Patent Application
20240272595
The present teachings and arrangements relate generally to novel systems and methods for management of a fluid infrastructure network. More particularly, the present teachings and arrangements relate to using a digital twin for enabling management by monitoring flow attributes within the fluid infrastructure network.
Fluid infrastructure network includes conveyance, pumps, diversions, and storage tanks and wells, treatment units, for delivering or processing fluid or water. Examples of fluid networks include sewage system, groundwater well system, aquafer, aqueduct network and rain harvesting system. These networks are designed to collect and deliver fluid streams and sometimes treat the delivered fluid streams. By way of example, infrastructure networks dedicated to collecting and delivering treated water for drinking and agricultural and industrial use must be properly managed as they are integral to human life. Unfortunately, regardless of whether fluid treatment is necessary or not, management of these networks is time consuming, arduous, and financially burdensome.
What is, therefore, required are novel systems and processes for management of fluid infrastructure network systems.
To achieve the foregoing, the present arrangements and teachings offer novel systems and methods, using a digital twin, for enabling management by monitoring flow attributes within a fluid infrastructure network.
In one aspect, the present teachings provide fluid infrastructure network management methods. One such exemplar management method includes: (1) obtaining a copy of a digital twin model at a server; (2) receiving, for an instance in time or for a period of time, a forecast of an environmental condition around an area surrounding a sensor; (3) estimating at the server, using the forecast of the environmental condition and the copy of the digital twin model, flow attribute inside the fluid infrastructure item for the instance in time or for the period of time to produce one or more server's estimated flow attribute values; (4) receiving or obtaining, from the sensor inside a remote telemetry unit, the sensor measurement of the flow attribute inside the fluid infrastructure item; and (5) reporting, after lapse of the instance in time or after lapse of the period of time, one or more of the server's estimated flow attribute values as being or being an estimate of the sensor measurement for the instance in time or for the period of time, if the server does not receive from the remote telemetry unit the sensor measurement. In one preferred embodiment of the present teachings, the above-mentioned reporting element is carried out, if the remote telemetry unit determines that an absolute value of a difference between the sensor measurement and one of the remote telemetry unit's estimated flow attribute values is equal to or less than a predefined threshold tolerance value, or determines that the sensor measurement, for the instance in time or for the period of time, falls within a predefined confidence interval of the remote telemetry unit's multiple estimated flow attribute values of the fluid infrastructure item.
In another aspect, the present teachings provide another type of fluid infrastructure network management methods. In this aspect, the remote telemetry unit and a sensor are remotely located to the server, the remote telemetry unit is communicatively coupled to a sensor that measures flow attribute inside the fluid infrastructure item such that the remote telemetry unit receives from the sensor a sensor measurement.
Further, an exemplar method of this aspect includes: (1) obtaining, inside each of the server and inside the remote telemetry unit, a digital twin model providing, for an instance in time or for a period of time, one or more estimates of a flow attribute inside a fluid infrastructure item; (2) receiving, for the instance in time or for the period of time, a forecast of an environmental condition around an area surrounding the sensor; (3) estimating, at the remote telemetry unit and using the forecast of the environmental condition and one of the digital twin models, flow attribute inside the fluid infrastructure item for the instance in time or for the period of time to produce one or more remote telemetry unit's estimated flow attribute values; (4) estimating, at the server and using the forecast of the environmental condition and one of the digital twin models, flow attribute inside the fluid infrastructure item for the instance in time or for the period of time to produce one or more server's estimated flow attribute values; (5) obtaining, using the sensor and at the remote telemetry unit, the sensor measurement of the flow attribute inside the fluid infrastructure item; (6) comparing, at the remote telemetry unit, the sensor measurement with one or more of the remote telemetry unit's estimated flow attribute values; (7) reporting, for the instance in time or for the period of time, one or more of the server's estimated flow attribute values as being or being an estimate of the sensor measurement, if an absolute value of a difference between the sensor measurement and one of the remote telemetry unit's estimated flow attribute values is equal to or less than a predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls within a predefined confidence interval of the remote telemetry unit's multiple estimated flow attribute values.
If, however, the absolute value of the difference between the sensor measurement and one of the remote telemetry unit's estimated flow attribute values is greater than the predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside the predefined confidence interval of the remote telemetry unit's multiple estimated flow attribute values, then the above-mentioned management methods of the present teachings, preferably, further comprise conveying, from the remote telemetry unit to the server, the sensor measurement for the instance in time or for the period of time.
Next the method of this aspect, preferably, includes an element of receiving, after lapse of the instance in time or of the period of time and at the server, information regarding a realized environmental condition that is realized in the area surrounding the sensor for the instance in time or during the period of time.
Based upon the information regarding the realized environmental condition and the digital twin model present at the server, the management methods of the present teachings determine one or more server's new estimated flow attribute values for the instance in time or for the period of time. This allows for comparing, at the server, the sensor measurement with one or more of the server's new estimated flow attribute values.
The management methods of the present teachings then proceed to an element of maintaining the digital twin model at the server, if the absolute value of the difference between the sensor measurement and one or more of the server's new estimated flow attribute values is equal to or greater than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside a check predefined confidence interval of multiple of the server's new estimated flow attribute values.
If, however, the absolute value of the difference between the sensor measurement and one of the server's new estimated flow attribute is less than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls inside the check predefined confidence interval of multiple of the server's new estimated flow attribute values, then the methods of the present teachings, preferably, proceed to an element of obtaining inside each of the server and the remote telemetry unit, a new digital twin model that is different from the digital twin model. In other words, the original digital twin model is updated, recalibrated, or reconstructed.
In yet another aspect, the present teachings provide yet another type of fluid infrastructure network management methods. In this aspect, the remote telemetry unit is communicatively coupled to a sensor, which measures a flow attribute inside the fluid infrastructure item such that the remote telemetry unit receives from the sensor a sensor measurement of the flow attribute. One such exemplar method of this aspect includes: (1) obtaining, inside a remote telemetry unit, a digital twin model providing, for an instance in time or for a period of time, one or more estimates of the flow attribute inside the fluid infrastructure item; (2) receiving a forecast of an amount of precipitation that is expected to be received in an area surrounding the sensor at the instance in time or during the period of time; (3) estimating at the remote telemetry unit, for the instance in time or for the period of time, based upon the forecast of the amount of precipitation and the digital twin model, one or more remote telemetry unit's estimated flow attribute values for the fluid infrastructure item; (4) receiving, at the remote telemetry unit, the sensor measurement for the instance in time or for the period of time and receiving an amount of realized precipitation that is realized in an area surrounding the sensor for the instance in time or during the period of time; (5) comparing the sensor measurement with one or more of the remote telemetry unit's estimated flow attribute values; (6) determining at the remote telemetry unit, using the digital twin and the amount of realized precipitation, one or more new remote telemetry unit's estimated flow attribute values for the instance in time or for the period of time, if the absolute value of a difference between the sensor measurement and one of the remote telemetry unit's estimated flow attribute values is greater than a predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside a predefined confidence interval of the remote telemetry unit's multiple estimated flow attribute values; (6) comparing, at the remote telemetry unit, the sensor measurement with one or more of the new remote telemetry unit's estimated flow attribute values; and (7) maintaining the digital twin model at the server, if the absolute value of the difference between the sensor measurement and one of the new remote telemetry unit's estimated flow attribute values is equal to or greater than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside a check predefined confidence interval of multiple of the new remote telemetry unit's estimated flow attribute values.
In yet another aspect, the present arrangements provide fluid infrastructure network management systems. One such exemplar management system comprises: (1) multiple sensors, each of which is designed to measure a flow attribute inside a fluid infrastructure item that is part of a fluid infrastructure network; (2) multiple remote telemetry units, each of which is communicatively coupled to at least one of the sensors and comprises: (i) a battery for providing power, (ii) a sensor module for communicatively coupling to the sensor and receiving the sensor measurement, (iii) a memory for storing sensor measurements and one of a digital twin model that provides, for an instance in time or for a period of time, one or more remote telemetry unit's estimated flow attribute values for the infrastructure item, (iv) a communications module connected to a network and designed to transmit information from the remote telemetry unit to the network, and (v) a microprocessor programmed to implement instructions comprising: (a) receiving, for the instance in time or for the period of time and using the communication module, a forecast of an environmental condition around an area surrounding the sensor; (b) estimating, for the instance in time or for the period of time, and using the forecast of the environmental condition and the digital twin model, one or more remote telemetry unit's estimated flow attribute values inside the fluid infrastructure item; (c) obtaining, using the sensor, a sensor measurement inside the fluid infrastructure item; (d) comparing the sensor measurement with one or more of the remote telemetry unit's estimated flow attribute values; (e) conveying, from the remote telemetry unit to a network, the sensor measurement for the instance in time or for the period of time, if the absolute value of the difference between the sensor measurement and one of the remote telemetry unit's estimated flow attribute values is greater than a predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside a predefined confidence interval of the remote telemetry unit's multiple estimated flow attribute values.
The exemplar management system also comprises a server communicatively coupled, through the network, to multiple of the remote telemetry units and having stored therein multiple digital twin models, each of which is a copy of a digital twin model stored in each of the remote telemetry units, and wherein each of the remote telemetry units and each of the sensors are remotely located to the server, wherein the server is programmed to implement instructions comprising: (a) receiving from the remote telemetry unit through the network, the sensor measurement for the instance in time or for the period of time; (b) receiving, after lapse of the instance in time or after lapse of the period of time and at the server, information regarding a realized environmental condition that was realized in the area surrounding the sensor for the instance in time or during the period of time; (c) determining, based upon the information regarding the realized environmental condition and the digital twin model present at the server, one or more server's new estimated flow attribute values inside the fluid infrastructure item for the instance in time or for the period of time; (d) comparing, at the server, the sensor measurement with one or more of the server's new estimated flow attribute values; (e) maintaining the digital twin model at the server, if an absolute value of the difference between the sensor measurement and one of the server's new estimated flow attribute values is equal to or greater than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside a check predefined confidence interval of multiple of the server's new estimated flow attribute values.
In yet another aspect, the present arrangements provide another type of fluid infrastructure network management systems. One such exemplar system comprises: (1) multiple sensors, each of which is designed to generate a sensor measurement inside a fluid infrastructure item; (2) multiple remote telemetry units, each of which is communicatively coupled to at least one of the sensors and has stored therein one of a digital twin model; (3) a server communicatively coupled to multiple of the remote telemetry units and has stored therein multiple digital twin models such that one digital twin model on the server is substantially similar to the digital twin model stored on the remote telemetry units; (4) a network through which the multiple remote telemetry units are communicatively coupled to the server that includes a microprocessor being programmed with instructions comprising: (a) receiving, for the instance in time or for the period of time, a forecast of an environmental condition around an area surrounding at least one of the sensors; (b) estimating, at the server and using the forecast of the environmental condition and one of the digital twin models, flow attribute inside the fluid infrastructure item for the instance in time or for the period of time to produce one or more server's estimated flow attribute values; (c) receiving or obtaining, from the sensor and at the remote telemetry unit, the sensor measurement of the flow attribute inside the fluid infrastructure item; and (d) reporting, after lapse of the instance in time or after lapse of the period of time, one or more of the server's estimated flow attribute values as being or being an estimate of the sensor measurement for the instance in time or for the period of time, if the server does not receive from the remote telemetry unit the sensor measurement.
The systems and methods of operation and effective compositions obtained from the present teachings and arrangements, however, together with additional objects and advantages thereof, will be best understood from the following descriptions of specific embodiments when read in connection with the accompanying figures.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present teachings and arrangements. It will be apparent, however, to one skilled in the art that the present teachings and arrangements may be practiced without limitation to some or all these specific details. In other instances, well-known process steps have not been described in detail in order to not unnecessarily obscure the present teachings and arrangements.
Effective monitoring systems and methods for management of fluid infrastructure network are crucial for implementing real time decision support systems. To this end, the present arrangements and present teachings realize that while advances in sensing, battery technology and cellular communication have made the installation and operation of such monitoring systems more cost effective and robust, little has been done to extend their battery life. Particularly when transmission of real time sensor measurements is deemed paramount, conventional monitoring systems suffer from constantly expending energy to collect and transmit real-time sensor data that undesirably requires replacing batteries that power communication modules such as cellular modems.
The present arrangements and present teachings solve this “constantly-replacing-batteries” problem by, in preferred embodiments, using a synchronized machine-learning model, which is alternately referred to as a digital twin model. In one configuration, one of the digital twin models operates on a remote telemetry unit (“RTU”) and another of the digital twin models operates on a server, which is communicatively coupled to and aggregates data from multiple RTUs distributed within a fluid infrastructure network. In this configuration, at least one RTU is connected to a sensor and the RTUs and the sensors are disposed at monitoring locations of fluid infrastructure network that are remote to the server. As a result, the server may be thought of as a central processor that aggregates remotely collected data and digital twins for management of the fluid network. Representative fluid infrastructure item. according to the present arrangements and present teachings, includes conduit, reservoir, orifice, weir, pump and wastewater treatment plant.
According to the present teachings, the digital twin model is used to produce estimates of the sensor measurements at an RTU and the server. In one embodiment of the present teachings, the RTU compares an actual sensor measurement with an estimate obtained from the digital twin model (operating at the RTU). If the actual sensor measurement and the estimate from the digital twin model are within a tolerable range, then the actual sensor measurement is not conveyed from the RTU (where it resides) to the server. In this circumstance, the server relies upon the digital twin model stored thereon and assumes that an estimate of sensor measurement obtained from the digital twin model (at the server) to be the actual sensor measurement. Obviating the need to transmit the actual sensor measurement from the RTU to the server represents an energy saving features of the present teachings that is not realized by conventional monitoring systems and methods. Furthermore, the energy saving features allows the present arrangements and present teachings to effectively monitor fluid infrastructure networks for several years without replacing the batteries powering the communication module.
Further still, the server's ability, by relying on the digital twin model, to provide an acceptable estimate of the actual sensor measurement allows the present arrangements and present systems to offer cost effective real-time decision support to effectively manage the fluid infrastructure network.
It is important that the digital twin models, present at the RTU and at the server, are synchronized to realize these advantages (e.g., energy savings and/or real-time responsive measures) offered by the digital twin models. To this end, if the actual sensor measurement and the estimate from the digital twin model are not within a tolerable range, then the actual sensor measurement is conveyed from the RTU to the server and a process to reconcile is initiated. At the server, information regarding a realized environmental condition (e.g., amount of precipitation) is received and is used to obtain an estimate of the sensor measurement from the digital twin model. If the difference between the actual sensor measurement and the estimate from the digital twin model, at the server, is outside a tolerable range, then the realized environmental condition is deemed an outlier. In this situation, the present teachings consider the digital twin models, at the RTU and at the server, to be synchronized and therefore, retains the digital twin model at the server.
If, however, the difference between the actual sensor measurement and the estimate of the sensor measurement, obtained from the digital twin model at the server, is within the tolerable range, then the digital twin models, at the RTU and at the server, are considered by the present teachings not to be synchronized. Accordingly, the digital twin model at the server is updated or reconstructed so that it is synchronized with the digital twin model present at the RTU.
The present teachings not only recognize the need for energy savings and real time response measures, but also provides structural provisions and process features for synchronizing the underlying digital twin models, at the RTU and at the server and that enable the advantages of energy savings and real time response measures so that those advantages continue and are not disrupted. These structural provisions and process features are described in detail below.
RTU 102 includes an RTU microprocessor 128 that is communicatively coupled to an RTU communication module 134, which is in turn communicatively coupled to network 110. Inside RTU 102, an RTU memory module is bi-directionally coupled to RTU microprocessor 128. A sensor module 136 receives data from a sensor 112 that is configured to measure a flow attribute of a remote fluid infrastructure item, which is located remotely with respect to server 104.
According to
In an alternate embodiment of the present fluid infrastructure monitoring system, sensor module 136 is directly coupled to RTU communication module 134 and a sensor measurement, taken at sensor 112, does not have to go through RTU microprocessor 128 to arrive at RTU communication module 134. In this alternate arrangement, the sensor measurement, taken at sensor 112, is conveyed from sensor module 136 directly to RTU communication module 134. Regardless of whether RTU microprocessor 128 is involved in communicating a sensor measurement, taken at sensor 112, the sensor measurement is conveyed from RTU communication module 134 through a network to server 104.
Server 104 includes a server communications module 138 that is bi-directionally coupled to a server processor 140, which is in turn bi-directionally coupled to a database 142. Server communications module 138, like RTU communication module 134, is communicatively coupled to network 110. In connection with a sensor measurement, taken at sensor 112, server communications module 138 receives the sensor measurement from network 110 and forwards it to server processor 140 that uses it in a comparison, for example, explained below.
Although
In one embodiment of the present arrangements, inside server 204, machine learning engine 206 uses sensor parameter information of the fluid infrastructure network, sensor measurements of inputs into and outputs from the fluid infrastructure item of interest, i.e., fluid infrastructure item for which the digital twin model is being constructed, and a probability distribution function of environmental information surrounding the sensor (e.g., sensor 212 of
Representative sensor measurements, in connection with a fluid inside a fluid infrastructure item, include flowrate data, flow level data, hydrostatic pressure data, flow velocity data, concentration data for one or more chemical species of interest, pH, turbidity, total suspended solids, residual chlorine, total chlorine, total organic carbon, dissolved oxygen, ammonia, nitrates, and phosphorous. Sensor measurements taken using a sensor are conveyed using an associated RTU therewith (e.g., RTU 214A, RTU 214B and RTU 214C if the fluid infrastructure items associated with each of these RTUs provide an input or input stream into, or an output or an output stream exiting RTU 202), through network 210, to server 204.
In connection with inputs into a fluid infrastructure item of interest (e.g., infrastructure item which has flow attributes measured by sensor 212), the input information used to construct digital twin model 208 includes at least one information chosen from a group comprising precipitation information surrounding the sensor (e.g., sensor 212 measurements of which are estimated by digital twin model 20), fluid level of one or more different inputs into a fluid infrastructure item and flow rate of one or more different the inputs into a fluid infrastructure item. By way of example, if RTU 214A is associated with a conduit, which serves as a fluid infrastructure item and conveys fluid from a storage tank, another fluid infrastructure item monitored by RTU 214B, to RTU 202, then a fluid flowrate sensor measurement of the conduit is conveyed from RTU 214A, through network 210, to server 204 and/or a fluid level sensor measurement of the storage tank is conveyed from RTU 214B, through network 210, to server 204. Further, these sensor measurements of the inputs into RTU 202 are conveyed through a network to a server for construction of digital twin model 208, one of which also ultimately resides in RTU 202.
As another example, server 204 receives precipitation data 216 from National Weather Service that forecasts the amount and time of precipitation at an area surrounding the location of a remote sensor (e.g., sensor 212) for the construction, using machine learning engine 206, of a digital twin model (e.g., digital twin model 208) of an RTU (e.g., RTU 202). In this manner, digital twin models of each RTU (e.g., RTU 202, RTUs 214A, 214B, or 214C) are constructed and preferably stored inside a server (e.g., server processor 140 of
Similarly, with respect to outputs exiting fluid infrastructure item of interest (e.g., infrastructure item which has flow attributes measured by sensor 212), the output information used to construct digital twin model 208 includes at least one information chosen from a group comprising evaporation information surrounding the sensor (e.g., sensor 212 measurements of which are estimated by digital twin model 20), fluid level of one or more different outputs exiting a fluid infrastructure item and flow rate of one or more different outputs exiting a fluid infrastructure item. Fluid flowrate and/or fluid level measurements of outputs or output streams exiting fluid infrastructure item of interest (being monitored by RTU 202) are conveyed from their respective RTUs, through network, to the server. Such sensor measurements of the outputs exiting RTU 202 are used in the construction of digital twin model 208.
Examples of sensor parameters include location of the sensor, (e.g., sensor 212), type of sensor, critical high/low transmission rates of sensor and/or of RTU (e.g., RTU 202, RTUs 214A, 214B, or 214C), battery level of sensor. Some or all these parameters are obtained from a sensor, an RTU or are input at a server.
In one embodiment of the present teachings, the digital twin model (e.g., digital twin model 208 of
In this context,
The present teachings recognize that in this manner, digital twin models for other infrastructure items, in the fluid infrastructure network, that are monitored by other RTUs, e.g., RTU 314A, RTU 314B and RTU 314C, are created either, preferably, in server 304 or in RTU 302 and then a copy of the digital twin model is conveyed to the other, i.e., either, preferably, to RTU 302 or server 304. Regardless of the approach of where an initial digital twin model is created, a server, e.g., server 304, has stored thereon a copy of a digital twin model of a digital twin model stored inside each of the RTUs that monitor different fluid infrastructure items in a fluid infrastructure network. More importantly, the present teachings recognize that a server, preferably, incorporates an infrastructure connection between RTUs (e.g., RTU 314B monitors a tank, which is located downstream from RTU 314A that monitors a conduit that is connected to a fluid infrastructure item that is monitored by RTU 302), to automatically determine—“a system-wide digital twin model.” To this end, the present teachings recognize also that such a “system wide digital twin model” may allow a few scheduled transmissions of actual sensor measurements from one or more RTUs to the server and gain a more precise estimate of the sensor measurements being monitored by one or more of the RTUs, without dramatically increasing the energy requirements associated with transmissions of sensor measurements from the RTU to the server. It is not necessary to create a “system wide digital twin model,” and the present teachings recognize that by simply constructing, in the server, different digital twin models, a copy of each also being stored on the RTU, the present arrangements offer the advantages of RTUs, without relying on the monitoring activities of the RTUs.
In the embodiment shown in
If the sensor measurement falls within a confidence interval of one or more RTU's estimated flow attribute values (e.g., that ranges in value from about ±10% of the mean of the probability distribution function to about ±15% of the mean of the probability distribution function), then the present teachings recognize that, as an energy savings measure, an RTU need not convey the actual sensor measurement to the server.
Server 404 also receives, through network 410, a forecast of precipitation data 416 (e.g., time and amount of precipitation which may be in the form a probability distribution function). Server 404 uses a server's digital twin model 408 to use precipitation data 416 and generate a probability distribution function, estimating flow attribute values for an infrastructure item. In preferred embodiments of the present arrangements, digital twin model 408 instructs a machine learning engine 406 to use precipitation data 416 and generate the probability distribution function that estimates flow attribute values. Regardless of the way in which the flow attribute sensor measurement is estimated, server 404 assumes the sensor measurement to be a mean value of the probability distribution function estimating flow attribute values and reports, at a server's user interface, the mean value of the server's probability distribution function as being the sensor measurement or being an estimate of the sensor measurement.
Such reporting may be periodic, e.g., every 5 minutes, and need not be carried out after the lapse of future instance in time, which forms the basis of the estimate, e.g., the probability distribution function, of the sensor measurement. In one preferred embodiment of the present teachings, after a lapse of future instance in time or after a lapse of future period in time and if server 404 does not receive from RTU 402 a sensor measurement, then server reports to a user the mean value of the server's probability distribution function as being the sensor measurement or being a reliable estimate of the sensor measurement.
Further, even if the actual sensor measurement falls within a confidence interval, RTU 402 may convey, through network 410, the actual sensor measurement to server 404. Under such conditions and for conserving energy, RTU 402 may so convey relatively infrequently, e.g., once a day or once a week, the actual sensor measurements. In certain embodiments of the present teachings, when server 408 does not receive, at predefined time, a sensor measurement from RTU 402, then at or after the predefined time, server 408 relies upon precipitation data to calculate server's estimated flow attribute values e.g., a probability distribution function, and reports a mean value of these estimated flow attribute values as being the sensor measurement or being a reliable estimate of the sensor measurement.
The present teachings also recognize the importance of a possibility that the sensor measurement falls outside the confidence interval of the RTU's one or more estimated flow attribute values (e.g., outside of the ranges in value from about ±10% of the mean of the probability distribution function to about ±15% of the mean of the probability distribution function), then the present teachings recognize that an additional condition must be tested, i.e., whether the actual amount of precipitation realized is an outlier or within the expected amount of precipitation (as discerned from precipitation data 516).
In the embodiment shown in
Precipitation data 516, e.g., forecasts of environmental conditions in areas surrounding one or more sensors such as sensor 512, is conveyed, preferably through network 510, to both RTU 502 and server 504. In this configuration, RTU 502, using machine learning engine 520 and forecasts of environmental conditions obtained from precipitation data 516, determines RTU's new estimated flow attribute values inside a fluid infrastructure item for an instance in time or for a period of time.
If the absolute value of a difference between the sensor measurement obtained from sensor 512 and sensor module 522 and the RTU 502's new estimated flow attribute values is greater than a predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside the predefined confidence interval of the RTU 502's multiple estimated flow attribute values, then the underlying sensor measurement used for computing the estimate is conveyed from RTU 502, through network 510, to server 504. At server 502, precipitation data 516, e.g., the actual amount of precipitation realized around areas surrounding one or more sensors, such as sensor 512, is used to compute one or more server 504's new estimated flow attribute values (e.g., see curve in graphical representation 526). This estimate is of a flow attribute, for an instance in time or for a period of time, inside a fluid infrastructure item being monitored by sensor 512. Server 504 compares, as shown in graphical representation 526, the sensor measurement taken from sensor 512 with one or more server 504's new estimated flow attribute values (e.g., a probability distribution function that includes multiple estimated values).
If the absolute value of a difference between the sensor measurement and one of the server's new estimated flow attribute values is equal to or greater than a check predefined threshold tolerance value, or if the sensor measurement, for an instance in time or for a period of time, falls outside the check predefined confidence interval of multiple server's new estimated flow attribute values, then the digital twin model 508 is maintained at server 504. In this scenario, the present teachings recognize that the actual precipitation realized in the area surrounding sensor 512 is an outlier relative to the forecasted amount of precipitation in that area. As a result, the present teachings do not deem digital twin model to be an inaccurate at estimating the flow attribute inside a remotely located fluid infrastructure item, rather understand that the forecast did not accurately predict the amount of precipitation.
If, however, the absolute value of the difference between the sensor measurement and one of the server's new estimated flow attribute values is equal to or less than the check predefined threshold tolerance value, or if the sensor measurement, for an instance in time or for a period of time, falls within the check predefined confidence interval of multiple server's new estimated flow attribute values, then digital twin model 508 is updated, recalibrated or reconstructed at server 504. A copy of this updated, recalibrated or reconstructed digital model is conveyed to RTU 502 so that the “twin” nature of the digital model is reestablished. The digit twin model is updated, recalibrated or reconstructed because the present teachings recognize that the previous digital twin models' estimated flow attribute values are unreliable, or otherwise unacceptable. As a result, the present teachings propose updating, recalibrating, or reconstructing, as may be needed, so that the new digital twin models' estimated flow attribute values are reliable and acceptable.
The present teachings also offer novel methods of infrastructure management. Although implementation of these novel methods do not require a specific system configuration, but novel methods implemented using the structural details described in
Element 606 may also include estimating, at the server, for the instance in time or for the period of time, using the forecast of environmental condition and the digital twin models, to generate one or more server's estimated flow attribute values for the fluid infrastructure item. In another embodiment, another element, which is different from element 606, in method 600 is implemented to arrive at one or more server's estimated flow attribute values for the fluid infrastructure item. As a result, generating these two estimates, i.e., the RTU's and the server's estimated flow attribute values of the fluid infrastructure item, may be thought of as two separate elements or as a single element of method 600.
Regardless of whether one or two elements are implemented to obtain these estimated flow attribute values, method 600 proceeds to an element of 608, which involves obtaining, using the sensor (e.g., sensor 112 of
Preferably, after element 610 has concluded, an element of 612 is carried out. According to the embodiment shown in
In another preferred embodiment of the present teachings, element 612 involves, after a lapse of the instance in time or a lapse of the period of time, reporting the server's estimated flow attribute being provided in place of the sensor measurement of the fluid infrastructure item, if a sensor measurement of a flow attribute inside the fluid infrastructure item is not received at the server. In one implementation of this embodiment, the RTU does not convey the sensor measurement of the fluid infrastructure item to the server, if the RTU determines that the absolute value of the difference between the sensor measurement and the RTU's estimated flow attribute values is equal to or less than a predefined threshold tolerance value, or determines that the sensor measurement, for the instance in time or for the period of time, falls within a predefined confidence interval of the RTU's multiple estimated flow attribute values of the fluid infrastructure item.
Prior to the implementation of elements 610 and/or 612, method 600, more preferably, does not include conveying the sensor measurement of the flow attribute inside the fluid infrastructure item, from the RTU to the server.
Accordingly, element 702 involves conveying, from the RTU to the server, the sensor measurement for the instance in time or for the period of time, if the absolute value of the difference between the sensor measurement and one of the RTU's estimated flow attribute values is greater than the predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside the predefined confidence interval of the RTU's multiple estimated flow attribute values. By way of example,
Method 700 includes an element 704 of receiving at the server, after lapse of the instance in time or the period of time, information regarding a realized environmental condition, i.e., realized around an area surrounding the fluid infrastructure item and the sensor for the instance in time or during the period of time. However, in one embodiment of method 700, element 704 of the present teachings may be carried out prior to the lapse of instance in time or prior to the period of time.
Next, method 700 involves carrying out an element 706 that involves determining, based upon the information regarding the realized environmental condition (obtained from element 704) and the digital twin model present at the server, a server's new estimated flow attribute values inside the fluid infrastructure item for the instance in time or for the period of time.
Then, method 700 includes an element 708 of comparing, at the server, the sensor measurement (obtained from element 702) with one of the server's new estimated flow attribute values (obtained from element 706).
Method 700 then advances to an element 710, which involves maintaining the digital twin model at the server, if an absolute value of the difference between the sensor measurement and one of the server's new estimated flow attribute values is equal to or greater than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside a check predefined confidence interval of multiple of the server's new estimated flow attribute values. By way of example,
The sensor parameter information includes at least one item chosen from a group comprising location of sensor, type of sensor, critical high flow attribute that is detectable by the sensor, critical low flow attribute that is detectable by the sensor, transmission rate of sensor data and battery level of the sensor.
The input information includes at least one information chosen from a group comprising precipitation information in an area surrounding the sensor associated with the RTU, fluid level of one or more different inputs into the fluid infrastructure item and flow rate of one or more different inputs into the fluid infrastructure item.
The output information includes at least one information chosen from a group comprising evaporation information around the sensor associated with the RTU, fluid level of one or more different outputs exiting the fluid infrastructure item and flow rate of one or more different outputs exiting the fluid infrastructure item.
In addition to fluid level and/or fluid flowrate, flow attribute of input information and output information may also include at least one attribute chosen from a group comprising hydraulic head, groundwater flux, rate of groundwater flux, pH, conductivity, concentration of total organic compounds, amount of total suspended solids and turbidity.
Building, as contemplated under element 602, preferably, includes generating, using a machine learning module, a probability distribution function of the estimated sensor measurements of the flow attribute inside the fluid infrastructure item.
Regardless of the manner in which the digital twin model is obtained, element 602 may include modifying, using the RTU and the sensor (e.g., sensor measurements provide insight into the reliability of the digital twin model's estimate) the digital twin model until it reliably estimates the flow attribute inside the fluid infrastructure item. By way of example, modifying of the digital twin model includes revising values of one or more parameters, associated with the fluid infrastructure item that is used to build an initial version of the digital twin model, until one or more of the estimates of the flow attribute obtained from the digital twin model are equal or close to the sensor measurement (e.g., sensor measurements obtained from sensor 212 of
In one preferred embodiment of the present teachings, modifying of the digital twin model includes determining that a Nash-Sutcliff coefficient of the digital twin model is greater than a predefined acceptable value, or determining that a root mean squared deviation, between the estimates of the flow attribute and sensor measurements of the flow attribute obtained for different instances in time or for different periods of time, is less than or equal to a predefined error value.
The present teachings recognize that building and modifying, if these elements are implemented, preferably, provides at least one of the digital twin models (e.g., digital twin model 208 of
According to certain preferred embodiments of the present teachings, once an acceptable digital twin model is present at the server, method 600 includes an element of conveying or transmitting a copy of the digital twin model from the server to the remote telemetry unit (e.g., server 304 conveys a copy of digital twin model 308, through network 310, to RTU 302 such that the copy is stored as digital twin model 318 therein as shown in
After an acceptable digital twin model is obtained at the server, method 600 of
In one aspect of element 604 shown in
According to preferred implementations of the present teachings, when implementing element 606 of method 600 shown in
Continuing with this example and in connection with element 608 shown in
Element 612, preferably, includes reporting, for the instance in time or for the period of time, a mean value of the server's probability distribution function as being the sensor measurement or being an estimate of the flow attribute inside the fluid infrastructure item, if the sensor measurement for the instance in time or for the period of time, falls within a predefined confidence interval of a mean value of the RTU's probability distribution function. In one embodiment, the present teachings recognize that the confidence interval spans a predefined range of estimated flow attribute values, which deviates from a mean value of the RTU's probability distribution function.
In one preferred implementation of method 600, element 604 includes receiving, at a periodic rate, the forecast of the environmental condition around an area surrounding the sensor. In one implementation of this embodiment, element 606 of the present teachings includes estimating, at the periodic rate or less than the periodic rate, and at the RTU flow attribute inside the fluid infrastructure item to generate one or more RTU's estimated flow attribute values. Continuing with this embodiment, element 608 of the present teachings includes obtaining, at a sampling rate of the sensor, the sensor measurement. In this embodiment, the sampling rate is greater than the periodic rate.
Further, if the absolute value of the difference between the sensor measurement and one of the RTU's estimated flow attribute values is greater than the predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls outside the predefined confidence interval of the RTU's multiple estimated flow attribute values, then the management method of the present teachings, preferably, further includes an element of conveying (e.g., element 702), at least at or less than the sampling rate, from the RTU to the server, the sensor measurement taken for the instance in time or for the period of time.
This preferred embodiment further includes an element of receiving (e.g., substantially similar to element 704 of
Then, at the server, management methods of the present teachings implement an element of comparing (e.g., substantially similar to element 708 of
In preferred embodiments, management methods of the present teachings include obtaining, inside each of the server and inside the RTU, a new digital twin model different from the (original) digital twin model, if the absolute value of the difference between the sensor measurement and one of the server's new estimated flow attribute values is less than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls inside the check predefined confidence interval of multiple of the server's new estimated flow attribute values.
In connection with method 700, the present teachings offer additional different embodiments for effective management of a fluid network. In one such embodiment, method 700 includes obtaining, inside the RTU, a new digital twin model, if the absolute value of the difference between the sensor measurement and one of the remote telemetry unit's new estimated flow attribute values is less than a check predefined threshold tolerance value, or if the sensor measurement, for the instance in time or for the period of time, falls inside the check predefined confidence interval of multiple of the remote telemetry unit's new estimated flow attribute values.
Continuing with this embodiment, method 700, preferably, further includes an element of conveying the new digital twin model, formed at the RTU, to the server. In this embodiment, the element of conveying is carried out using a communicative coupling between the RTU and the server. In this configuration, the server is communicatively coupled to multiple RTUs and has stored thereon one copy of the digital twin model that is also stored on one of the RTUs, Further, each RTU is coupled to a sensor that measures a particular type of flow attribute inside a fluid infrastructure item. The digital twin at the server allows the server to use, after lapse of the instance in time or the period of time, information regarding the amount of realized precipitation to arrive at the server's new estimates flow attribute values and operate, in certain instances, independently of the RTU, thereby conserving energy.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present teachings and arrangements. It will be apparent, however, to one skilled in the art that the present teachings and arrangements may be practiced without limitation to some or all of these specific details. In other instances, well-known process steps have not been described in detail in order to not unnecessarily obscure the present teachings and arrangements.
This patent application claims priority to U.S. provisional patent application No. 63/254,175 filed on Oct. 11, 2021, which is incorporated herein by reference for all purposes.
| Number | Date | Country | |
|---|---|---|---|
| 63254175 | Oct 2021 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 18686444 | Jan 0001 | US |
| Child | 18621142 | US |
