SE1151219A1 - Procedure for optimizing filter usage time between switches and systems for monitoring a ventilation system - Google Patents

Abstract

22ABSTRACT The present document discloses a method of determining an optimalfilter use time period between replacements of a filter in a ventilation system(1). The method comprises receiving a filter hardware value, representing anamount of a resource associated with at least production of the filter,receiving a filter use value, representing an amount or rate of said resourceassociated with use of the filter, receiving at least one measured datarepresenting a measured pressure drop over the filter, and determining saidoptimal filter use time period based on said filter hardware value, said filteruse value and said measurement data. Elected for publication: Fig. 1

Description

METHOD OF OPTIMIZING FILTER USE TIME BETWEEN REPLACEMENTSAND SYSTEM FOR MONITORING A VENTILATION SYSTEM Technical Field The invention relates to a method and system for optimizing filter usetimes between filter replacements in ventilation systems.

The invention provides a method for improving the overall filter costand/or carbon dioxide impact of ventilation systems.

BackgroundVentilation systems used in buildings, ships or other major structures typically comprise ventilation ducts and a fan arranged to drive air through theventilation ducts. However, many systems also comprise further components,such as heat exchangers and moisture exchangers, operating to providing adesired indoor climate. ln order to reduce exposure to particles, both for users and for theventilation system components, the system normally comprises one or morefilters, arranged to filter incoming and/or outgoing air.

Such filters are typically consumables and need to be replaced atcertain intervals. As there is a cost associated with the filter itself, and with thelabor required to exchange it, there is a general motivation to replace filters asseldom as possible, and preferably only at the very end of their technicalservice life.

Moreover, as a filter is being used, it gradually fills up with particleswhich have been separated from the air being filtered. As the filter fills up, itwill provide greater resistance to air flowing through it, thus requiring the fanto work harder. As the fan works harder, its energy consumption will increase.Hence, there is also a motivation to replace filters as often as possible, inorder to minimize energy consumption.

Various methods for monitoring filter status are disclosed inFR2770788A, US2005247194A, US2008014853A, JP201 1 191017A,US6035851A, US2007146148A and US5036698A. However, these methodsare directed at predicting the filter's technical service life, i.e. for how long the 2 filter will work good enough. They fail to consider the fact that while a filterworks good enough, i.e. provides a good enough filtration and low enoughpressure drop, the actual total resource consumption would actually bedecreased by replacing the filter well before the end of its technical servicelife.

Hence, there is a need for a method of optimizing the time periodbetween filter replacements in order to achieve an overall improved resourceutilization.

Summarylt is an object of the present disclosure to provide a method and system for optimizing filter use time between replacements in ventilationsystems.

The invention is defined by the appended independent claims, withembodiments being set forth in the appended dependent claims, in thefollowing description and in the drawings.

According to a first aspect, there is provided a method of determiningan optimal filter use time period between replacements of a filter in aventilation system. The method comprises receiving a filter hardware value,representing an amount of a resource associated with at least production ofthe filter, receiving a filter use value, representing an amount or ratio of saidresource associated with use of the filter, receiving at least one measureddata point representing a measured pressure drop over the filter, anddetermining said optimal filter use time period based on said filter hardwarevalue, said filter use value and said measurement data.

The term “filter” should be construed as the assembly which is actuallybeing replaced, and may thus include only the filter medium, or the filtermedium and a frame on which the filter medium is mounted. Moreover, theterm “filter” may comprise both a single filter unit and a filter assembly of twoor more individual filters which are connected in series and/or in parallel.

The “filter hardware value” can be seen as the resource consumptionassociated with changing a filter, i.e. primarily the cost (or other resourceconsumption, such as C02 impact) of the filter itself, but it may also extend to 3 costs such as the cost of transporting the filter from a distribution location tothe filter use location, the cost for the operator performing the filter change,possibly also including any travel cost, and the costs for disposal of the usedfilter.

The “measured pressure drop over the filter” is a measurement of thepressure drop provided by the actual filter when it is in place of operation.

The present disclosure is thus based on the recognition that, whilefrom a filter cost perspective, it is desirable to replace filters as seldom aspossible, from a fan drive energy perspective; it may be desirable to replacefilters more often. ln addition to the cost perspective, there is also the carbon dioxideperspective, since the production (and distribution) of the filter itself gives riseto some carbon dioxide impact on the environment, as does the energy usedto drive the fan.

Hence, the present invention provides a method which enablesplanning of filter changes with a view to minimizing a resource consumption,such as cost or carbon dioxide impact.

The method may further comprise estimating a total resourceconsumption for a given time period based on the filter hardware value andthe filter use value, and determining the optimal use time period so as tosubstantially minimize the total resource consumption.

The total resource consumption for the given time period may bedetermined by a first factor, according to which the resource consumption isinversely proportional to the time period, and a second factor, according towhich the resource consumption is directly proportional to the time period.

The first factor may be determined as a product of the filter hardwarevalue, an inverse of the use time period, of an uptime ratio, and optionally oneor more constants.

The second factor may be determined as a product of the filter usevalue, an airflow through the filter, an average pressure dropover the filter, thetime period, an uptime ratio an inverse of a fan efficiency, and optionally oneor more constants. 4 The airflow is generally a nominal one, i.e. it is determined by the capacity of the system or of the system setting.The average pressure drop may be determined based on at least two measured data points.

The average pressure drop may be determined based on at least oneestimated pressure drop data item.

The estimated pressure drop may have the form Pa(t) = startpa*eb*t,where startpa is an initial measured pressure drop, b is an environment coefficient and t is the time.

The initial measured pressure drop may be determined as an averageof at least two measured pressure drops. ln one embodiment, or initially, the environment coefficient may beempirically determined. ln another embodiment, or after an initial phase, the environmentcoefficient may be determined based on a regression analysis matching atleast some of the measured data points to an exponential function. ln particular, the environment coefficient may be determined accordingto b = Lolog (ïj-'OÜ/tn, wherein po is an initial pressure drop, and wherein t”is the time of the n:th data point, p” is the pressure drop of the n:th data point and N is the number of data points.The environment coefficient (b) may be determined according to b _ N*Zg=0 tn*pn_Z1I”\L,=0 tTL ZII”\I.I=O pni 2N*Eg=o tizæ"(zy=o fn) p” is the pressure drop of the n:th data point and N is the number of data , wherein t” is the time of the n:th data point, points.

The method may further comprise determining whether the pressuredrop over the filter increases substantially linearly over time.

The method may further comprise determining coefficients of a linearequation based on at least some of the measurement data points, preferablyall but the latest measurement data points, inserting a time of a latestmeasurement data point into the linear equation, comparing a thus estimatedpressure drop with a corresponding measured pressure drop, if thedifference between the estimated pressure drop and the measured pressure 5drop is less than a predetermined threshold value, determining that the pressure drop increases substantially linearly over time, and if the differencebetween the estimated pressure drop and the measured pressure drop isgreater than the predetermined threshold value, determining that the pressuredrop increases substantially exponentially over time.lf it is determined that the pressure drop increases substantially exponentially, at least two different environment coefficients may beestimated based on different numbers of measurement data points, and thethus calculated coefficient having the greatest value is used for determiningthe estimated pressure drop. ln one embodiment, the resource may be a cost. ln another embodiment, the resource may be a carbon dioxide impact.

The method may further comprise receiving a filter replacement value,representing an amount of said resource associated with replacement of thefilter, wherein the optimal filter use time period is determined also with respectto the filter replacement value.

The method may further comprise presenting the optimal filter use timeperiod in a user-readable format. For example, the optimal time period maybe expressed as a part of a graph, as a date, or as a time period remaininguntil the optimal filter replacement time. ln the method, the at least one measured data point may comprise atleast one data point which is higher than an initial pressure drop and lowerthan a final pressure drop of the filter, and which is at least about 5, preferablyabout 10, about 20 or about 30 days apart from both a time of the initialpressure drop and a time of the final pressure drop.

Hence, the data points may be distributed over the life of the filter, andnot limited to periods immediately before or after a filter change.

According to a second aspect, there is provided a method of optimizingreplacements of filters at a plurality of filter locations. The method comprisesperforming the method as described above for each of at least two of thefilters, which are of different types. The method also comprises providing areplacement schedule for said at least two of the filters based on theirrespective optimal filter use time periods. 6 Hence, the optimization achieved by the described method may beused as input for a scheduling system, which can be used for scheduling filtermaintenance. Thus, filter replacements may be scheduled based on theoptimal filter use time for each filter, with due consideration taken to e.g. thefact that it may be more advantageous to replace two filters at one location (oradjacent locations) than travelling separately for each filter replacement.

According to a second aspect, there is provided a system formonitoring a ventilation system, comprising means for receiving a filterhardware value, representing an amount of a resource associated with atleast production of the filter, means for receiving a filter use value,representing an amount of said resource associated with use of the filter,measurement means for measuring a pressure drop over a filter arranged insaid ventilation system, and processing means for determining an optimalfilter use time period based on said filter hardware value, said filter use valueand said measurement data.

The system may be configured to implement the above-describedmethod or methods.

The processing means may have the form of a mobile hand-helddevice.Hence, the processing means may be in the form of a “smartphone”, alaptop or a tablet computer (such as an iPad®).

Brief description of the Drawinqs Fig. 1 schematically illustrates a ventilation system 1, to which thepresent disclosure is applicable.

Fig. 2 is a schematic diagram of a software structure for implementingthe technology according to the present disclosure.

Figs 3-7 show different views of the software's user interface.

Figs 8-12 show graphs representing calculation results based onexemple data.

Description of EmbodimentsFig. 1 schematically illustrates a ventilation system 1, which can beused to provide air to/from the rooms of e.g. a building. The system 1 7 comprises ventilation ducts 20, 22, a fan 21 for driving air through theventilation ducts and a filter module 10, adapted for receiving a replaceablefilter cartridge 11. ln the filter module 10, there is provided a measurementdevice 12a, 12b for measuring a pressure drop over the filter. For examplethe measurement device may comprise first and second pressure sensors12a, 12b. The measurement device may be connected to a controller 30,which may be adapted for receiving measurement data from the pressuresensors 12a, 12b.

The controller 30 may be arranged to receive a respective pressurevalue from the sensors 12a, 12b and to calculate the pressure drop. As analternative, the sensors may be arranged to directly measure the pressuredrop, and thus to provide a single value to the controller 30.

The controller 30 may be arranged to read values from the sensors12a, 12b continuously or at predetermined intervals and to store the receiveddata in a memory. As an alternative, the controller 30 may be arranged toread values from the sensors only when being polled.

The controller 30 may be arranged to communicate with a remote unit31, which may be a computer, a mobile phone/smart phone, etc. ln one embodiment, the controller 30 may be provided in the form of adedicated unit having a sensor interface and a communications device, whichmay be arranged to communicate via e.g. a text messaging service (“SMS” -Short Message Service) or Wi-Fi (ftp, e-mail, etc.). The unit may be arrangedto send sensor data at predetermined intervals or only when polled. Forexample, the unit may be arranged to automatically reply to an incoming textmessage by sending the current sensor value. The unit thus need not haveany memory or processing device at all, but merely the necessary interfacesand an a/d converter. ln this embodiment, all data storage and processingmay take place in the remote unit 31, possibly with a backup function beingprovided. ln another embodiment, a software may be provided for performing themethods disclosed herein, which is either in the form of a program stored onthe controller, or in a computer which is in communication with the controller,and which is accessible by the remote unit, e.g. via a web browser. As 8 another option, the software may be provided in the form of a downloadableapplication software (known as an “app”), which is downloaded to the remoteunit. ln one embodiment, the software is an application software which isconfigured to run on a smartphone, such as an iPhone®, or tablet-type PC,such as an iPad®, which is in direct communication with the controller andwhich has a backup function provided either through a docking function with ahost computer or through a cloud-based service (such as iC|oud®).

The software may be structured as out|ined in Fig. 3, with a main menu40, an overview window 41, a filter entry window 42, which for each filterprovides environment data entry 43, filter-specific data entry 44, pressuredrop entry 45, life cycle cost optimization 46 and C02 impact optimization 47.

The filter entry window 42 is adapted for adding a new filter, changingor adding data relating to an existing filter. The new filter entry window mayhave first portion 43 or tab for inputting system data for the specific filterlocation. The system data may include airflow 431 (in e.g. m3/hour), fan type432 (e.g. rpm regulated or not) uptime 433 (in percentage or hours per day,etc.), air type 434 (e.g. classified as “central city", “city”, “countryside”), filterstep 435 (where there are multiple filters arranged in series), energy type 436(e.g. “Swedish mixture”, “European mixture” or “green energy”), energy cost437 (e.g. as a price per kWh) and fan efficiency 438, e.g. as a percentage.

A second portion 44 or tab may be provided for entering filter data for aparticular system. The second portion 44 may include an input window 441and a table/overview window 442. The filter data may include data such asnumber of filters 443, filter type 444, filter article number 445, filter cost 446and filter associated carbon dioxide impact value 447. The filter data mayalso, if desired, include data on e.g. filter size (area) and filter material. Theremay also be an entry allowing the user or system to indicate whether andwhen a filter has been replaced. lt is possible to provide a menu of variousfilter types, the data of which are preprogrammed.

A third portion or tab 45 may be provided for entry of pressure dropdata for a particular filter. The third portion may include a status field 451showing a date for a next filter replacement and another field 452 showing theprojected pressure drop at the replacement date. The entry may cause the 9 system to poll the respective controller 30 or sensor 12a, 12b, it maycomprise a poll scheduling function, a presentation function for data (push)sent by the controller 30, or a manual data entry function. Typically, eachentry would include a date, optionally a time, and a pressure drop (e.g. in Pa)or other value indicative thereof. The third portion may also present apressure drop forecast, which may be derived as will be described below. Theforecast may be presented as a graph 453 showing pressure drop as afunction oftime. Measured data points may be presented as points in thegraph, and the graph may also show a date for a critical pressure drop (i.e.end of filter life) and/or a date for a next recommended (or optimal) filterreplacement. A table 454 may be provided showing the measured data pointsas pressure drops with an associated date.

A fourth portion 46 or tab may be provided for presenting anoptimization result with respect to a first resource, such as cost. The fourthportion may include a status field 461 showing a date for a next filterreplacement and another field 462 showing the projected pressure drop at thereplacement date. The result may be presented as a graph 464 showing atotal cost Cfofa, as a function of replacement time T. The optimum replacementtime T may be indicated in the graph, as may the remaining life of the currentfilter. lt is also possible to present, in a separate graph 463 and/ornumerically, the components f(T), e(T) forming the total cost Cfofai. The costmay be summarized over a predetermined period of time, such as one year.

A fifth portion 47 or tab may be provided for presenting an optimizationresult with respect to a second resource, such as carbon dioxide impact. Thefourth portion may include a status field 471 showing a date for a next filterreplacement and another field 472 showing the projected pressure drop at thereplacement date. The presentation may be analogous to that of the fourthportion 46, with the optimization being based on carbon dioxide impactinstead of cost, and with the annual carbon dioxide impact being presented at473, 474 in terms of e.g. kg of C02 per year. lt is noted that the input and output functions may be varied in terms offunction and appearance.

The description will now be directed to the optimization algorithm.Based on the data input as described above, a first part of the optimizationalgorithm will provide a prediction of the pressure drop development. lnitially, when there is no, or only one pressure drop data pointavailable, the pressure drop will be estimated according to Equation 1 below,wherein Pa(t) is the estimated pressure drop at time t(which may beexpressed in e.g. hours); startpa is the initial measured pressure drop; airtypeis a constant which may be empirically derived and specific to a type or classof air (e.g. “central city air", “city air”, “countryside air”, or “exhaust air”).

Pa(t) = startpa =i< eairtype” (Equation 1) The airtype constant for a specific air type (or even for a specificlocation) may be derived empirically, by testing filters in accordance with agiven standard, such as EN 13779, at a specific location.

From experience, the present inventors have learned that in the firstpart of a filter's useful life, the pressure drop tends to vary, which may resultin a decrease in the accuracy of the prediction. Hence, as more measuredpressure drops become available, the startpa may be calculated as anaverage of these pressure drops. lt is possible to provide a weighting of themeasured pressure drop, such that later measured pressure drops are givena greater weight. This averaging of the startpa may be practiced e.g. duringthe first approximately 1-45 (preferably about 15-25, about 25-35 or about30) days from the installation of a new filter. This period of averaging will bereferred to as the “initial period”. lt is also possible to calculate startpa with help from series of measuredpressure drops from previous associated filter replacements. Information fromprevious associated replacements may also be used for the pressure dropprediction, to increase the accuracy of the prediction when the filters areidentical and at the same filter position in the system.

After the initial period, and as more measured data become available,the pressure drop will be estimated according to Equation 2 below, where b isa constant derived from Equation 3. 11 Pa(t) = startpa =i< eb” (Equation 2) b = Loiog (ï-*oö/fn (Equation 3) ln Equation 3, N is the number of measured data, and p” is themeasured pressure drop at time t”. Hence, Equation 3 effectively fits theexponential curve of Equation 2 to the measured pressure drop data points.

The resource consumption (e.g. cost or carbon dioxide impact) for a given time period Tof the filtration is provided by two components: i) the resource consumption of the filter or filter unit itself over thattime period (i.e. the resource consumption for manufacturing,distributing and/or recycling or disposing of the filter, and ii) the resource consumption of driving air through the filter, i.e. theresources consumed by the fan. lt is noted that the resource can be a cost or e.g. a carbon dioxide impact on environment.

The resource consumption of the filter or filter unit is given by Equation 4 below, wherein f(T) is the resource consumption, Cfi/fe, is the resourceconsumption for the filter or filter unit, RUP is the uptime (expressed in e.g.hours per day).

Cfilter*365*RupT f(T) = (Equation 4)The resource consumption of running the filter is given by Equation 5 below, wherein e(T) is the resource consumption, Cuse is the resourceconsumption per time unit of use, Q is the airflow, pavg is the averagepressure drop as derived by Equation 6 below and fyfan is the efficiency of thefan.

Q* Pai/ge(T) I Cuse iT]fan*1000 =i< (365 =i< Rup) (Equation 5) pmm = åff parodi: (Equation e) 12 Hence, the average pressure drop over a specific time period T isderived from the Equation 2, which provides the pressure drop for the entirefilter life based on the measured data, whereas the expected future pressuredrop is estimated.

The total resource consumption, Cfofa, is the sum of the twocomponents according to Equation 7 below: Ctowl = f(T) + e(T) (Equation 7) As is evident from the disclosure above, the first component f(T)decreases with an increase in filter life, i.e. it is inversely proportional to thetime T. On the other hand, the second component e(T) increases with anincrease in filter life, i.e. it is directly proportional to the time T.

An example based on realistic data will now be given with reference toFigs 8-12.

The example starts from the following assupmtions: air flow, Q=0.9444m3/h; filter up time Rup=16 h; filter cost, Cfi,,e,=550 SEK, filter carbon dioxideimpact, CO2fi,,e,=13 kg; energy cost, Cuse=1.1 SEK/kWh; fan efficiency,f7fan=50%; airtype=0.00015. ln a first case, where the measured pressure drop is 80 Pa, theresulting pressure drop prediction according to Equation 1 will be asillustrated in Fig. 8, with the time in hours on the horizontal axis.

The total annual filter cost will according to Equation 4 will be asillutrated in Fig. 9.

The average pressure drop for various replacement intervals Twill becalculated according to Equation 6 and illustrated in Fig. 10.

The energy cost of as a function of the replacement interval T iscalculated according to Equation 5 and illustrated in Fig. 11.

The total cost Cfofa, is calculated according to Equation 7 and illustratedin Fig. 12. From Fig. 12, it appears as if the lowest cost is achieved with areplacement interval Tof approximately 5000 hours, which with an uptime RUPof 16 h/day means filter replacement every 11 months. 13 Fig. 12 thus schematically illustrates a possible appearance of the twocomponents, as well as of the sum. As can be seen, there is a clear minimumin the composite curve Cfofaf, which represents the optimal time between filterreplacements. The optimization will thus involve finding that time T.

The optimization results achieved may be used for planning of filterreplacement in one or more ventilation systems. By combining data on filterreplacement times Tand calendar data, it is possible to determine at whatpoints in time filter replacements should be made, which, in turn can be usedin workforce allocation planning.

For example, a graph (not shown) may be provided showing thenumber of filters (e.g. over all number of filters or filters per site) having aparticular status with respect to remaining filter life. For example, filters havingmore than 20 weeks life left may form a first group, filters having four to 20weeks life left may form a second group, filters having zero to four weeks leftmay form a third group and filters being overdue for filter replacement mayform a fourth group. ln a further embodiment, data relating to the geographic position ofeach filter and/or system may be provided. Such data may be used to basethe filter replacement planning not only on the predicted condition of eachfilter, but also on the cost and/or carbon dioxide impact associated with theoperation of replacing the filter. For example, the labor and transportation costand/or carbon dioxide impact for replacing a particular filter may becalculated, and used as a factor in the optimization operation. lt is also possible to implement a route planning algorithm aiming atminimizing the transportation work for filters and/or travel times of operatorsassociated with filter replacement at a plurality of sites.

The present disclosure finds particular application in ventilation systemscontrolled to provide a constant air flow. Such ventilation systems are typicallyused in buildings and other large and/or fixed structures, such as off-shoreplatforms and sea-going vessels. Filters where this disclosure is applicablemay be filters of the types G3-F9 according to EN779 and H10-H14 accordingto EN1822. Such filters may be formed as bag filters, pocket filters, filter pads,panel filters or pleated compact filters. Typically, the invention is useful for 14filters having relatively large airflows, such as more than about 0.1 m3/s ormore than about 0.47 m3/s.lt is possible to refine the algorithms of the present disclosure byadding additional sensor data. For example data on air particle contents may be used.Applicant has noted that over the life of a filter, the pressure drop may initially increase substantially linearly with time, but may later on begin toincrease substantially exponentially instead. ln order to increase the accuracyof the filter life prediction, it may be desirable to determine whether a specificfilter is in its “linear phase” or in its “exponential phase”.

One way of making this determination is by fitting a series of Nmeasurement data sets, each comprising a time and a pressure drop, to thegeneral linear equation, as per Equation 8. y(t)=k*t+m. (Equation 8) The measurement data, typically all but the latest data point, areinserted into Equation 9 below, whereby coefficients m and k are calculated. ri = l ” m” ”ll * l lk Zz=0 tn z=0 trzt Zz=0 pn * tn By inserting the coefficients m, k derived through Equation 9 and the (Equation 9) latest time data point into Equation 8, the estimated pressure drop for thatpoint can be calculated. This estimated pressure drop can be compared withthe actual pressure drop (i.e. of the latest data point) for that time and thedifference can be determined, e.g. as a percentage.

This difference may be subjected to a threshold determination,whereby the filter is determined to be in its linear phase if the difference isless than a predetermined threshold value. This threshold value may be e.g.about 10% to 30%, preferably about 20% for the filter's first 2000 hours ofuse.

The threshold value may vary with the filter's use time. For example,the threshold value may decrease with an increasing use time. Thus, the threshold value may be about 10%-30% for the filter's first 2000 hours of use, about 5%-15% (preferably about 10%) for the filter's next 2000 hours of useand about 2%-7% (preferably about 5%) when the filter has been used 4000hours or more. lt should be noted that it is possible to provide an arbitrary number ofthreshold values, and in one embodiment, the threshold value may becalculated as a function of the use time.

Hence, the algorithm may be made increasingly sensitive to deviationsfrom the linear equation as the filter use time increases.

The number of of measurement data points used for determining theenvironment coefficient b of Equation 2 may also vary depending on whetherthe filter is in the linear phase or not. lf the filter is determined as being in the linear phase, all measurementpoints may be used for estimating the coefficient b. lf the filter is determined as being in the exponential phase, a smallernumber of points may be used for estimating the coefficient b. For example,separate estimates may be made based on the last two, three, four or fivedata points, thus providing two or more different values of the coefficient b.The highest value of b is typically selected for use in the prediction (Equation2).

Equation 10 below illustrates a determination of the coefficient b for N measurement data sets. b _ N*Z'II”\LI=O tn*pn_Z1I”\L,=0 tTL Z1l¥=Opni 2N*Eg=o tizæ"(zy=o fn) (Equation 10) The coefficient found in Equation 10 is inserted into Equation 2' toestimate the pressure drop.p(t) = lastPointpmmwe* em, 0 < t< maxTime (Equation 2') The time for each estimated pressure drop data point is added to the time of the last measurement. 16 [t + lastPointn-me, p(t)] Finally, it is noted that the average pressure drop for a filter use time Tmay, as an alternative, be calculated according to Equation 6' below, where Nis the number of measurement data sets to the time T. parfym = äïkäunn - rn) (Equation fr) The resulting pavg from Equation 6' may be inserted into Equation 5.

The resource on which the optimization is based may, as mentionedabove, be cost or carbon dioxide impact. Other non-limiting examples ofresources may be transportation work or other environmental effects (such asexhaust substances).

Claims (24)

1. A method of determining an optimal filter use time period (Topf)between repiacements of a filter in a ventilation system, comprising: receiving at least one filter hardware value, representing an amount ofa resource associated with at least production of the filter, receiving at least one filter use value, representing an amount or rateof said resource associated with use of the filter, receiving at least one measured data point (tn, pn) representing ameasured pressure drop over the filter, and determining said optimal filter use time period (Topf) based on said filter hardware value, said filter use value and said measurement data.

2. The method as claimed in claim 1, further comprising: estimating a total resource consumption for a given time period (T)based on the filter hardware value and the filter use value, and determining the optimal use time period (Tom) substantially minimizing the total resource consumption.

3. The method as claimed in claim 2, wherein the total resourceconsumption for the given time period (T) is determined by: a first factor ( f(T) ), according to which the resource consumption isinversely proportional to the time period (T), and a second factor ( e(T) ), according to which the resource consumptionis directly proportional to the time period (T).

4. The method as claimed in claim 3, wherein the first factor ( f(T) )is determined as a product of the filter hardware value (Cfi/fef), an inverse ofthe use time period (T), of an uptime ratio (Rup), and optionally one or moreconstants.

5. The method as claimed in claim 3 or 4, wherein the second factor ( e(T) ) is determined as a product of the filter use value (Cuse), an 18airflow (Q) through the filter, an average pressure drop (pavg) over the filter, the time period (T), an uptime ratio (RUP) an inverse of a fan efficiency (fyfan),and optionally one or more constants.

6. The method as claimed in claim 5, wherein the averagepressure drop is determined based on at least two items of said measureddata points (tn, pn).

7. The method as claimed in claim 5 or 6, wherein the averagepressure drop is determined based on at least one estimated pressure dropdata point.

8. The method as claimed in claim 7, wherein the estimatedpressure drop has the form Pa(t) = startpêfebfl, where startpa is an initialpressure drop, b is an environment coefficient and tis the time.

9. The method as claimed in claim 8, wherein the initial pressuredrop (startpa) is determined as an average of at least two measured data (tn, PH)-

10. The method as claimed in claim 8 or 9, wherein the environmentcoefficient (b) is empirically determined.

11. The method as claimed in claim 8 or 9, wherein the environmentcoefficient (b) is determined based on a regression analysis matching at leasttwo measured data points (tn, pn) items to an exponential function.

12. The method as claimed in claim 8, wherein the environmentcoefficient (b) is determined according to b = LO log (ï-*U/tn, wherein to is0 an initial pressure drop, and wherein tn is the time of the n:th data point, p” isthe pressure drop of the n:th data point and N is the number of data points. 1913. The method as claimed in claim 8, wherein the environment

N*Ey=0 tn*pn_zg=o fn Ey=0 pn2N*Eg=o tizæ"(zy=o fn) wherein t” is the time of the n:th data point, p” is the pressure drop of the n:th coefficient (b) is determined according to b = data point and N is the number of data points.

14. The method as claimed in any one of the preceding claims, furthercomprising determining whether the pressure drop over the filter increasessubstantially linearly over time.

15. The method as claimed in claim 14, further comprising: determining coefficients of a linear equation based on at least some ofthe measurement data points, preferabiy all but the latest measurement datapoints, inserting a time of a latest measurement data point into the linearequaüon, comparing a thus estimated pressure drop with a correspondingmeasured pressure drop, if the difference between the estimated pressure drop and themeasured pressure drop is less than a predetermined threshold value,determining that the pressure drop increases substantially linearly over time,and if the difference between the estimated pressure drop and themeasured pressure drop is greater than the predetermined threshold value,determining that the pressure drop increases substantially exponentially over time.

16. The method as claimed in claim 14 or 15, wherein if it isdetermined that the pressure drop increases substantially exponentially, atleast two different environment coefficients (b) are estimated based ondifferent numbers of measurement data points, and the thus calculatedcoefficient (b) having the greatest value is used for determining the estimated pressure drop.

17. The method as claimed in any one of the preceding claims, wherein the resource is a cost.

18. The method as claimed in any one of claims 1-16, wherein theresource is a carbon dioxide impact.

19. The method as claimed in any one of the preceding claims, furthercomprising receiving a filter replacement value, representing an amount ofsaid resource associated with replacement of the filter, wherein the optimalfilter use time period (Topf) is determined also with respect to the filter replacement value.

20. The method as claimed in any one of the preceding claims, furthercomprising presenting the optimal filter use time period (Topf) in a user-readable format.

21. The method as claimed in any one of the preceding claims,wherein said at least one measured data (tn, pn) comprises at least one datapoint which is higher than an initial pressure drop and lower than a finalpressure drop of the filter, and which is at least about 5, preferably about 10,about 20 or about 30, days apart from both a time of the initial pressure dropand a time of the final pressure drop.

22. A method of optimizing filter changes in a plurality of filter locations,comprising: performing the method as claimed in any one of the preceding claimsfor each of at least two of the filters, said filters being of different types, and determining a replacement schedule for said at least two of the filtersbased on their respective optimal filter use time periods (Topf).

23. A system for monitoring a ventilation system, comprising: 21means for receiving a filter hardware value, representing an amount of a resource associated with at least production of the filter, means for receiving a filter use value, representing an amount of saidresource associated with use of the filter, measurement means for measuring a pressure drop over a filterarranged in said ventilation system, and processing means for determining an optimal filter use time period(Topf) based on said filter hardware value, said filter use value and said measurement data.

24. The system as claimed in claim 23, wherein the processing meanshas the form of a mobile hand-held device.