid pressure to be applied at the pressure control station during the second time period to meet the fluid pressure conditions at the one or more predetermined points in the FDN.

25. A computer program comprising instructions which, when said program is run by a computer, cause said computer to carry out the method according to claim 1, 19, 21 or 23.

Images