SE2150085A1 - System and method for predictive maintenance for a District Heating Substation - Google Patents

Abstract

Method for predictive maintenance for a district heating, DH, substation using a prediction system, having a plurality of sensor units configured to collect measurements in the DH substation and being in communication with at least one central unit, wherein the DH substation comprise a heat exchanger configured to transfer heating energy from a DH network to a user heating system. The method comprises measuring, at each sensor unit of the plurality of sensor units in the DH substation, at least one parameter; transmitting the measured at least one parameter to at the at least one central unit; and determining at least one malfunction prediction at the DH substation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

Description

System and method for predictive maintenance for a District Heating Substation.

TECHNICAL FIELD The present invention generally relates to the field of District Heating, DH, and more particularlyto a system and a method for predictive maintenance for at least one DH substation by analyzingdata captured by a plurality of sensor units in the DH substation, being in communication with at least one central unit.

BACKG ROUN D ART Many countries, including Sweden, has a long tradition for using district heating in urban areas.District heating is a system for distributing heat generated in a centralized location through asystem of insulated pipes for residential and commercial heating requirements such as space heating and water heating.

District Heating systems mainly comprise of two primary components, a heat generation facilityand a DH substation, these are connected by a distribution network or DH network. The heatgeneration facility may be centralized and in connection with a plurality of DH substations via theDH network. The heat may for example be generated in any type of boiler, solar energy,geothermal source, electricity, and/or thermal energy developed as a by-product of electricalgeneration. The DH network may be a network of insulated pipes that conveys the energygenerated at the heat generation facility to a DH substation at an end users' locations, i.e. residential and commercial spaces.

An ideal DH substation may comprise hundreds of components including heat exchangers,radiators, valves and circulation pumps among others. Due to extreme pressure andtemperature, the components of the DH substation may be under a lot of stress, which may lead to undesired operation or faults in the components. The typical faults and common malfunctions e.g. internal leakage, excessive heat, expansion of pipes, etc., are difficult and costly to detect according to a service technician's experience.

Given the huge number of components present in the DH substation and the current systemarchitecture, it is difficult to detect where a fault actually occurs. Moreover, to identify wherethe fault has occurred may involve a number of steps including dismantling various parts of thesubstation, which can demand a lot of downtime and also be exhaustive to the technicians. lf thecondition and operation ofthe components are not checked on time or when required, the entiresystem may malfunction, which will attract a big amount of investment in repair or replacement of the components. Therefore, timely maintenance of a DH substation is of utmost importance.

Currently, service technicians pay regular visits to customer su bstations for maintenance. Servicetechnicians may even receive calls from customers requiring inspection due to malfunctions.With the current DH substation and their implementations, it is difficult for service techniciansto get access to information regarding historical/current performance of substations. Access tosuch information can not only efficiently facilitate the service technicians to perform themaintenance work but also would enable conditional monitoring of substations and benefit maintenance service which leads to increased customer satisfaction.

Thus, the inventor of the present invention has identified a need for improvement inmaintenance of DH substations, which delivers additional values than traditional maintenance methods.

SUMMARY OF THE INVENTION An object of the present invention is to provide a system and a method for predictivemaintenance for a district heating, DH, substation which overcomes or at least mitigate the drawbacks identified above.

A further object of the present invention is to provide a system with advantages relating to the reliability of the prediction.

Aspects of the present invention are defined in to the independent claims. Preferred embodiments are set forth in the dependent claims.

According to one aspect of the invention a method for predictive maintenance for a districtheating, DH, substation using a prediction system is provided. The control system has at least aplurality of sensor units configured to co||ect measurements in the DH substation and being incommunication with at least one central unit, wherein the DH substation comprises a heatexchanger configured to transfer heating energy from a DH network to a user heating system.

The method comprises the steps of: Measuring, at each sensor unit of the plurality of sensor units in the DH substation, at least one pa rametef.

Transmitting, from each sensor unit of the plurality of sensor units, the measured at least one pa rametef.

Receiving, at the at least one central unit, the measured at least one parameterfrom each sensor unit of the plurality of sensor units.

Determining, at the at least one central unit, at least one malfunction prediction at the DHsubstation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

BRIEF DESCRIPTION OF THE DRAWINGS The invention is described in the following illustrative and non-limiting detailed description of exemplary embodiments, with reference to the appended drawings, wherein: Figure 1 is a schematic illustration of a basic system for predictive maintenance according to an aspect of the present disclosu re; Figure 2 is a schematic illustration ofa method for predictive maintenance according to an aspect of the present disclosure; Figure 3 is a schematic illustration of a system for predictive maintenance according to an aspect of the present disclosure; Figure 4a is a schematic illustration of a system for predictive maintenance according to an aspect of the present disclosure; Figure 4b is a schematic illustration of an alternative system for predictive maintenance according to an aspect of the present disclosure; Figure 4c is a schematic illustration of yet another system for predictive maintenance according to an aspect of the present disclosure; and Figure 5 is a schematic illustration of a central unit for predictive maintenance according to an aspect of the present disclosure.

All figures are schematic, not necessarily to scale, and generally only show parts that arenecessary in order to elucidate the invention, wherein other parts may be omitted or merelysuggested. Throughout the figures the same reference signs designate the same, or essentially the same features.

DETAILED DESCRIPTION The present invention can be used for predictive maintenance for a DH substation whichovercome or at least mitigate the problems of the prior art and with an improved functionality in a more robust system that provides reliability benefits.

A DH substation consists of a multitude of components, and each component may generate afault causing malfunctions. A typical DH substation used for heating purposes, such as domestichot water heating and/or space heating, may comprise one or more heat exchangers, one ormore radiators, one or more control valves, one or more circulation pumps and one or more security valves among others. Additionally, the typical DH substation may comprise one or more filters and one or more expansion vessels. lt is hard to detect where the fault actually occursgiven the large number of components and the complex design of the substation. To identifywhere the fault has occurred may involve a number of steps including opening and/ordismantling various parts of the substation, which can be cumbersome and may also demand alot of downtime. Moreover, once the system malfunctions, the entire substation may have to berepaired or at least one or more components where the fault has occurred may have to be replaced. This involves significant cost and downtime. lt is therefore desired that such malfunction may be prevented before it could damage the DHsubstation, and which could save cost of repair or replacement to a considerable extent. This canbe achieved by enabling a prediction of faults or anomalies by analyzing data that indicatesperformance of the DH substation. The prediction may be an indication that maintenance isrequired for the DH substation or that maintenance is required for certain components of the DH substation.

Current systems may for example have water meters and/or flow meters implemented, thesecan be used to detect leakage or flow of water at various locations in the DH substation. Watermeters can measure the usage of water at different times of the day, month or year and flowmeters may measure the water flow volume and pressure at various locations in the pipes of theDH substation. However, currently used meters do not have the technical capability or possibilityto record and analyze data that indicates performance of the DH substation. Thus, prediction ofany anomaly or fault in the DH system may not be achieved with current water meters and flow metefS.

For a system to predict faults or anomalies, the system may require a set of readings that can becompared against a set of historical data indicating the desired pressure, usage and temperatureof water at various locations in the pipes. lf the data patterns do not match then it may beinferred that a fault exists in the network ofthe DH system. Similarly, if the data patterns match a failure pattern, it may be inferred that a fault exists in the network of the DH system.

One possible way to improve could be to implement various sensors and advanced meters in the DH substation, which incorporate methods that may identify a fault. However, many meters are not able to modify their existing operation. Thereby, a modification of a meters software mightbe a complex solution that might require significant investment. Further, the sensors could onlyprovide a set of current data/readings and the issue still remains as to what historical data thesensor data would be compared against, how and where an analysis may be performed toprovide a prediction and where will such historical data be generated and stored. Therefore,without required data and a platform for analysis and comparison, predictive maintenance is impossible to perform.

The existing drawbacks associated with fault determination and maintenance may be reduced ifa service technician or the main DH service center is provided with information regarding historical/current performance of one or more DH substations.

From this, it has been realized that any possible improvement should be done at a central application.

The term predictive maintenance refers to the use of data-d riven, proactive maintenancemethods that are designed to analyze a condition of an equipment and assist in predicting whenmaintenance should be performed on the equipment. The invention is described in the followingillustrative and non-limiting detailed description of exemplary embodiments, with reference to the appended drawings, wherein: Figure 1 shows a schematic illustration of a system 100 for predictive maintenance for a DHsubstation using a prediction system, according to an aspect of the present disclosure. Thesystem 100 comprises one DH substation 120 and one central unit 160. The DH substation 120may be part of a centralized system where heat may be generated and distributed through anetwork of insulated pipes for residential and commercial heating requirements such as spaceheating and/or water heating. The illustrated DH substation 120 may communicate with thecentral unit 160 via a data communication link 125. The DH substation 120 as shown in Figure 1comprises a heat exchanger 131 configured to transfer heating energy from a DH network to auser heating system. The DH substation 120 further comprises a DH supply pipe 111 and a DHreturn pipe 112 connected to the DH network. The DH substation 120 further comprises a pipe 132 connected to the heat exchanger 131 for transport of heat energy in the user heating system.

The pipe 132 may be connected to one or more other pipes through which the pipes convey theenergy generated at a heat generation facility, transferred via the DH network, to the DH substation 120 and the end users' locations, i.e. residential and commercial spaces.

The heating energy from the DH network is transported to the DH substation 120 on a fluid carriervia the DH supply pipe 111. The fluid carrier transfers heating energy to the DH substation 120 atthe heat exchanger 131. The fluid carrier is returned to the DH network via the DH return pipe112.The heat exchanger 131 is placed between the DH supply pipe 111 and the DH return pipe112. The heat exchanger 131 is further connected to the pipe 132, which conveys the energyfrom the DH network to the user heating system. The pipe 132 may further be connected viacontrol valves and circulation pumps (shown in Figure 3) installed within for controlled distribution of heated water or heat to the user heating system.

The DH substation and the heat and water distribution pipes may collectively be referred to as the DH network in some embodiments.

The term user heating system may be used to refer to a system comprising a hot water system (DHW) and/or a space heating system (RAD) in some embodiments. lt may be noted that Figure 1 illustrates a basic DH substation architecture used for either waterheating or space heating. A DH substation may also comprise a multitude of components inaddition to those described in reference with Figure 1. A more elaborate DH substation architecture will be described with reference to Figure 3.

A DH substation, such as the DH substation 120 may further comprise a plurality of sensor units(shown in Figure 3) configured to collect measurements in the DH substation 120 and being incommunication with the central unit 160. Each sensor unit of the plurality of sensor units in theDH substation 120 may be configured to measure at least one parameter. The measured at leastone parameter is transmitted to the central unit 160 via the communication link 125. The centralunit 160 is configured to receive the measured at least one parameter from each sensor unit of the plurality of sensor units and determine at least one malfunction prediction at the DH substation 120 based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

The at least one measured parameter transmitted from the sensor units may be referred to assensor data. The measured parameters may be real time sensor data indicative ofthe operationof the various components in the DH substation 120. The sensor data from the sensor units aretransmitted to the central unit 160. By transmitting sensor data to the central unit 160, it isensured that appropriate real time data is made available for analysis of this data for predictivemaintenance. Such data may be classified as normal data or faulty data. This enablesidentification of whether data received from the DH substation are within normal operatingvalues and ranges or not. Such identification can be used to predict if there is any indication of failure to one or more components.

Prediction of faults in the components of the DH substations thereby enables conditionalmonitoring of the components and makes it possible and easier for the service technicians toperform maintenance operations efficiently when required without inviting consequences that involves significant cost and downtime.

The central unit 160 may be configured with computational capabilities to analyze and classifysensor data from the DH substation into various classes. Each class may be associated withdifferent maintenance requirements and actions to be performed. The analysis and classification enables prediction of malfunction and thereby determining maintenance requirements.

The sensor unit may communicate with the central unit via a data communication link 125. Thedata communication link 125 may for an example utilize one or a plurality of different types ofwired links or wireless links, such as for example xDSL, 2G, 3G, 4G, 5G TCP/IP, WiFi, Bluetooth,WiMax, WLL, PSTN, PLC, zigbee, optical fibre or a combination thereof.

Figure 2 shows a schematic illustration of a method 200 for predictive maintenance for a DHsubstation, such as the DH substation 120, using a prediction system according to an aspect ofthe present disclosure. The prediction system may have a plurality of sensor units configured to collect measurements in the DH substation 120. The DH substation 120 is in communication with the at least one central unit 160 via a data communication link. The DH substation 120 maycomprise a heat exchanger such as the heat exchanger 131 configured to transfer heating energyfrom a DH network to a user heating system. The method 200 comprises measuring 210, at eachsensor unit of the plurality of sensor units in the DH substation, at least one parameter. Themethod further comprises transmitting 220, from each sensor unit of the plurality of sensor units,the measured at least one parameter. The method further comprises receiving 230, at the atleast one central unit, the measured at least one parameter from each sensor unit of the pluralityof sensor units. The method further comprises determining 240, at the at least one central unit,at least one malfunction prediction at the DH substation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

The at least one measured parameter transmitted from the sensor units may be referred to assensor data. The measured parameters may be real time sensor data indicative ofthe operationof the various components in the DH substation 120. The real time data from the sensor unitsare transmitted to the central unit 160. By transmitting sensor data to the central unit 160, it isensured that appropriate real time data is made available for analysis of this data for predictivemaintenance. Such data may be classified as normal data or faulty data. This enablesidentification of whether data received from the DH substation are within normal operatingvalues and ranges or not. Such identification can be used to predict if there is any indication of failure to one or more components.

The at least one parameter that is being measured may primarily comprise temperature,pressure and flow of water, among others. These parameters may be sensed or measured usingany of the technologies widely known and used for sensing. For example, temperature sensorsmeasure the amount of heat energy or even coldness that is generated by an object or system,allowing one to ”sense” or detect any physical change to that temperature producing either ananalogue or digital output. A pressure sensor may measure the difference between pressure experienced at an input port and an output port of a tube or pipe.

Sensor data/measured parameters may be collected in analog form, which may be converted into digital form and transmitted over a data communication link to the central unit 160.

Sensor data/measured parameters may be encoded in one or more standard formats and transmitted over the data communication link to the central unit 160.

Sensor data/measured parameters from each sensor unit may be transmitted at differentintervals over the data communication link to the central unit 160 and the central unit 160 hasknowledge which parameters are transmitted at what time interval. By doing this, processing time for arranging and sorting data at the central unit may be reduced.

Sensor data/measured parameters from each sensor unit may be transmitted with time stamps indicating the time at which the data is recorded by the sensor unit.Sensor data/measured parameters may be transmitted over an interval.

Sensor data/measured parameters may be transmitted over an interval that may be between one second and one day.

Sensor data/measured parameters may be transmitted over an interval of one second, twoseconds, three seconds, four seconds, five seconds, six seconds, seven seconds, eight seconds, nine seconds, or ten seconds.

Sensor data/measured parameters may be transmitted over an interval of one minute, twominutes, three minutes, four minutes, five minutes, six minutes, seven minutes, eight minutes, nine minutes, or ten minutes.

Sensor data/measured parameters may be transmitted over an interval of one hour, two hours, three hours, four hours, five hours, six hours, seven hours, eight hours, nine hours, or ten hours.

Sensor data/measured parameters may be transmitted over an interval that may vary based on a determined transmission interval. The determined transmission interval may vary over time.Sensor data/measured parameters may be transmitted over an interval of 24 hours.

The data communication link as described earlier may for an example utilize one or a plurality ofdifferent types of wired links or wireless links, such as for example xDSL, 2G, 3G, 4G, 5G TCP/IP,WiFi, Bluetooth, WiMax, WLL, PSTN, PLC, zigbee, optical fibre or a combination thereof.

The central unit 160 may store received data, for example as comprised in the communication from the plurality of sensor units.

The sensor data received at the central unit 160 may be stored in a database of the central unit160. Additionally, the sensor data may be mapped to the associated components and themapping relation may be stored in the database. Further, the sensor data received at the central unit 160 may be sorted based on a timestamp and stored.

Alternatively, in another embodiment, the method 200 for predictive maintenance for a DHsubstation, such as the DH substation 120 may comprise one or more additional steps. Themethod 200 may comprise providing, a plurality of sensor units in the DH substation, whereinthe plurality of sensor units is configured to collect measurements in the DH substation and beingin communication with at least one central unit. The method 200 comprises measuring 210, ateach sensor unit of the plurality of sensor units in the DH substation, at least one parameter. Themethod further comprises transmitting 220, from each sensor unit of the plurality of sensor units,the measured at least one parameter. The method further comprises receiving 230, at the atleast one central unit, the measured at least one parameter from each sensor unit of the pluralityof sensor units. The method further comprises determining 240, at the at least one central unit,at least one malfunction prediction at the DH substation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

Figure 3 shows a schematic illustration of a system 300 for predictive maintenance for a DHsubstation 320 using a prediction system, according to an aspect of the present disclosure. Thesystem 300 comprises one DH substation 320 and one central unit 360. The DH substation 320may be a part of a centralized system where heat may be generated and distributed through anetwork of insulated pipes for residential and commercial heating requirements such as space heating and/or water heating.

The DH substation 320 comprises two heat exchangers connected in parallel between DH supplyand DH return pipes. One of the heat exchangers 331 may supply a district water heating, DHW,circuit (VVC) for water heating, and one of the heat exchangers 341 may supply a radiator circuit (RAD) for space heating. The DH substation 320 further comprises a DH supply pipe 311 and a DH return line pipe 312 connected to the DH network. The heat exchanger 331 for water heatingmay be referred to as a DHW heat exchanger 331 and the heat exchanger 341 for space heatingmay be referred to as a RAD heat exchanger 341. The DHW may further comprise a DHW controlvalve 314, a DHW circulation pump 333, a DHW security valve 334, and interconnecting pipes,such as pipe 332. The RAD may comprise a RAD control valve 315, a RAD circulation pump 345, aRAD security valve 344, a RAD expansion vessel 343, and interconnecting pipes, such as pipe 342.

The illustrated DH substation 320 further comprises a cold water supply.The heat exchangers 331, 341 may in some cases also be connected in parallel.

The DH substation 320 further comprises a plurality of sensor units 350 configured to collectmeasurements in the DH substation 320 and being in communication with the central unit 360.Each sensor unit of the plurality of sensor units 350 in the DH substation 320 may be configuredto measure at least one parameter. The measured at least one parameter is transmitted to thecentral unit 160 via a data communication link 325. The central unit 360 is configured to receivethe measured at least one parameter from each sensor unit of the plurality of sensor units 350and determine at least one malfunction prediction at the DH substation 320 based on the measured at least one parameter from each sensor unit of the plurality of sensor units 350.

For DHW i.e. domestic hot water, the DHW heat exchanger 331 may be connected the DH supplypipe 311 and the DH return pipe 312. A filter may be installed in the DH supply pipe 311, suchthat the water that may be supplied for heating is filtered off dust, chemicals and minerals. TheDHW heat exchanger 331 may further be connected to the DHW control valve or the DHWcirculation valve 324 through a pipe, which connects with the DH return pipe 312. A DHWcirculation pump 333 may be connected to the DHW security valve 334 that in turn may connectto the hot/cold water supply pipe 332. The cold and hot water may be mixed, heated up tosetpoint by the DHW heat exchanger 331, transported by the pipes, and (re)circulated by runninga pump when needed, when scheduled and/or continuously. The fixtu res, e.g. valves and faucets, may be used to control the flow of water in the system.

Similarly, for space heating the RAD heat exchanger 341 may be connected between the DH supply pipe 311 and the DH return pipe 312. The RAD heat exchanger 341 may further be connected to the RAD control valve 315 or the circulation valve 315 through a pipe, whichconnects with the DH return pipe 312. A RAD circulation pump 345 may be connected to the RADsecurity valve 344. For space heating system, radiators are the most common type of roomheaters, described as heat exchanger. The heat exchanger may be controlled by a control valve,which controls the amount of water that passes through the heat exchanger, and the water maybe circulated using a circulation pump. The expansion vessel 343 may be necessary to smoothout the pressure and volume variation in the water due to changing water temperature. The set point of water temperature mainly depends on outdoor climate.

The DHW circulation pump 333 may further comprise one pump, a pump arrangement, and/or a plurality of pumps. The plurality of pumps may be arranged in parallel and/or in series.

For DHW i.e. domestic hot water, the DHW heat exchanger may be connected between a filterinstalled in the DH supply pipe and the DH return pipe. The DHW heat exchanger may further beconnected to the DHW circulation valve through a pipe. The DHW circulation pump may beconnected to the DHW security valve that in turn may connect to the hot/cold water supply. Thecold and hot water may be mixed, heated up to setpoint by the DHW heat exchanger, transportedby the pipes, and (re)circulated by running a pump when needed, when scheduled and/orcontinuously. The fixtures, e.g. valves and faucets, may be used to control the flow of water in the system.

Similarly, for space heating the RAD heat exchanger may be connected between a filter in the DHsupply pipe and the DH return pipe. The RAD heat exchanger may further be connected to theRAD circulation valve through a pipe. The RAD circulation pump may be connected to the RADsecurity valve. For space heating system, radiators are the most common type of room heaters,described as heat exchanger. The heat exchanger may be controlled by a control valve, whichcontrols the amount of water that passes through the heat exchanger, and the water may becirculated using a circulation pump. The expansion vessel may be necessary to smooth out thepressure and volume variation in the water due to changing water temperature. The set point of water temperature mainly depends on outdoor climate. ln some embodiments, the RAD circulation pump 345 may further comprise one pump, a pumparrangement, and/or a plurality of pumps. The plurality of pumps may be arranged in parallel and/or in series.

The pipes 332, 342 connecting the components of the DH substation 320 may be referred to as the pipe network or pipeline network.

One or more sensors may be placed at various locations in the pipe network. The one or moresensors may comprise temperature sensors, pressure sensors, flow sensors and |eak detectors among other types of sensors.

A temperature sensor may be mounted on a pipe next to the DHW control valve 314 at a distanceranging from 1 cm to 10 cm from the DHW control valve 314. The sensor may also be mounted in connection with to the DHW control valve 314.

A temperature sensor may be mounted on a pipe next to the DHW control valve 314 at a distanceof less than 10 cm from the DHW control valve 314. The sensor may also be mounted in connection with to the DHW control valve 314.

A temperature sensor may be mounted on a pipe next to the RAD control valve 315 at a distanceranging from 1 cm to 10 cm from the RAD control valve 315. The sensor may also be mounted in connection with to the RAD control valve 315.

A temperature sensor may be mounted on a pipe next to the RAD control valve 315 at a distanceof less than 10 cm from the RAD control valve 315. The sensor may also be mounted in connection with to the RAD control valve 315.

A temperature sensor may be mounted on a pump casing or pipe next to the RAD circulation pump 345.

A temperature sensor may be mounted on the DHW circuit return temperature pipe at a distanceof less than 20 cm, or alternatively approximately 10 cm, after the DHW circulation pump 333.

The sensor may also be mounted in connection with to the DHW circulation pump 333.

A temperature sensor may be mounted on a pipe for the cold water supply.

A temperature sensor may be mounted on a pipe 311 of the DH supply between the DH supply and the heat exchanger 331.

A temperature sensor may be mounted on the pipe for cold water supply next to the DHWsecurity valve 334. The sensor may also be mounted at a distance of less than 10 cm from the DHW security valve 334.

A temperature sensor may be mounted on a radiator electrical motor surface.

A temperature sensor may be mounted on the radiator circuit supply temperature next to theRAD heat exchanger 341. The sensor may also be mounted at a distance of less than 10 cm from the RAD heat exchanger 341.

A pressure sensor may be installed at the expansion vessel 343.

A temperature sensor and a pressure sensor may be mounted at local manometer supply linenext to the filter 313. The sensor may also be mounted at a distance of less than 10 cm from the filter 313.

A temperature sensor and a pressure sensor and a flow sensor may be mounted at local manometer on the return pipe 312.

A temperature sensor and a flow sensor may be installed in the meter.

A leak detector may further be installed on the floor under the system valves. ln a specific non limiting example, three temperature sensors (TD1, TD2, TD3), three pressuresensors (PD1, PD2, PD3), a meter device (MD) and a leak device (LD) are installed at variouspositions in a DH substation 320 along the pipe network. These sensors provide at least sixteenmeasurements in total. Table 1 below illustrates the various sensors their positions and the meaSLlfementS.

Sensor Measurements Sensor PositionDHW contro| va|ve The temperature sensor is mounted on the pipe nextTD1_S1 temperature to control valves at about 1 to 10 cm from the DHWcontrol valveRadiator c0ntr0| va|ve The temperature sensor is mounted on the pipe nextTD1_S2 temperature to control valves at about 1 to 10 cm from the RADcontrol valveRadiator pump casing _ _TDl S3 temperature The temperature sensor |s mounted on pump casmg_ or pipe next to the RAD pumpDHW circuit return _temperature The temperature sensor |s mounted on the DHWTDZ-Sl circuit return temperature pipe approximately 10cm after the DHW pumpCold water supply _ _temperature The temperature sensor |s mounted on the p|pe forTD2_S2cold water supplyDHW circuit supply _TDZ S3 temperature The temperature sensor |s mounted on the DH supply_ pipe between the DH supply and the heat exchangerSecurity valve _ _temperature The temperature sensor |s mounted on the p|pe forTD3_S1cold water supply next to the security valveRadiator pump motor _ _temperature The temperature sensor |s mounted on Rad|atorTD3_S2 electrical motor surface Radiator circuit supply The temperature sensor is mounted on Radiator TD3 S3 temperature circuit supply temperature next to heat exchangerRadiator expansion vesselPD1 pressure Sensor at expansion vessel suitable connectionDH supply pressure on _. . Sensor at local manometer Supply l|ne next to thePDZ primary sidefilterDH return pressure onPD3 prlmary Slde Sensor at local manometer on return pipeDH supply flowMD_ST temperature Sensor installed in the meter in the DH supply pipeDH return flowMD_RT temperature Sensor installed in the meter in the DH return pipeDH return flow rate _ _ _ _Sensor |nstalled |n the meter |n the DH return p|peMD FR (accumulated)_ next to MD_RTLD Water leak detector On the floor Table 1. Sensor positions and measurements The Table 1 above is only exemplary and for the purpose of illustration and does not by any means limits the sensor positions and measurements to only those described in the table.

The sensors may be installed on, or in connection with, a body to be measured. The sensors mayalso be installed inside the body to be measured. A sensor may also be installed in an additional segment that may be operationally attached to the system.

The at least one sensor units may be chosen from: a thermal sensor arranged on the DHW controlvalve 314, a thermal sensor arranged on the RAD control valve 315, a thermal sensor arrangedon the RAD circulation pump 345, a thermal sensor arranged on a pipe 332 prior to the DHWcirculation pump 333, a thermal sensor arranged on the cold water supply, a thermal sensorarranged on a pipe 332 subsequent to the DHW heat exchanger 331, a thermal sensor arrangedon the DHW security valve 334, a thermal sensor arranged on the RAD pump motor, a thermalsensor arranged on a pipe prior to the RAD heat exchanger, a pressure sensor arranged on theRAD expansion vessel 350, a pressure sensor arranged on a manometer in the DH supply, athermal sensor arranged on the DH supply flow, a pressure sensor arranged on a manometer inthe DH return, a thermal sensor arranged on the DH return flow, and a flow sensor arranged on the DH return pipe 312.

The measured at least one parameter may comprise a temperature of a control valve 314, 315,a temperature of the DH return pipe 312, a temperature of the DH supply pipe 311, atemperature of a security valve 334, 344, a temperature of a circulation pump 333, 343 and aflow of the DH return pipe 112, among others. lt may be noted that the system 300 may be configured with capabilities to measure more parameters than those described above. ln one example, the plurality of sensor units 350 may comprise the thermal sensor arranged onthe DHW control valve 314, the thermal sensor arranged on the RAD control valve 315, thethermal sensor arranged on the pipe prior to the DHW circulation pump, the thermal sensorarranged on the cold water supply, the thermal sensor arranged on the DHW security valve, thethermal sensor arranged on the RAD pump motor, the pressure sensor arranged on the RADexpansion vessel, the thermal sensor arranged on the DH return flow, and the flow sensor arranged on the DH return.

The measured at least one parameter may for example comprise a temperature of the DHW control valve 314, a temperature of the RAD control valve 315, a temperature of the DHW circuit return pipe 312, a temperature of the cold water supply, a temperature of the security valve 334,a temperature of the RAD pump 345, a pressure of the RAD expansion vessel 343, a temperature ofthe DH return flow, and a flow of the DH return.

Determining the at least one malfunction prediction may in this example comprise determining leakage. ln one example, the plurality of sensor units 350 may comprise a thermal sensor arranged on theDHW control valve 314, a thermal sensor arranged on the RAD control valve 315, a thermalsensor arranged on a pipe prior to the DHW circulation pump 333, a thermal sensor arranged onthe cold water supply, a thermal sensor arranged on the pipe 332 subsequent to the DHW heatexchanger 331, a thermal sensor arranged on the RAD pump motor, a thermal sensor arranged on the DH return flow, and a flow sensor arranged on the DH return.

The plurality of parameters may in this example comprise a temperature of the DHW controlvalve 314, a temperature of the RAD control valve 315, a temperature of the DHW circuit return,a temperature of the cold water supply, a temperature of the DHW circuit supply, a temperature ofthe RAD pump 345, a temperature ofthe DH return flow, and a flow of the DH return.

Determining the at least one malfunction prediction may in this example comprise determining a malfunctioning of valves and DHW system. ln one example, the plurality of sensor units may comprise a thermal sensor arranged on theDHW control valve 314, a thermal sensor arranged on the pipe 332 prior to the DHW circulation pump 333, and a thermal sensor arranged on the RAD circulation pump 345.

The plurality of parameters may in this example comprise a temperature of the DHW control valve 314, a temperature of the DHW circuit return, and temperature of the RAD pump 345.

Determining the at least one malfunction prediction may in this example comprise determining malfunctioning pumps. ln one example, the plurality of sensor units 350 may comprise a thermal sensor arranged on the RAD control valve 315, a thermal sensor arranged on the pipe 342 prior to the RAD heat exchanger 341, a pressure sensor arranged on the RAD expansion vessel 343, a pressure sensorarranged on a manometer in the DH supply 311, a pressure sensor arranged on a manometer inthe DH return 312, a thermal sensor arranged on the DH supply flow, a thermal sensor arranged on the DH return flow, and a flow sensor arranged on the DH return.

The plurality of parameters may in this example comprise a temperature of the RAD control valve315, a temperature of the RAD circuit supply 311, a pressure ofthe RAD expansion vessel 343, adifferential pressure based on a pressure of the DH supply 311 and a pressure of the DH return312, a temperature of the DH supply flow, a temperature of the DH return flow, and a flow of the DH return.

Determining the at least one malfunction prediction may in this example comprise determining low filter performance. ln one example, the plurality of sensor units 350 may comprise a thermal sensor arranged on theRAD control valve 315, a thermal sensor arranged on the RAD circulation pump 345, a thermalsensor arranged on a pipe prior to the RAD heat exchanger 341, and a pressure sensor arranged on the RAD expansion vessel 343.

The plurality of parameters may in this example comprise a temperature of the RAD control valve315, a temperature of the RAD pump motor, a temperature of the RAD circuit supply, and a pressure of the RAD expansion vessel 343.

Determining the at least one malfunction prediction may in this example comprise determining a malfunctioning heat exchanger and/or pumps.

The at least one of the at least one sensor unit of the plurality of sensor units may be a surfacemounted sensor. Alternatively, the at least one of the at least one sensor unit ofthe plurality ofsensor units may be embedded within the body of the respective component where the respective parameter is measured.

The at least one parameter may be chosen from: temperature, pressure, and flowrate.

Additionally, at least one of the at least one sensor unit of the plurality of sensor units may be an infrared sensor, a gas sensor, a smoke sensor and a humidity sensor, among others.

The malfunction prediction may be chosen from: leakage, low filter performance, valves and DHW system, fouling of heat exchanger, malfunctioning pumps, and/or a combination thereof.

Figure 4a shows a schematic illustration of an example system 400 for predictive maintenancefor a DH substation 420 using a prediction system, according to an aspect of the presentdisclosure. The system 400 comprises at least one DH substation 420 at least one first centralunit 461 and at least one second central unit 462. The DH substation 420 is in communicationwith the at least one first central unit 461 via a data communication link 425. The at least onefirst central unit 461 is in communication with the at least on second central unit 462 via a datacommunication link 465. The DH substation 420 may be part of a centralized system where heatmay be generated and distributed through a network of insulated pipes for residential and commercial heating requirements such as space heating and/or water heating.

The DH substation 420 further comprise a plurality of sensor units configured to collectmeasurements in the DH substation 420 and being in communication with at least the firstcentral unit 461. Each sensor unit of the plurality of sensor units in the DH substation 420 maybe configured to measure at least one parameter. The measured at least one parameter istransmitted to at least the first central unit 461. The first central unit 461 is configu red to receivethe measured at least one parameter from each sensor unit of the plurality of sensor units anddetermine at least one malfunction prediction at the DH substation 420 based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

The at least one measured parameter transmitted from the sensor units may be referred to assensor data. The at least one measured parameter transmitted from the sensor units may bereferred to as sensor data. The measured parameters may be real time sensor data indicative ofthe operation of the various components in the DH substation 420. The real time data from thesensor units are transmitted to the central unit 461. By transmitting sensor data to the centralunit 461, it is ensured that appropriate real time data is made available for analysis of this data for predictive maintenance. Such data may be classified as normal data or faulty data. This enables identification of whether data received from the DH substation are within normaloperating values and ranges or not. Such identification can be used to predict if there is an indication of failure to one or more components.

Prediction of faults in the components of the DH su bstations thereby makes it possible and easierfor the service tech nicia ns to perform maintenance operations efficiently when required without inviting consequences that involves significant cost and downtime.

The first central unit 461 may be configured with computational capabilities to analyze andclassify sensor data from the DH substation 420 into various classes. Each class may be associatedwith different maintenance requirements and actions to be performed. The analysis andclassification enables prediction of malfunction and thereby determining maintenance requirements.

The first central unit 461 may comprise one or more memory and one or more processing units.The memory unit may comprise instructions to be executed by the processing units to performdetermining at least one malfunction prediction at the DH substation 420 based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

The first central unit 461 may be configured to receive the sensor data at frequent intervals fromthe plurality of sensor units. The sensor data comprises real time data corresponding to each ofthe parameters. The first central unit 461 may comprise a preconfigured database containinghistorical data that may indicate normal operation ofthe DH subsystem. The first central unit 461may further comprise a preconfigured database containing historical data that may indicate faulty operation of the DH subsystem.

The first central unit 461 may be connected to a display unit (not shown). The first central unit461 may display the determined malfunction prediction at the display unit. Additionally, thedisplay unit may also display a maintenance requirement or maintenance type or an actionrequired to be performed depending on the determined malfunction prediction. The display unit may be a centralized display unit.

The first central unit 461 may further generate a notification based on the determinedmalfunction prediction and can be displayed on the centralized display unit. The operators at thedistrict heating office or quarters may monitor the display unit and then may notify or inform theservice technicians to check at least one component or related components of the DH substationfor maintenance requirements. Alternatively or additionally, the notification may be sent to adevice of a service technician. Alternatively or additionally, the notification may be sent to a device of a customer.

The second central unit 462 may be in communication with the first central unit 461 via the data communication link 465.

The first central unit 461 may transmita determined malfunction prediction to the second centralunit 462. The first central unit 461 may transmit the determined malfunction prediction to thesecond central unit 462 with an instruction to display the determined malfunction prediction ata centralized display unit in communication with the first central unit 461 and the second centralunit 462. The second central unit 462 may further generate a notification based on thedetermined malfunction prediction and can be displayed on the centralized display unit. Theoperators at the district heating office or quarters may monitor the display unit and then maynotify or inform the service technicians to check at least one component or related componentsof the DH substation for maintenance requirements. Alternatively or additionally, the notificationmay be sent to a device of a service technician. Alternatively or additionally, the notification may be sent to a device of a customer.

The processing load for computation and determination of malfunction prediction may bedistributed efficiently between the first central unit 461 and the second central unit 462. Suchdistribution of processing load may enable avoiding network congestions and delivering fastercomputations. Other processing may also be distributed or shared, such as for example filtering,determinations, analysis, learning algorithms, historic data, and/or a combination thereof. The first central unit 461 may be a cloud server.

The first central unit 461 may be a physical server.

The first central unit 461 may be a physical server installed within a remote location.

The second central unit 462 may be a cloud server.

The second central unit 462 may be a physical server installed within a remote location.

The first central unit 461 may be the central unit that the sensor units are in communication with.

The second central unit 462 may be the central unit that the sensor units are in communicationwith. ln this case, the first central unit 461 may sort and arrange sensor data in a predefined orderand transmit the data to the second central unit 462 for computation and analysis. The secondcentral unit 462 may receive instructions from the first central unit 461 on performing the computation and determining the malfunction predictions.

The sensor unit may communicate with the first central unit 461 via a data communication link.The data communication link may for an example utilize one or a plurality of different types ofwired links or wireless links, such as for example xDSL, 2G, 3G, 4G, 5G TCP/IP, WiFi, Bluetooth,WiMax, WLL, PSTN, PLC, zigbee, optical fibre or a combination thereof.

The sensor unit may communicate with a proxy unit via a data communication link. The proxyunit may in turn communicate with the central unit. The data communication link may for anexample utilize one or a plurality of different types of wired links or wireless links, such as forexample xDSL, 2G, 3G, 4G, 5G TCP/IP, WiFi, Bluetooth, WiMax, WLL, PSTN, PLC, zigbee, optical fibre or a combination thereof.

The sensor unit 462 may communicate with a first central unit 461 via a data communication link.The first central unit 461 may in turn communicate with the second central unit 462. The datacommunication link may for an example utilize one or a plurality of different types of wired linksor wireless links, such as for example xDSL, 2G, 3G, 4G, 5G TCP/IP, WiFi, Bluetooth, WiMax, WLL,PSTN, PLC, zigbee, optical fibre or a combination thereof. ln embodiments, at least one of theplurality of sensor units may be a smart sensor unit that comprise additional computational andanalytical capabilities. By this, parts of the steps performed by the central unit 461 may be performed by the sensor unit.

Figure 4b shows another schematic illustration of the example system 400 for predictivemaintenance for a DH substation 420 using a prediction system, according to an aspect of thepresent disclosure. The system 400 comprises three DH substations 420, 421 and 422 and at leastone central unit such as the first central unit 461. The DH substation 420 is in communicationwith the at least one first central unit 461 via a data communication link 425. The DH substation421 is in communication with the at least one first central unit 461 via a data communication link426. The DH substation 422 is in communication with the at least one first central unit 461 via a data communication link 427.

The system 400 as illustrated in Figure 4b shows three DH substations in communication with thefirst central unit 461 for the purposes of ease of explanation and does not in any way limit thedisclosure to three DH substations. ln practice, the first central unit 461 may be capable of receiving, storing and computing sensor data incoming from more than three DH substations. ln a scenario where multiple DH substations are in communication with a first central unit, theDH substations may be provided with a unique identity code. The sensor data may be transmittedwith the unique identity code identifying the DH substations. The sensor data also be transmitted with geo tags identifying the geographical location where a DH substation is installed.

The first central unit 461 may be configured with computational capabilities to analyze andclassify sensor data from the DH substation into various classes. Each class may be associatedwith different maintenance requirements and actions to be performed. The analysis andclassification enables prediction of malfunction and thereby determining maintenance requirements.

Figure 4c shows another schematic illustration of the example system 400 for predictivemaintenance for a DH substation 420 using a prediction system, according to an aspect of thepresent disclosure. The system 400 comprises three DH substations 420, 421 and 422 and at leastone first central unit 461 and at least one second central unit 462. The DH substation 420 is incommunication with the at least one first central unit 461 via a data communication link 425. TheDH substation 421 is in communication with the at least one first central unit 461 via a data communication link 426. The DH substation 422 is in communication with the at least one first central unit 461 via a data communication link 427. The at least one first central unit 461 is in communication with the at least on second central unit 462 via a data communication link 465.

The system 400 as illustrated in Figure 4c shows three DH substations in communication with thefirst central unit 461. The use of three substations is an example for the purposes of ease ofexplanation and does not in any way limit the disclosure to three DH substations. ln practice, thefirst central unit 461 may be capable of receiving, storing and computing sensor data incoming from more than three DH substations. ln a scenario where multiple DH substations are in communication with a first central unit, theDH substations may be provided with a unique identity code. The sensor data may be transmittedwith the unique identity code identifying the DH substations. The sensor data also be transmitted with geo tags identifying the geographical location where a DH substation is installed.

The first central unit 461 may be configured with computational capabilities to analyze andclassify sensor data from the DH substation into various classes. Each class may be associatedwith different maintenance requirements and actions to be performed. The analysis andclassification enables prediction of malfunction and thereby determining maintenance requirements.

The first central unit 461 may for example transmit a determined malfunction prediction to thesecond central unit 462. The first central unit 461 may transmit the determined malfunctionprediction to the second central unit 462 with an instruction to display the determinedmalfunction prediction at a centralized display unit in communication with the first central unit461 and the second central unit 462. The second central unit 462 may further generate anotification based on the determined malfunction prediction and that notification and/orcorresponding information can be displayed on the centralized display unit. The operators at thedistrict heating office or quarters may monitor the display unit and then may notify or inform theservice technicians to check at least one component or related components of the DH substationfor maintenance requirements. Alternatively or additionally, the notification may be sent to adevice of a service technician. Alternatively or additionally, the notification may be sent to a device of a customer.

The processing load for computation and determination of malfunction prediction may bedistributed efficiently between the first central unit 461 and the second central unit 462. Suchdistribution of processing load may enable avoiding network congestions and delivering faster computations.

Figure 5 shows a schematic illustration of a system 500 used for predictive maintenance for DHsubstation according to an aspect of the present disclosure. The system 500 is illustrated tocomprise a central unit 560. The central unit 560 may for example be the first central unit 461 orthe second central unit 462. Alternatively, the central unit 560 may encompass the first centralunit and the second central unit within. The central unit 560 may further comprise storagedevices or memory 571 and one or more processing units 572. The memory includes instructionsfor processing data. The processor executes the instructions stored in the memory and facilitatesdetermining at least one malfunction prediction at the DH substation based on the measured atleast one parameter from each sensor unit of the plurality of sensor units. The central unit 560may further comprise one or more communication units 573 that enable communication of thecentral unit 560 with one or more DH substation 120, 220, 320, 420 and the plurality of the sensor units. Additionally, the central unit 560 may comprise a Machine Learning, ML, unit 574.

The memory 571 may comprise one or more preconfigured databases. One of the databases maycomprise historical data indicative of the operation of the components of the DH substation 120,220, 320, 420. The central unit 560 may store received data, for example as comprised in the communication from the plurality of sensor units.

As an example, all the historical data collected previously may be arranged in one super set andthe super set may comprise one or more subsets. Some of the one or more subsets maycorrespond to data indicative of normal operation ofthe components of the DH substation 120,220, 320, 420 and some of the one or more subsets may correspond to data indicative of faultyoperation of the components of the DH substation 120, 220, 320, 420. ln another example, allthe historical data collected previously may be arranged as one master set of data and there may be various patterns in the data set that may indicate faulty operation of the components of the DH substation. lt may be noted that historical data may be stored and arranged in various sequences and orders and the example above only shows one of the ways.

The central unit 560 may be configured to receive the sensor data continuously and/or atfrequent intervals from the plurality of sensor units. The sensor data comprises real time datacorresponding to each of the parameters. The preconfigured database may comprise historicaldata that may indicate normal operation of the DH substation. Alternatively, the preconfigured database may contain historical data that may indicate fault in the DH substation.

One or more databases may be configured to store real time sensor data. Additionally oralternatively, the real time sensor data may be stored momentarily in a cache memory. Thesensor data received at the central unit 560 may be stored in a database of the central unit 560.Additionally, the sensor data may be mapped to the associated components and the mappingrelation may be stored in the database. Further, the sensor data received at the central unit 560 may be sorted based on a timestamp and stored.The malfunction prediction may be based on the historic data.The historic data may comprise historic malfunctioning predictions.

Historic malfunctioning predictions may be used to estimate or set one or more thresholdsagainst which new real time sensor data may be measured. The new/real time sensor datareceived from the sensor units may be analysed by the ML unit 574 by comparing the new dataagainst the one or more thresholds. Accordingly, the new data also gets classified into one or more classes of malfunctioning predictions.

The processing unit 572 of the central unit 560 may be configured to estimate the one or morethresholds against which new real time sensor data may be measured to identify whether newmeasurements are normal or not and predict if there is any indication of failure to component.The thresholds may be estimated based on the historic data or the historic malfunctionpredictions. lf the value of a parameter/sensor data is below a certain threshold, a normal operation may be indicated. lf the value of a parameter is above the certain threshold, an indication of possible failure may be predicted. Different thresholds may be set or estimated for each parameter measured by the sensor units.

The data compared against threshold may later be classified into various classes using thetraining models of the ML unit 574. Each class is associated with a maintenance requirement type or an action to be performed.

Alternatively, various data patterns of historic malfunctioning predictions may be stored in thecentral unit. There may be at least two data patterns corresponding to one component of the DHsubstation. One of the at least two data patterns may indicate normal operation and the otherdata pattern may indicate faulty operation. The incoming sensor data can be compared againstthe data patterns. As an example, one data pattern may indicate normal operation of the RADpump and another data pattern may indicate faulty operation of the RAD pump. lf the incomingsensor data corresponding to the RAD pump matches the data pattern indicating normaloperation of the RD pump, it may be inferred that the RAD pump is operating under normaloperation. lf at least one value/bit in the incoming sensor data corresponding to the RAD pumpmatches the data pattern indicating faulty operation of the RAD pump, it may be inferred that the RAD pump has faults.

Historic data may include data previously analyzed and stored in the central unit 560.Additionally, historic data may include data, corresponding to system failure, which are generated, studied by experts and analyzed over the years.

Alternatively or additionally, the ML unit 574 may be configured to estimate one or morethresholds to identify whether new measurements are normal or not and predict if there is any indication of failure to component.

Exploratory analytics may be carried out to find meaningful insights in pools of sensor data andlearn from failure pattern. The recognized data patterns are reflected in predictive models withclassification approaches. Models are built, trained and properly evaluated. Moreover, the best- fit model/algorithm is selected, and also regular updated and tested for accuracy.

The central unit may be configured to estimate one or more thresholds to identify whether new measurements are normal or not and predict if there is any indication of failure to component.

The ML unit may be configured to estimate one or more thresholds to identify whether new measurements are normal or not and predict if there is any indication of failure to component. lf the value of a parameter is below a certain threshold, a normal operation may be indicated. lfthe value of a parameter is above the certain threshold, an indication of possible failure may be predicted.Different thresholds may be set or estimated for each parameter measured by the sensor units.

A first parameter of the plurality of parameters, measured at a first sensor of the plurality of sensor units, may be measured at a plurality of times.

The central unit 160, 360, 461, 462, 560 may determine the at least one malfunction prediction based on a trend analysis on the plurality of measurements of the first parameter.

The at least one malfunction prediction may be determined based on at least one of: time seriesanalysis, exponential smoothing method, univariate anomaly detection, multivariate anomalydetection, real-time anomaly detection, seasonal models, heat demand forecasting, trend analysis and/or limit adjustment.The central unit may comprise the ML unit.The ML unit may be a standalone unit remotely communicating with the central unit.

The sensor data is cleaned and validated at the ML unit to ensure consistent and accurate data are the foundation for exploratory analysis and modelling.The sensor data may be cleaned and/or validated using one or more known validation algorithms.

The sensor data may be cleaned using filter functions and/or validated using one or more validation algorithms.

Given all the interaction and/or connections, the predictions may be summarized and arrangedinto multiple classes in terms of size, duration and frequency of deviations. Each class may be associated with different maintenance requirements or actions to be taken.

The central unit 560 may determine a degree of a determined malfunction prediction, wherein the degree is based on historic data.

Historic data may include data previously analyzed and stored in the central unit. Additionally,historic data may include data, corresponding to system failure, which are generated, studied by experts and analyzed over the years.

The central unit 560 may determine a degree of a determined malfunction prediction, wherein the degree is based on previously determined malfunction predictions.

A degree of a determined malfunction prediction may indicate how accurate or reliable thedetermined malfunction prediction is. The degree of a determined malfunction prediction maybe a value between a range of values. The degree of a determined malfunction prediction may range from, for example, 1-5. ln an example, if the degree of malfunction predictions is 1, the prediction may not be too reliableor accurate. lf the degree of malfunction predictions is 3, the prediction may be reliable to acertain extent. lf the degree of malfunction predictions is 5, the prediction may be accurate or COfTeCt.

Severity of the maintenance tasks may also be associated with the degree of malfunctionpredictions. ln an example, if the degree of malfunction predictions is 1, maintenance may notbe required. lf the degree of malfunction predictions is 3, immediate maintenance may not berequired but may be monitored for the next few days. lf the degree of malfunction predictions is , immediate maintenance may be required.

The degree of malfunction prediction may be used in training ML models at the ML unit 574.Continuous training and sufficient data for training may gradually improve the algorithms and thereby efficient training models may be used for used for evaluation and classification of the sensor data. An improved training model may be beneficial to achieve very reliable, accurate and faster predictive maintenance.

Over time, sensor data received from the DH substation may become historic data and can be used for training ML models for malfunction classification and prediction.

The central unit 560 may be connected to a display unit (not shown). The first central unit 560may display the determined malfunction prediction at the display unit. Additionally, the displayunit may also display a maintenance requirement or maintenance type or an action required tobe performed depending on the determined malfunction prediction. The display unit may be a centralized display unit.

The central unit 560 may further generate a notification based on the determined malfunctionprediction and can be displayed on the centralized display unit. The operators at the districtheating office or quarters may monitor the display unit and then may notify or inform the servicetechnicians to check at least one component or related components of the DH substation formaintenance requirements. Alternatively or additionally, the notification may be sent to a deviceof a service technician. Alternatively or additionally, the notification may be sent to a device of a CUStOmef.

The term ”sensor unit” in this context means a unit comprising at least one, or a plurality, ofsensors that is configured to detect events or changes in its environment. The sensor unit as suchmay be a unit comprising the sensors, or several units having a sensor located to measureproperties at a similar location. Hence, the sensorunit can be an array of sensors, each measuringproperties on different phases, the sensor unit can also be a unit to which different phases are connected to.

The term ”central unit” in this context means a unit that suitable to be configured to receivereports from the first sensor unit or a plurality of sensor units. The central unit is furtherconfigured to perform computations, analysis, and to provide predictive maintenance. The functionality of the central unit may for example be implemented in a unit comprised in a DH system. The central unit may also be implemented as a specific unit in communication with the DH system.

The term ”indicated” is used when a fault is indicated by one or more of the plurality of sensorunits means that the sensor unit provides parameters from which predictive maintenance may be based.

The sensor unit may communicate with the central unit 560 via a data communication link. Thedata communication link may for an example utilize one or a plurality of different types of wiredlinks or wireless links, such as for example xDSL, 2G, 3G, 4G, 5G TCP/IP, WiFi, Bluetooth, WiMax, WLL, PSTN, PLC, zigbee, optical fibre or a combination thereof.

The system 100, 300, 400 and 500 may further comprise a computer program for performingpredictive maintenance for a DH substation comprising instructions which, when executed on atleast one processor, cause the at least one processor to carry out measuring, at each sensor unitof the plurality of sensor units in the DH substation, at least one parameter, transmitting, fromeach sensor unit of the plurality of sensor units, the measured at least one parameter, receiving,at the at least one central unit, the measured at least one parameter from each sensor unit ofthe plurality of sensor units and determining, at the at least one central unit, at least onemalfunction prediction at the DH substation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

The system 100, 300, 400 and 500 may further comprise a computer-readable storage medium carrying a computer program for performing predictive maintenance for DH substation.

Figures 1 to 5 show a simplified illustration of an embodiment of the DH substation, the centralunit and their contents. The simplified illustration is intended to convey understanding of thegeneral idea of storing different program functions in the system and lower-level details that arenot necessary to understand the techniques are omitted. A memory segment within the centralunit stores program code for controlling the central unit to perform operations described herein.ln some embodiments, the central unit executes as a monolithic application on a single device (e.g., a server) or as a distributed application one a plurality of computing devices (like in a microservices architecture), in some cases with replicated instances of various componentsexecuting behind load balancers at the direction of orchestration tooling configured to elasticallyscale the number of instances according to demand. Although the simplified illustration ofFigures 2 shows pseudo code, it is to be understood that the program code may be constitutedby machine code or any level program code that can be executed or interpreted by the centralunit and/or a user terminal. The program code may when run on the central unit and/or the firstuser terminal and will cause the central unit and/or the first user terminal to perform a functionsuch as a method described herein. The method may comprise an advanced mathematicalprocessing of the data. According to embodiments of the invention the program code is adaptedto cause a processor means to perform signal processing functions and methods described in this document. ln block diagrams, illustrated components are depicted as discrete functional blocks, butembodiments are not limited to systems in which the functionality described herein is organizedas illustrated. The functionality provided by each of the components may be provided bysoftware or hardware modules that are differently organized than is presently depicted, forexample such software or hardware may be intermingled, conjoined, replicated, broken up,distributed (e.g. within a data center or geographically), or otherwise differently organized. Thefunctionality described herein may be provided by one or more processors of one or morecomputers executing code stored on a tangible, non-transitory, machine readable medium. lnsome cases, notwithstanding use of the singular term "medium," the instructions may bedistributed on different storage devices associated with different computing devices, forinstance, with each computing device having a different subset of the instructions, animplementation consistent with usage of the singularterm ”medium” herein. ln some cases, thirdparty content delivery networks may host some or all of the information conveyed overnetworks, in which case, to the extent information (e.g., content) is said to be supplied orotherwise provided, the information may be provided by sending instructions to retrieve that information from a content delivery network.

As used throughout this application, the word ”may” is used in a permissive sense (i.e., meaninghaving the potential to), rather than the mandatory sense (i.e., meaning must). The words H ll ”include , including", and ”includes” and the like mean including, but not limited to. As used H ll throughout this application, the singularforms ”a, an," and ”the” include plural referents unlessthe content explicitly indicates otherwise. Thus, for example, reference to ”an element” or "theelement" includes a combination of two or more elements, notwithstanding use of other termsand phrases for one or more elements, such as ”one or more." The term "or" is, unless indicatedotherwise, non-exclusive, i.e., encompassing both "and" and "or." Terms describing conditionalrelationships, e.g., "in response to X, Y," "upon X, Y,", ”if X, Y," "when X, Y," and the like,encompass causal relationships in which the antecedent is a necessary causal condition, theantecedent is a sufficient causal condition, or the antecedent is a contributory causal conditionof the consequent, e.g., "state X occurs upon condition Y obtaining" is generic to "X occurs solelyupon Y" and "X occurs upon Y and Z." Such conditional relationships are not limited toconsequences that instantly follow the antecedent obtaining, as some consequences may bedelayed, and in conditional statements, antecedents are connected to their consequents, e.g.,the antecedent is relevant to the likelihood of the consequent occurring. Statements in which aplurality of attributes or functions are mapped to a plurality of objects (e.g., one or moreprocessors performing steps A, B, C, and D) encompasses both all such attributes or functionsbeing mapped to all such objects and subsets of the attributes or functions being mapped tosubsets of the attributes or functions (e.g., both all processors each performing steps A-D, and acase in which processor 1 performs step A, processor 2 performs step B and part of step C, andprocessor 3 performs part of step C and step D), unless otherwise indicated. Further, unlessotherwise indicated, statements that one value or action is ”based on” another condition orvalueencompass both instances in which the condition or value is the sole factor and instances in whichthe condition or value is one factor among a plurality of factors. Unless otherwise indicated,statements that ”each” instance of some collection have some property should not be read toexclude cases where some otherwise identical or similar members of a larger collection do nothave the property, i.e., each does not necessarily mean each and every. Limitations as to sequence of recited steps should not be read into the claims unless explicitly specified, e.g., with explicit language like ”after performing X, performing Y," in contrast to statements that might beimproperly argued to imply sequence limitations, like ”performing X on items, performing Y onthe X'ed items," used for purposes of making claims more readable rather than specifyingsequence. Statements referring to ”at least Z of A, B, and C," and the like (e.g., ”at least Z of A,B, or C"), refer to at least Z of the listed categories (A, B, and C) and do not require at least Z unitsin each category. Unless specifically stated otherwise, as apparent from the discussion, it isappreciated that throughout this specification discussions utilizing terms such as ”processing,” H ll ”computing, calculating," ”determining” or the like refer to actions or processes of a specificapparatus, such as a special purpose computer or a similar special purpose electronicprocessing/computing device. The terms "first", "second", "third," ”given” and so on, if used inthe claims, are used to distinguish or otherwise identify, and not to show a sequential ornumerical limitation. As is the case in ordinary usage in the field, data structures and formatsdescribed with reference to uses salient to a human need not be presented in a human-intelligible format to constitute the described data structure or format, e.g., text need not berendered or even encoded in Unicode or ASCII to constitute text; images, maps, and data-visualizations need not be displayed or decoded to constitute images, maps, and data- visualizations, respectively; speech, music, and other audio need not be emitted through a speaker or decoded to constitute speech, music, or other audio, respectively.

While specific embodiments have been described, the skilled person will understand thatvarious modifications and alterations are conceivable within the scope as defined in the appended claims.

Claims (22)

1. Method (200) for predictive maintenance for a district heating, DH, substation (120, 320,420) using a prediction system, having a plurality of sensor units configured to collectmeasurements in the DH substation (120, 320, 420) and being in communication with at leastone central unit (160, 360, 461, 560), wherein the DH substation (120, 320, 420) comprise a heatexchanger (131, 331, 341) configured to transfer heating energy from a DH network to a user heating system, and wherein the method comprises the steps of: measuring (210), at each sensor unit of the plurality of sensor units in the DH substation, at leaSt One parametef) transmitting (220), from each sensor unit of the plurality of sensor units, the measured at leaSt One pafametef; receiving (230), at the at least one central unit, the measured at least one parameter from each sensor unit of the plurality of sensor units; determining (240), at the at least one central unit, at least one malfunction prediction atthe DH substation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

2. Method according to claim 1, wherein the user heating system comprise a hot watersystem (DHW) and a space heating system (RAD), and the DH substation (120, 320, 420) comprisea DHW heat exchanger (131, 331) configured to transfer heating energy from the DH network tothe DHW, and a RAD heat exchanger (341) configured to transfer heating energy from the DH network to the RAD.

3. Method according to claim 2, wherein the DHW comprise a DHW control valve (313), a DHW circulation pump (333), a DHW security valve (334), and interconnecting pipes (332, 342); the RAD comprise a RAD control valve (315), a RAD circulation pump (345), a RAD security valve(344), a RAD expansion vessel (343), and interconnecting pipes (332, 342); the DH substation(120, 320, 420) further comprise a DH supply (111, 311), a DH return (112, 312), and a cold water supply; and wherein the at least one sensor units is chosen from: a thermal sensor arranged on the DHWcontrol valve (313), a thermal sensor arranged on the RAD control valve (315), a thermal sensorarranged on the RAD circulation pump (345), a thermal sensor arranged on a pipe prior to theDHW circulation pump (333), a thermal sensor arranged on the cold water supply, a thermalsensor arranged on a pipe subsequent to the DHW heat exchanger (331), a thermal sensorarranged on the DHW security valve (334), a thermal sensor arranged on the RAD pump motor,a thermal sensor arranged on a pipe prior to the RAD heat exchanger (341), a pressure sensorarranged on the RAD expansion vessel (343), a pressure sensor arranged on a manometer in theDH supply, c, a thermal sensor arranged on the DH supply flow, a thermal sensor arranged on the DH return flow, and a flow sensor arranged on the DH return.

4. Method according to claim 3, wherein: the plurality of sensor units comprise the thermal sensor arranged on the DHWcontrol valve (313), the thermal sensor arranged on the RAD control valve (314), the thermalsensor arranged on the pipe prior to the DHW circulation pump (333), the thermal sensorarranged on the cold water supply, the thermal sensor arranged on the DHW security valve (334),the thermal sensor arranged on the RAD pump motor, the pressure sensor arranged on the RADexpansion vessel (343), the thermal sensor arranged on the DH return flow, and the flow sensor arranged on the DH return; the measured at least one parameter comprise a temperature of the DHW control valve(313), a temperature of the RAD control valve (314), a temperature of the DHW circuit return, a temperature of the cold water supply, a temperature of the security valve, a temperature of the RAD pump (345), a pressure of the RAD expansion vessel (343), a temperature of the DH return flow, and a flow of the DH return; and determining the at least one malfunction prediction comprise determining leakage.

5. Method according to claim 3, wherein: the plurality of sensor units comprise a thermal sensor arranged on the DHWcontrol valve (313), a thermal sensor arranged on the RAD control valve (314), a thermal sensorarranged on a pipe prior to the DHW circulation pump (333), a thermal sensor arranged on thecold water supply, a thermal sensor arranged on a pipe subsequent to the DHW heat exchanger(331), a thermal sensor arranged on the RAD pump motor, a thermal sensor arranged on the DH return flow, and a flow sensor arranged on the DH return; and the plurality of parameters comprise a temperature of the DHW control valve (313),a temperature of the RAD control valve (314), a temperature of the DHW circuit return, atemperature of the cold water supply, a temperature of the DHW circuit supply, a temperature ofthe RAD pump (345), a temperature of the DH return flow, and a flow of the DH return; and determining the at least one malfunction prediction comprise determining a malfunctioning of valves and VVS system.

6. Method according to claim 3, wherein: the plurality of sensor units comprise a thermal sensor arranged on the DHWcontrol valve (313), a thermal sensor arranged on a pipe prior to the DHWcirculation pump (333), and a thermal sensor arranged on the RAD circulation pump (345), the plurality of parameters comprise a temperature of the DHW control valve (313), a temperature of the DHW circuit return, and temperature of the RAD pump (345), and determining the at least one malfunction prediction comprise determining malfunctioning pumps.

7. Method according to claim 3, wherein: the plurality of sensor units comprise a thermal sensor arranged on the RAD controlvalve (314), a thermal sensor arranged on a pipe prior to the RAD heat exchanger (341), apressure sensor arranged on the RAD expansion vessel (343), a pressure sensor arranged on amanometer in the DH supply, a pressure sensor arranged on a manometer in the DH return, athermal sensor arranged on the DH supply flow, a thermal sensor arranged on the DH return flow, and a flow sensor arranged on the DH return; the plurality of parameters comprise a temperature of the RAD control valve (314),a temperature of the RAD circuit supply, a pressure of the RAD expansion vessel (343), adifferential pressure based on a pressure of the DH supply and a pressure of the DH return, atemperature of the DH supply flow, a temperature of the DH return flow, and a flow of the DH return; and determining the at least one malfunction prediction comprise determining low filter performance.

8. Method according to claim 3, wherein: the plurality of sensor units comprise a thermal sensor arranged on the RAD controlvalve (314), a thermal sensor arranged on the RAD circulation pump (334), a thermal sensorarranged on a pipe prior to the RAD heat exchanger (341), and a pressure sensor arranged on the RAD expansion vessel (343), the plurality of parameters comprise a temperature of the RAD control valve (314),a temperature of the RAD pump motor, a temperature of the RAD circuit supply, and a pressure of the RAD expansion vessel (343), and determining the at least one malfunction prediction comprise determining a malfunctioning heat exchanger and/or pumps.

9. Method according to any preceding claim, wherein at least one of the at least one sensor unit of the plurality of sensor units is a surface mounted sensor.

10. Method according to any preceding claim, wherein the at least one parameter is chosen from: temperature, pressure, and flow.

11. Method according to any preceding claim, wherein the malfunction prediction is chosenfrom: leakage, low filter performance, valves and DHW system, fouling of heat exchanger, and malfunctioning pumps.

12. Method according to any preceding claim, wherein the malfunction prediction also is based on historic data.

13. Method according to claim 12, wherein the historic data comprise historic malfunctioning predictions.

14. 4214. Method according to any preceding claim, further comprising determining, at a central unit (160, 360, 461, 462, 560), a degree of a determined malfunction prediction, wherein the degree is based on previously determined malfunction predictions.

15. Method according to any preceding claim, further comprising transmitting, from a firstcentral unit (160, 360, 461, 560), a determined malfunction prediction to a second central unit (462).

16. Method according to claim 9, wherein the first central unit (160, 360, 461, 560) is a cloud SerVef.

17. Method according to any preceding claim, wherein a first parameter of the plurality ofparameters, measured at a first sensor of the plurality of sensor units, is measured at a plurality of times; and the central unit (160, 360, 461, 462, 560) determines the at least one malfunction prediction based on a trend analysis on the plurality of measurements of the first parameter.

18. Method according to any preceding claim, wherein the at least one malfunctionprediction is determined based on at least one of: time series analysis, exponential smoothingmethod, univariate anomaly detection, multivariate anomaly detection, real-time anomaly detection, seasonal models, heat demand forecasting, trend analysis and/or limit adjustment.

19. System for predictive maintenance for a district heating, DH, substation (using a prediction system, wherein the DH substation (120, 320, 420) comprise a heat exchanger (131, 331, 341) configured to transfer heating energy from a DH network to a user heating system, and wherein the system comprise:at least one central unit (160, 360, 461, 560); a plurality of sensor units (350) configured to collect measurements in the DH substation and being in communication with the at least one central unit (160, 360, 461, 560); and wherein each sensor unit of the plurality of sensor units in the DH substation (120, 320,420) is configured to measure at least one parameter and to transmit the measured at least oneparameter, and the at least one central unit (160, 360, 461, 560) is configured to receive themeasured at least one parameter from each sensor unit of the plurality of sensor units anddetermine at least one malfunction prediction at the DH substation based on the measured at least one parameter from each sensor unit of the plurality of sensor units.

20. System according to claim 18, that is further configured to perform the method of anyclaim 1 to 18.

21. A computer program for performing predictive maintenance for a district heating, DH, substation (120, 320, 420) comprising instructions which, when executed on at least oneprocessor, cause the at least one processor to carry out the method according to any of claims 1 to 18.

22. A computer-readable storage medium carrying a computer program for performing predictive maintenance for a district heating, DH, substation (120, 320, 420) according to claim.