Scribd Logo
Upload

Welcome to Scribd!

  • US11579576B2

    Uploaded by

    muhammad.arslan
    7 views15 pages
    Patents

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    Patents

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    7 views15 pages

    US11579576B2

    Uploaded by

    muhammad.arslan
    Patents

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 15
    2) United States Patent Agrawal et al. (Sd) SYSTEMS AND METHODS FOR OPTIMIZING PERFORMANCE PARAMETERS OF AIR HANDLING UNITS, IN INFRASTRUCTURES (71) Applicant: Tata Consultaney Services Limited, Mumbai (IN) (72) Inventors: Supriya Agrawal, Pune (IN); Sagar ‘Verma, Pune (IN); Ramasubramanian Suriyanarayanan, Chennai (IN). Srinarayana Nagarathinam, Chonnsi (IN): Rajesh Jayaprakash, Chennai (IN): Venkatesh Ramanathan, Pune (IN); Amand Sivasubraman Chennai (IN) ‘Tata Consultaney Services Limited ‘Mumbai TN) (73) Assignee (#) Notice: Subject wo any disclaimer, the term of this tended of adjusted under 35 cays. (21) Appl. No. 160775089 (22) Filed: Jan. 28, 2020 ws) Prior Publication Data US 202010249645 Al Aug. 6, 2020 G0) Foreign Application Priority Data Jan. 31, 2019 GN) 21921003884 (1) Incr. Gose 19042 GO6N Suz (2006.01) (202801), GOSB 19042 (2013.01); GOBN 527 (2013.01); GaSe 2219/3614 (2013.01) USOI1S79576B2 US 11,579,576 B2 Feb. 14, 2023 (10) Patent No.: (4s) Date of Patent: (58) Field of Classification Search cre (GOSB 19/042: GOB 2219/2614; GOsB 2219/2642; GOSB. 15/02; GO6N 5/027 ‘ce application file for complete search history 60) References Cited US. PATENT DOCUMENTS. 2000246685 AL 1/2005. Seshad &a aoig0rri4ed AL 102012 Dempster eta ao1nu0}s70 AL |22013. Songeeal suigun724se Al* 122018 ae cosa 2902 POREIGN PATENT DOCUMENTS os 22018 * cited by examiner Primary Examiner — Michael W Choi (74) Attorney, Agent, or Firm — Finnegan, Henderson, Farabow, Garett & Dunner, LLP on ABSTRACT Subssystems of air handling units in infrastructures face ‘unresolved problem of eontlct in the rules that activate in a conirlictory manner at the same time resulting in sub- ‘optimal performance ofthe subsystems. The present dsclo- sre provides a system and method for optimizing perfor ‘mance parameters of air handing units in infrastructures. Ril sets having conflicting conditions are identified ater verification of ries which are specific to air handling units. Furr, frequency of the rule ses having conflicting con: ditions is determined to generate a ranked list ofthe rule sets having conflicting conditions. Another ranking procedure is ‘implemented forthe rales comprised in the ranked list of the rile sets having conflicting conditions. The system dynam cally optimizes one or more parameters specific 10 the performance criteria hased on the ranking of rules 6 Claims, 4 Drawing Sheets AC temperature] | Static and || Control Rules ee i Variables Test Case Conflicting Generation] | Conficting| Framework) | Cpp Funetion Energy Model | } Python Actor Patent Feb. 14,2023 Sheet 1 of 4 US 11,579,576 B2 Device 100 Processor 106 VO interface 104 Memory 102 Optimization module 108 Data repository 110 FIG.1 U.S. Patent Feb. 14, 2023 Sheet 2 of 4 US 11,579,576 B2 Receiving, via one or more hardware processors, a plurality of | 202 rules comprising one or more static and dynamic variables of one or more operations of at least one subsystem ‘Verifying, using a test case generation framework, the received | 204 plurality of rules comprising one or more static and dynamic variables of the one or more operations to identify one or more tule sets having conflicting conditions 206 Determining using an integrated framework, frequency of the one or more rule sets having conflicting conditions, based on the frequency of the one or more rule sets | 208 having conflicting conditions, a ranked list of the one or more rule sets having conflicting conditions 210 Identifying, one or more rules based on a selection of at least one rule set from the ranked list of the one or more rule sets and a performance criteria Dynamically optimizing, one or more parameters specific to | = 212 the performance criteria based on the identified one or more rules FIG.2 U.S. Patent Feb. 14, 2023 Sheet 3 of 4 US 11,579,576 B2 Static and, Test Case JAC temperature! | “Dynamic || Generation) | Costing ‘ontrol Rules! | Variables Framework ae Cp Function Python Actor [Energy Model FIG.3 U.S. Patent Feb. 14, 2023 Sheet 4 of 4 US 11,579,576 B2 US 11,579,576 B2 1 SYSTEMS AND METHODS FOR OPTIMIZING PERFORMANCE PARAMETERS OF AIR HANDLING UNITS. IN INFRASTRUCTURES. PRIORITY CLAIM This US. patent application claims priority under 35 US. $119 to: India Application No. 201921008884, fied ‘on Jan, 31, 2019, The entire coatents of the aforementioned pplication are incorporated herein by reference. ‘TECHNICAL FIELD ‘The disclosure herein penerally relates to Air Handling Units (AHUs) in infrastrucnes, and, more particularly, 10 systems and methods for optimizing performance param- ‘ters of AUS in infrastrmctures BACKGROUND Air handling units (AHUs) ae integral part of infiastc- ‘ure ecosystem whieh provide an intelligent mectnism for keeping building ambienUcondition(s) within a specified range based on an occupancy schedule, monitoring perfor mance and device fulures in all systems and providing malfunetion alarms. During peak usage times, redvetion in building energy helps in reducing the energy costs of oper- ating a building’s heating, ventilation, and air conditioning (HVAC), lighting, levator, and other subsystems. Most, Air handling units include 2 control layer as a supervisory layer that manages multiple underlying sub-systems suc as HVAC sub-systems and acaess control sub-systems. Imple- mentation oflogi and intelligence bebind supervisory con- teol layer i expressed in the form ofa sequence of “if then* res, Which leverage data from multiple systems for imple- mentation. These rules are mostly given by operators. The operators, despite of having operational knowledge and an intuitive oF heuristic understanding ofthe system behavior, may not be ‘expert in the respective domains. Funer. there can be possibilities of confit in the rules (two rules tht activate the devices in a contradictory manner at the same time) as the rules may be provided by multiple stakeholders. Hence. there may be possibility of constraints violation with respect to the oocupants’ comfort and adverse effet on the equip- ‘ent die 19 too-feuent duty eyeling or adjustment of the devices, SUMMARY, Embodiments of the present disclosure present techno logical improvements a8 solutions Io one or more of the above-mentioned technical problems recognized by the inventors in conventional systems. For example, in one ‘embodiment, thore is provided a processor implemented method, comprising receiving, via one oF more hardware processors, a plurality of rules comprising one or more static ‘and dynamic Variables of one or more operations of atleast ‘one subsystem; verifying, using a test ease generation Jramework, the roeived plurality of ules comprising one of more static and dynamic variables of the one ot more ‘operation to identify one or more rue sets having confliet- ing conditions; determining, using an integrated framework, Frequency’ of the one oF more rule sets having contlicting ‘conditions generating, based on the frequency af the one oF ‘more rule sets having conflicting conditions, a ranked lst of 0 o 2 the one of more rule sets having conlliet ‘dentfying, one or more rules based on a selection ofat least ‘one rule se om the ranked list ofthe ane ar more re sets anda perfomance eritera. In an embodiment, the selection ‘of at least one ruleset rom the raked list oF the one or more rile ses is based on a highest coniict frequency criteria. In ‘an embodiment, the method further comprises dynamically ‘optimizing, one or more parameters spovifie to the perfor ‘mance ertria based on atleast one of (the identified one ‘or more rule. In an embodiment, the one oF more param- ters specific to the performance criteria comprises at least ‘one of () a consumed energy and (i) a discomfort level Tn another embodiment, there is provided «system com- prising: a memory storing instructions; one or more com- ‘iinicaton interfaces; and ane or more hardware processors ‘coupled to the memory through the one or more communi- cation interfaces, wherein the ane or more handware pro- cessors are configured by the instretions to receive, a plurality of rales comprising one or more state and dyna variables of one or more operations of atleast one subsys- tem; verify, using a test ease generation framework, the received plurality of roles compising one or more static an ‘dynamic Variables ofthe one of more operations to identify fone of more rule sets having conilicting conditions; date ine, using an integrated framework, frequency of the one for more mule sets having conflicting conditions; generate, based on the frequency ofthe one oF more rule ses having conilcting canditions, a ranked list of the one or more rule sets having conflieting conditions; identify, one or more niles based on a selection of atleast one rile set from th ranked list of the one or more rile seis and a performance criteria. In an embodiment, the selection ofa least one rule ‘set fom the ranked list ofthe one or more rule sts is based fm a highest conflict frequency erteria, In an embodiment the one or more hardware processors are further configured by the instructions to dynamically optimize, one or more parameters specific tothe performance erteria based on at Teast one of (i) the identified one oF more rules. In an embodiment, the one or more parameters specific t0 the performance eriteria comprises at least one of () consumed energy’ and (i) discomfort level. Tn yet another embodiment, there are provided one or ‘more non-iransitory machine readable information storage ‘mediums comprising one or more instructions which when executed by ane or more hardware processors catsereceiv~ ing via one or more hardware processors, a plurality of ules ‘comprising one or more statie and dynamie variables of one for more operations of at least one subsystem; veriying, sing. test ease generation framework, the received plural= ity of miles comprising one of more static and dynamic variables of the one or more operations to identify one or more rue sets having eonfiting conditions: determining, ‘using an intograte framework, fraqueney ofthe one or more rile sets having conflicting conditions; generating, based on the frequency of the one or more rile sets having conflicting ‘coneitions, a ranked list of the one of more re ses having conflicting conditions; identifying, one or more rules based ‘na selection of atleast one rule set from the ranked list of the one or more rule sels and performance ertri. la an embodiment, the selection of at feast one rue set fom the ranked lst ofthe one or more rle ses is based on a highest conlliet frequeney erteria. In an embodiment, the instue- tions may further cause dynamically optimizing, one or ‘more parameters specifi to the performance criteria based fo at least one of () the identified one oF more rules. In an embodiment, the one or more parameters specific to the US 11,579,576 B2 3 performance criteria comprises at least one of () a com: sumed energy and (i) a discomfort level Its to be understood that both the foregoing general ‘description and the following detailed description are exem- plary und explanatory only and are not restrictive of the ‘vention, 88 claimed, BRIBE DESCRIPTION OF THE DRAWINGS ‘The accompanying drawings, which are inconporated in ‘and constitute a par of this disclosure, illustrate exemplary ‘embodiments and, together with the description, serve t0 ‘explain the disclosed principles TG, 1 illustrates a funcional block diagram ofa system for optimizing performance parameters of Air Handling Units (AHUs) in infrastructures, according to some embod ments of the present disclosure IG. 2 is an exemplary flow diagram of a processor Jmplemented method for optimizing performance param- ‘ters of the AITUs in infstrictures, in accondance with some embodiments of the present diselosure IG, 3 illustrate a high level architecture of an integrated ‘ramework inthe system for optimizing performance param- ‘eters of air handling units in infrastructures, according t0 some embodiments of the present dislosure FIG, 4 illustrates a simulation path for mule Ripping by picking the coatlicting rules with highest froqneney 10 ‘achieve a balance between energy and comfort, aecording (0 some embodiments of the present diselosure. Itshould be appreciated by those skilled in he at that any block diagrams herein represent conceptual views of ilus- trative systems and devices embodying the principles of the present subject mater. Similarly, it will be appreciated that ‘any ow chars, Mow diagrams, and the lke represent various processes which may be substantially represented in ‘computer readable medium and so executed by a computer ‘or processor, whether o¢ not such computer oF processor is ‘explicitly shown, DETAILED DESCRIPTION Exemplary embodiments are described with reference to the secompanying drawings. In the figures, the lft-most ‘digit(s) ofa reference number identifies the figure in which the reference number fist appears. Wherever convenient, the same reference numbers are used throughout the dravs= ings to refer to the same or like pars. While examples and Jeatures of disclosed principles are described herein, mod Fieations, adeplations, and other implementations are pos- sible without departing from the spirit and scope of the disclosed embodiments, It is intended that the following ‘detailed description be considered as exemplary only, with the tnve scope and spirit being indicated by the following claims. ‘The embodiments herein provide systems and methods for optimizing performance parameters of Air Handling Units in infrastructures. The typical interpretation of results ‘obtained from conventional building automation systems ths sen mf vole a probim of uly eo reduction where «formal specification based verification ‘automation resis performed followed by a rules prior zation mechanism for optimizing parameters affecting per formance of AHUs. The rules prostization mechanism is performed based on ranking ofa plurality of rae sets having ‘conflicting conditions and ranking of rules comprised inthe ranked rule ses having conflicting conditions. The ranking ‘fre sets having conflicting conditions is performed based 0 4 on thee frequency. Further, based on rule prioritization, one ‘or more parumeters specific to the performance criteria such as consumed energy ankl discomiort level, and the like are ‘Synamically optimized, Referring now to the drawings, and more particlaely t0 PIGS. 1 through 4, where similar reference characters denote corresponding features consistently throughout the figures, there are shown prefered emboximents and these embodiments are deseribed in the context ofthe following ‘exemplary system andlor method. FIG, 1 illustrates functional block diggram ofa system or optimizing performance parameters of ai handling units in infrastruc, aeeording t© some embodiments of the present disclosure. The system 100 includes or is otherwise Jn communication with one oF more hardware processors such as a processor 196, an VO interface 104, atleast one ‘memory stich asa memory 102, and an optimization module 108. In an embodiment, the optimization 108 can be imple- ‘mented as a standalone unit in the system 100. In another embodiment, the optimization module 108 can be imple- ‘mented as a module in the memory 102. The processor 106, the 10 interface 104, and the memory 102, may be coupled by a system bus The VO interlace 104 may include a variety of sofware ‘and hardware interfaces, for example, a web interface, a raphical user interface, and the ike. The interfaces 104 may include a variety of software and ardware interfaces, for ‘example, interfaces for peripheral devices), such 2s a key- ‘board, a mouse, an external memory, a camera device, and printer. The interfaces 104 can faite multiple coma fications within a wide variety of networks and protocol types, including wired networks, for example, local area setwork (LAN), cable, ct, and wireless networks, such as Wireless LAN (WLAN), cellular, or satellite. For the pur- pose, the interfaces 104 may include one oF more ports for ‘connecting numberof computing systems with one another fof to another server computer. The TO interface 104 may include one or more ports for connecting a number of devices to one another or to anther server. The hardware processor 106 may be implemented as one for more microprocessors, microcomputers, microcon- troller, digital signal processors, central processing unit, state machines, logic cicuiries, andlor any devices that ‘manipalste signals based on "operational instructions Among other capabilites, the hardware processor 106 is ‘configured to fetch and execute computerreadable instrve- ‘ions stored in the memory 102, ‘The memory 102 may inchule any computerreadable ‘medium knowa in the art including, for example, volatile ‘memory, sich as static random access memory (SRAM) and dynamic random access memory (DRAM), andor non- volatile memory, such as read only memory (ROM), eas- able programmable ROM, flash memories, hard disks, opt cal disks, and magnetic tapes. In an embodiment, the ‘memory 102 includes an optimization module 108 and a repository 110 for storing data processed, received, and zenensted by the optimization module 108. The optimization ‘module 108 may include routines, programs, objets, com poneats, data structures, and so on, Which perform particular tasks or implement particular abstract datatypes. “The data repository 110, amongst other things, inludes 2 system database and other data, The ther data may include data generated as a result of the execution of one or more ‘modules in the optimization module 108, The system data- base stores one or more static and dynamie variables includ- ‘ng input and ontpt variales. and performance parameters ‘are generated as a result of the execution of one or US 11,579,576 B2 5 ‘more modules in the optimization module 108. The data stored in system database can be leant to further provide ‘increased number of intelligent rules. In an embodiment, the optimization module 108 can be ‘configured to oplimize performance parameters of ir han- ding units in infrastrctures. Optimizing performance Parameters of air handling units in infrastructures can be ‘carted out by using methodology, described in conjunction With FIG. 2 and use case examples FIG, 2, with reference to FIG. 1, is an exemplary flow 0 and RAT is desired temperature (desTemp), set Flowrate to MAX. Rule 2 I OpCht is <10, set Flowrate to Zero. TABLE 2 S80_Ofemr_RAT Currin Guntime __RAT Table 2 specifies the above requirements using EDT, in which each row maps directly to one of the requirements ‘The column headers specify three input signals: OpCat, RAT and CurfTime, and three output signals: CurfTime. RATand Flowrate. tis noted that CureTime and RAT occurs jn both input and output signals, thus called local or UO signals. All input cells ae connected through Logical pattern ‘expressions where pattem expressions in eacl input cell, speci the input valuo(s) that will match the requirements of that coll TABLE 3 “Tine MIN) apa Sima Ouput Sami Remy o oeare Gettin 1 Row ouput wees Row oupe : Fowisle=-MAX Row Soup Curtin =2 Row Loupe Ween Rov loupe Howie MAX Row Dou Fawkes. Row Sauna 12 be seen from Table 3, for the default values of desired temperature and occupancy count assumed to be 22 ‘and 0 respectively, the retum air temperature (RTE) is calculated by the fiction ealeRAT( ) shen row 1 of Table 2 matches. At time 2, value of occupancy count (OpCat) is 5 that matches bei rows 2 and 3 of Table 2, Also RTE is ‘calculated as 24:5, which is less than desired temperature Second inpit of row 3 of Table 2 also matebes, Since all inputs of rows 2 and 3 of Table 2 matches, both rows enerate diffrent outputs at the same time, thus the eon- Slicting condition is identified. In another embodiment, the rue sets having conflieting conditions are identified whe: 8 vore than one rules are fired at the same time ad that may’ Tead to set the How mite variable to diferent values, Por ‘example, a subset of rules fred at the same time are shown below Ri: [Occupancy Count(OC)-0 for 2 hours, set flow rate to 1 Retm Air Temperature(RAT) Threshold3&&——Threshold2 and current time is either END_TIME, set flow rate to five 3: IFRAT is desired_temperature (DS)+TEMP_TOLER- ANCE, set flow rate to 10, Ré: IFeecupancy count~0 and ambient emp (AT)>RAT, set Aw rate to 20. RS: IF RAT END TIME 33:10 RAT ie > dsied_tenpernze + TEMP_TOLERANCE stow 141A is > Theol? R& Onl nent ep (AT) > RAT at at to 1: IRAT ig > Thesol2 RA <= Threshold ued 1: {f motondteson (MD = 0 fo 1 Bou Bo at fo 216 [IDs IF RAT < DS end OC > sete How it 5, US 11,579,576 B2 u TARL Es} =Tivesols Tio 4 “Tel dee Treetol 5 tos rc * oft Bi) It ps Outputs of the 12 rules specified in Table Sare as provided in Table 7 showa below: ABLE 7 Ti, Re an RO Roan R12 i RI at RS Table 8 provides one or miore rules sets which are iden- fled as the rule sets having eoniieting conditions. TABLE 8 Rulcet Rule wt Conti Freensy re As can be seen from Table 8, four rule sets are identified asthe rues sts having coniicting conditions since the rules in corresponding rule sets are fired atthe same time and setting output (eg, flow rate) to different valves. Further, it is depicted in Table $ that rule set 1 (RI, R2 and R3) cconfics 100 number of times, rule set 2 (R4 and RS) conticts 80 times, rule set 3 (R6, R7, and RB) coniiets 65 times, and rule set 4 (RO and RIO) conflicts 52 times. So, based on the frequency of occurrence of eoafiet(alternae tively refered as confit frequency seore) among rules in the coresponding rule ses, a ranked lst of rules sets having ‘conflicting conditions is generated. An example of the ranked list ofthe rules sets having conflicting conditions is shown in Table 8. Further, as depicted in step 210 of FIG. 2, the one or more hardware processors (106) are configured to ident, one oF ‘more niles based on a selection of at Teast one re st Irom the ranked ist ofthe one oF mare rule sets anda performance ‘riteria. In an embodiment, the selection of atleast one cule (staRT_TIME END-TIME} SSCART_TIME > ENDLTIME 0 o 12 2» is so fom the ranked list ofthe one or more rule sts i based ‘ona highest conflict frequency criteria. Inother words, a rule set with highest conflict frequency score is selected from the ranked list of the rules sets having conflicting conditions. For example, as shown in Table 8 rule sot 1 (R1, R2, and 3) has the highest conflict frequency score of 100. Thus, rile st | is selected from Table 8 showing an example ofthe ranked list of the rules sets having conflicting conditions. Tuner, one or more rales are sclected from all rules ‘comprised inthe selected rule set (eg, rule set 1), Selection fof ane or more rules comprised in he selected rule set (eg. rule st 1) is done based on generation of another ranked list. Another runked list is generated by providing « ninking ‘score to each rule comprised in the selected rile set (62. rile set 1), TABLE 9 Por example, as depicted in Table 9, the making score provided to RI is 30, R2 is 35 and R3 is 25. Thus, R2 is Selected from the another ranked ist a depicted in Table 9. Here, the runking score is provided based on a performance criteria, wherein a mapping between each rule comprised in the selected rule set (ex, rule set 1) is performed with one ‘oF more parameters specific to the performance criteria In fan embodiment, the one oF more parameters specific to the performance criteria comprises at least ane of () a eon- ‘sumed energy and (i) a discomfort level. The rules showing better performance are assigned lamer score. Here, the performance eriteria for rule is defined as reduction in the value of consumed energy while maintaining discomfort level of occupants with ina predefined permissible range for ext T steps, Here, the permissible range could he bist not Jimited w 0 wo 15. Ia sn embodiment, the ranking seore pets ‘updated after each iteration, In each iteration, performance US 11,579,576 B2 13 of a seleciod rule is observed based on the values of ‘consumed energy and discomfort level for X steps afer the seloted rule is fred, The ranking score is disetly propor- ‘ional to consumed energy and ution in discomfort level Based om the umber by which the energy and discomfort Fevel increases or decrease, the ratking seore is update. The raking score of all rues in previous iteration is used ia next iterstion and also gets uplated afer the next iteration is Tn another embodiment, more than one rule sets are selected fom the ranked ist ofthe one or more nue sts In ‘an embostimeat, all rule sets fom the ranked list ofthe one ‘or more rule sets could also be selected. Here. some of the selected rules ean have high confit frequency scores, while some other rules can have medium conic lrequency seores snd remaining can have low confit Krequency scores. Por ‘each role set, 4 machine learning, meckanism is applied to Jeam score by making each rae of the selected ruleset a the deciding rule oa some dataset and monitoring the change ia, ‘energy consumed and diseomfor level. Thus, based on, ‘monitoring the change in energy consumed and discomfort level, a relative score is provided foreach rule inal the rule sets comprised inthe ranked list ofthe one or more rule sets Por example, itis assumed that 2 rule sets (RI, R2 and R3) ‘and (R4 and RS) are selected. Now, based on the perfor- mance criteria, the impact of each rue is analyzed when a ‘conflict i there in these data sets, If conflict arise in rle set (RI R2 and R3), then a rule (RI or R3 or R3) is selected with some probability and impact of selected rule is mea- sured. Ths impact of the chosen rule will be measured as a Junetion of energy consumed and discomfort level for next stops. Inn embodiment, for those rules sets from the ientitied ‘one or more rule ses that compre (Wo rules only, a rule Slipping mechanism is performed. Rule Ripping mechanism refers to a process of changing the soquence of rules ia Expressive decision table (EDT) when a eonfliet occurs Generally, execution in EDT follows a top-to-bottom (oF bottom to top in some scenarios) approach due to which the lust rule overwrites the previous ale, However, the rule flipping mechanism proposed in the present disclosure as implemented by the system 100 is modified based on ‘mapping the rules withthe performance entra, In case of ‘arule set comprising only a pur of rules (wo rules only, the rule providing better performance is flipped with another rule irrespective of whether the other rule is lst rule which js normally given priority in traditional system. For ‘example, Table 3 depicting execution in Expressive decision ‘able in traditional system shows that last value (output of row 2) overwrites the previous value (output of row 3) for flow rate. Hence, the final value taken for Further processing Js 0. However, n preset disclosure the value of flow rate for row 2 and row 3 ate mapped with energy value and discomfort level. Based om mapping, it is analyzed that ‘exocution of row 3 provides better performance than row 2. Thins, row 2 and row 3 are interchanged and row 2 of previous EDT (shown in Table 3) becomes row 3 of new EDT (shown in Table 10). Itean be scen from Table 10 that the value of flow rate is changed due to rule flipping. The flipped rues are further used for next executions. 14 TABLE 10 Tae in) pws —_Oupu | Rents © Oem) 1 opca=0 atime ow up Rares Row loupe FowRite = MAK Row 2 opt Punter, as depicted in FIG. 2, at step 212, the one or more hardware processors (106) are configured to dynamically ‘optimize, one oF more parameters specific to the perfor ‘mance eriteria based onthe identified one or more rules. The identified one or more rules are the rules having highest ranking seore which are dynamically identified, in one ‘example embodiment, Further, ranking of the rues is per ormed based on mapping with the one or more parameters (Gnthis case, consumed energy and discomfort level) specific to the performance criteria. Thus, hased on dynamic iden- ‘ilfcation of rules having high nk from the ranked list ofthe ‘one of more male sets, the performance ofthe system 100 (or AHIUs under consideration) and the one o more parameters specific to the performance eriteria are dynamically opti nized, Experimental Results Te present disclosure utilizes an experimental set up ‘where experiments are conducted using the BCVTB co- simulation framework ona real-world office seup. A list of all possible rules from a facility HVAC manager is obtained, ‘specifically 27 rules that woud st the fla rate 10 0,4 2 Yor | of MAX (rated AHU flow rate). These 27 rules were converted to EDT rows. Further, experiments ae performed ‘ona real ine daa set with an integrated setup of EDT-Tool ‘and the building energy model. The historical logs of the input parameters to EDT framework such as OpCat and ambientTemp are obtained for 90 days in the form of test ceases. Approximately 40 unigue conflits per test case are identified, Also, with the variety in the input profile, the rules that conflicted varied aeross time and test case, To ‘analyze the impact of these confit, the rule sequences in EDT are flipped for further experiments. The rule pairs conllicting with the highest frequency are selected. and flipped for the next execution. Now, the test cases afer execution 2 of flipped rls are diferent from test cases of ‘execution 1 and they had different confit, despite having same input values. Again, the rules conflicting the highest ‘numberof times for execttion two are selected and flipped for next execution. Flipping has been executed for a depth of wo, In an embodiment, the present disclosure utilizes the ota ‘consumed energy (kWh) and time-average PPD metric (6) ‘during the occupied hours aeross the 90 days as comparison metrics. The discomfort level is measured in terms of thermal comfort whieh is further quantified using Fanger's predicted percentage dissatisfied (PPD) metric. It ix noted that PPD<15% is recommended for thermal comfort. It is ‘observed thatthe thermal comfort of occupant is within the predefined/pemnssible limits for the entire duration of the ‘Working hours (0900-1800 Hour). US 11,579,576 B2 15 Table 11 provides results of the Benchmark set of rules TABLE 11 S.No Siamision 1 Flot Deh ery evel F 7 ar yy 4 24 tani 3 “4009 “The present disclosure i based on some standard assummp= tions, wherein frst standard assumption is dhat energy is ‘directly proponional to low mite and. second standard assumption is that occupants discomfort level will decrease 2S Tetras = RAT ‘when more energy is consumed. Hence simulation was ran for the system described in the present disclosure, where flow rate with minimum values is prefered among. all ‘conflicting rules for any ine insta. This is easily achieved, by sorting the rule set in decreasing order of Now rate and eeding to EDT-Tool, Table 11 shows comesponding con- sumed energy and discomfort level in second simulation. Based on a comparison of proposed disclosure with Bench- ‘ark set of ries, it is observes that the consume energy has “dropped by. 5.38% while discomfort level has gone up 10 17.5% which is notin the prefered mange. It holds above stated standard assumptions but at same time, also indicates that trading off between consumed energy and occupants ‘comfort level is a hard problem, In onder t0 achieve a balance betwoen energy and comfor, differnt paths for rule flipping are opted by picking the conflicting rules with highest frequency. A path is defined as a sequence of directed edges which connect a sequence of simulations (states). These edges will store the pur of rules flipped for ‘particular transition. The relation between sequences of states have been represented using sate identities (IDs). FIG. 4 illustrates «simulation path for rule flipping. by Picking the conflicting rules with highest taquency to achieve a balance between energy and comfort, aecording to some embodiments ofthe present disclosure. In FIG. 4, 5, and S, represent three simulation instances ofa path with simulation 1D 0, 1.1 and 1.2. S, was attained by flipping 2 pair of rues (1, 2) from benchmark. Rules (2, 3) from S, preference rle lst are lipped again to asin S.. Depth af & Sates defined as numberof edges between the root (Bench- mark) and current state. Initially for benchmark simulation the top three most confiting pir of rls are chosen one by ‘one and simulation is w-run by giving preference to one rule ‘over other through flipping. Flipping of rules results in three new preference orders forthe set of rules. Similarly, sima- Jation for the secon depth is executed which means tat simulation is executed for new preferred onder for rules. ‘While going for third depth flip, it is observed that the new preference orders are already covered in previous simulu- tions. Hence, there is no need to go for third. depth of simulation “Table 12 shows the st of rales Bipped during simelatons ABLE 12 s ss stntally. In simulation to, it is observed that appro 0 16 ‘TABLE 12-continued SUF RAT > Tlrsld and dot Wrkingious = Curae = nol Mockisturs THEN FlowRate = NEG STP Opcan sb sat RAT = anbntenp THEN RowRate » MAX heats THEN Few ate = MAK 8.1 Opeat™0 and RAT = drfenp THEN Hw tate= MAR Coverage of all aes through EDT for the above simu Jations ar also measured. For the benchmark preference rule list its found that one rule (eg, “IT RAT is (22.23}, st flow rate 10 0°) is not covered throughout the 90 days It is aso jobserved that attempting the coverage of this rule by gen- crating test cases purely through EDT is also not success indicating that there my he some ambiguous nile priority that is responsible for the lack of coverage. The execution etals are analyze to gain insight ofthe reason for uneov- temility It is observed that thre of the rales are trying t9 in RAT around 24 with a moderate flow rate of S hile rule 4 in Table 12 is tying to set Bow rate to MAX. Since priority of rule 4 is lower than at least one of these three rules, resultant flow rate is always 5 in ease RAT become 24, So, RAT never drops below 23. Fence, when priority of rule four is set higher than ssid three rales, it results in postive rate of decrease in RAT leading RAT to drop below 25 and finally covering the rein simlation 5 In Table 1 the impact of conflicts on consumed energy and discomfart level is shown. Due to the flips, comfort tem- perature and consumes! energy of the system changed sub- mately 68% reduction in consumed energy whi discomfort level went up by 2.1% when compared with the ‘benchmark. Further, for simulation five, its observed that consimed enemy goes up by 26% while discomfort level is approximately same. Thus, the best trado-ofl between fenergy and occupants comfort is achieved for simulation three. ‘The written description deseribes the subject matter hersin to enable any person skilled inthe art to make and use the embodiments. The soope ofthe subject matter embod- ‘ments is defined by the claims and may inchide other modifications that oocur those skilled in the at, Such ‘other modifications are intended o be within the scope ofthe claims if they have similar elements that do not differ fro the literal language ofthe claims or if they inelude equiva- lent elements with insubstantial differences from the iveral language of the claims. Tis to be understood thatthe scope of the protection is extended to such @ program and in addition to @ computer readable means having # mestage therein; such computer readable storage means contain program-code means for ementation of one or more steps of the method, when US 11,579,576 B2 17 the program runs on a server or mobile device or any suitable programmable device. The hardware device can be ‘any’ kind of device which can be programaied including ny kind of computer like « server or a personal compiler, ‘or the like, or any combination thereof, The device may also Jnchide means which could be eg. hantware means like eg. an application-spectic integrated cireuit (ASIC), a feld- Progrimmable gate amy (FPGA), or & combination of hardware and software means, eg. an ASIC and an FPGA, ‘ort least one microprocessor and at Jest one memory with software modules located therein, Thus, the means can include both hardware means and software means, The method embodiments described herein could be imple- mented in handware and software, The device may also include software means. Alternatively, the embodiments may be implemented on different hardware deviees, eg using a plorality of CPUs. The embodiments hercin can comprise hardware and software elements. The embodiments that are implemented in software include but are aot Hmited to, fiver, resident 2 software, microcode, ete. The functions performed by vari- ‘ous modules described herein may be implemented in other ‘modules ar combinations of other meduls, For the purposes ‘of this description, a computer-usable or computer readable medium ean be any apparatus that can comprise, store, ‘communicate, propagate, or transport the program for use by ‘or in connection with the instruction execution system, apparatus, or device “The illustrated steps are set out to explain the exemplary ‘embovdiments shown, and it shouldbe anticipated that ongo™ ing technological development will change the manner in which particular funetions are performed. These examples are presented herein for purposes of illustration, and not Timitation, Furler, the boundaries ofthe fanetional building blocks have boon arbitrarily defined herein for the conver rience of the description. erative boundaries can be fined so long asthe specified functions and relationships thereof are appropriately performed, Altematives (including ‘equivalents, extensions, variations, deviations, etc of those

    You might also like