JP6440525B2 - Equipment performance diagnostic apparatus and equipment performance diagnostic method - Google Patents

Description

  The present invention relates to a technique for diagnosing equipment performance.

  A method for diagnosing deterioration of equipment performance in monitoring of building equipment is known.

  Patent Document 1 describes that air conditioning equipment plots data on two axes of primary side received integrated heat quantity and secondary side supplied integrated heat quantity, and diagnoses deterioration by comparing the data. Here, it is assumed that an integrator is installed. When there is no accumulator, “If it is not accumulated data, it is converted to accumulated data in minutes, hours, days, and months in order to reduce the processing speed when creating the graph. “It is preferable to memorize”.

  In Patent Document 2, for an abnormality in the operation monitoring of a plant such as a sewage treatment facility, a combination of data is determined in advance from measurement data, and then mathematical and statistical modeling is performed. Originally, an example in which an abnormality is diagnosed using a threshold value is described. For example, data is modeled in the form of a principal component using a part of process variables, and a Q statistic is calculated using a score matrix that represents a principal component score for the measured data and a loading matrix that represents a coefficient of the principal component. It is described that a T2 statistic of Hotelling is obtained and an abnormality is detected from this value.

JP 2006-318405 A JP 2013-061853 A

  The totalizer cannot always be used for all data. In addition, even if instantaneous data is accumulated, if the measurement cycle of instantaneous data is a long time interval such as one hour, it only captures the momentary state in one hour, and air conditioning and mechanical equipment in the meantime It is not always possible to obtain data highly correlated with the power consumed or the fuel consumed by the operation. In particular, when evaluating the efficiency of equipment, the energy input to the equipment is mainly electric power, and an integrator is often used. On the other hand, the process output related to heat and work, which is the output of the equipment, is instantaneous data. There are many cases. For this reason, since it is necessary to calculate by combining these data in the evaluation of efficiency, the data measurement method and the measurement cycle greatly affect the results.

  When modeling as described above, it is necessary to prepare and learn clean data in advance offline. In the distribution assumed from the data learned in advance, if the value measured online is a value that occurs only with a very low probability, it is not considered normal, because it is the concept of statistical testing. is there. For this reason, when the data learned in advance is noisy, only a large change exceeding it can be detected from the values measured online.

  As described above, when the accumulated data cannot be used in some cases, or when the noise generated by the mixture of the accumulated data and the instantaneous data has a large fluctuation range with respect to the data indicating the performance of the facility, the deterioration of the performance can be detected. I can't.

  In order to solve the above problems, an equipment performance diagnosis apparatus according to an aspect of the present invention includes a storage device, a communication device connected to a communication network, the storage device and a processing device connected to the communication device, Is provided. The communication device acquires a plurality of measurement data respectively measured at a plurality of time points within a monitoring period by a sensor provided in the facility via the communication network. The processing device calculates an input equivalent amount indicating the amount of energy input to the facility based on each measurement data, and corresponds to an output indicating the amount of energy supplied from the facility based on each measurement data Calculating an amount, calculating a load-related amount related to the load of the facility based on each measurement data, and relating efficiency related to the efficiency of the facility based on the input equivalent amount and the supply equivalent amount corresponding to each measurement data Calculating a quantity, calculating a feature quantity indicating a relationship between the load-related quantity and the efficiency-related quantity corresponding to the plurality of time points, storing the calculated feature quantity in the storage device, and obtaining a preset value Based on the condition, a past feature quantity that is a feature quantity of a past period is acquired, and the calculated feature quantity is compared with the past feature quantity.

  A change in the performance of the facility can be detected from the measurement data in which the integrated value and the instantaneous value are mixed.

The structure of an equipment performance diagnostic device is shown. Indicates map data. The characteristic curve extraction process is shown. Feature vector information is shown. A comparison screen is shown. Indicates diagnostic processing. The input screen of the variable allocation definition data 114 is shown. The map data of Example 2 are shown. The feature vector of Example 3 is shown. The characteristic curve extraction process of Example 5 is shown. Deterioration factor determination processing is shown. Shows the map screen.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 shows the configuration of an equipment performance diagnosis apparatus.

  The target facility 101 performs heat supply and work using supply energy such as electricity, gas, and fuel.

  The collection unit 103 acquires measurement data from the diagnosis target equipment 101 using the sensor 102, aggregates the data, and transmits the measurement data to the equipment performance diagnosis apparatus 110 through the communication network 105. A set of the target equipment 101 is called a target system. The information processing device 104 is connected to the communication network 105, stores other information when the target facility 101 is in operation, and transmits it to the facility performance diagnostic device 110.

  The facility performance diagnosis apparatus 110 of this embodiment includes an input energy equivalent amount calculation unit 112, a supply energy equivalent amount calculation unit 113, a variable assignment definition data 114, an efficiency mapping unit 115, a characteristic curve extraction unit 116, and a feature. A vector DB (database) 117, a characteristic change detecting means 118, and an output device 119 are included.

  The input energy equivalent amount calculation means 112 converts the collected measurement data into amounts related to the input energy amount (input energy equivalent amount, input equivalent amount). The delivery energy equivalent amount calculation means 113 converts the collected measurement data into amounts related to the delivery energy amount (delivery energy equivalent amount, delivery equivalent amount). The efficiency mapping unit 115 calculates an amount related to efficiency (equivalent amount after conversion, efficiency-related amount) from these amounts, and an amount related to loads such as outside air enthalpy (equivalent amount for mapping load, load-related amount). Map to space in association with. Based on this result, the characteristic curve extraction means 116 extracts a feature quantity representing the relationship between the load-related quantity and the efficiency-related quantity as a feature vector, and stores this result in the feature vector DB 117. The characteristic change detecting means 118 compares the feature vector stored in the past with the newly calculated feature vector, and detects the change. The output device 119 displays the feature vector and the measurement data based on the feature vector in a space of a quantity related to the load and a quantity related to the efficiency. The output device 119 may be a display device. The output device 119 may be provided outside the equipment performance diagnostic device 110, or may be connected to the equipment performance diagnostic device 110 via a communication network.

  For example, if the sensor 102 is an integrated watt-hour meter, the input energy equivalent amount calculation means 112 calculates the difference between the integrated power amounts in a preset sampling period as the input energy equivalent amount. When the measurement data by the sensor 102 is fuel, the input energy equivalent amount calculation means 112 may use the consumption amount within the sampling period as the input energy equivalent amount. When the measurement data by the sensor 102 is exhaust heat, the input energy equivalent amount calculation means 112 needs to calculate the amount of heat supplied. For example, if the exhaust heat is exhaust hot water, the input energy equivalent amount calculation means 112 can calculate the supply heat amount by multiplying the product of the flow rate of the exhaust hot water and the temperature difference between the entrance and exit by specific heat. Since the input energy equivalent amount calculation means 112 calculates the input energy equivalent amount, the specific heat need not be applied if a plurality of heating media are not used. For example, when the sampling period of the measurement data by the equipment performance diagnosis apparatus 110 is 1 unit time, if the collection period of the collection unit 103 is shorter than that, for example, 0.01 unit time, the input energy equivalent amount calculation unit 112 is used. As described in Patent Document 1, the calculated supply heat amount may be integrated for 100 samples, and the integrated value may be obtained as the input energy equivalent amount, but may simply be downsampled to 100 samples once. .

  Further, when power, gas, heat, etc. are mixed in the measurement data, the input energy equivalent amount calculation means 112 converts these into numerical values of a common unit such as the amount of heat and adds them.

  The delivery energy equivalent amount calculation means 113 calculates the delivery energy equivalent amount from the measurement data related to heat, work, etc. delivered by the target facility 101. For example, when the target facility 101 is air conditioning, the delivery energy equivalent amount calculation means 113 uses the product of the flow rate of the heat medium and the return temperature difference as the delivery energy equivalent amount. When the target facility 101 is a pump, the delivery energy equivalent amount calculation unit 113 can obtain the product of the flow rate and the discharge pressure as the delivery energy equivalent amount.

  Similarly to the explanation of the input energy equivalent amount calculation means 112, when the collection energy equivalent amount calculation means 113 is sufficiently shorter than the sampling period of the equipment performance diagnostic device 110, the integrated value is equivalent to the supply energy equivalent to the explanation of the input energy equivalent amount calculation means 112. It may be used as a quantity.

  The efficiency mapping means 115 calculates the efficiency using the input energy equivalent amount and the supplied energy equivalent amount calculated by the input energy equivalent amount calculation means 112 and the supplied energy equivalent amount calculation means 113, and uses it for the map based on the efficiency. The post-conversion efficiency equivalent amount is calculated. Since the efficiency is generally the amount of energy that can be delivered with respect to the unit input energy amount, the efficiency is calculated by the equivalent amount of supplied energy / the equivalent amount of input energy.

  Another amount used in the map is a mapping load equivalent amount that correlates with the efficiency and is related to the load of the target facility 101. The mapping load equivalent amount is, for example, the power consumption amount of the target facility per unit time, the outside air enthalpy, or the like. When the mapping load equivalent amount is the outside air enthalpy, for example, if the measurement data by the sensor 102 includes the outside air temperature (dry bulb temperature) and the relative humidity, the efficiency mapping unit 115 calculates the Wexler-Hyland equation from the outside air temperature. The saturated vapor pressure is determined using the above and the absolute humidity is determined from the relative humidity and the absolute humidity based on the definition of the absolute humidity and the relative humidity. Specific enthalpy is the amount of heat required to bring the air from 0 degrees to the temperature (dry air mass = 1 kg x temperature x isobaric specific heat) and the heat of evaporation required to evaporate the moisture in the air (water content (= absolute (Humidity x dry air mass = 1 kg) x water evaporation heat) and the amount of heat required to bring the water to that temperature (moisture content x specific heat degree of water). Further, if the measurement data by the sensor 102 includes the dry bulb and wet bulb temperatures of the outside air, the outside air enthalpy can be calculated using these.

  Based on the map of the mapping load equivalent amount and the converted efficiency equivalent amount, the measurement data used to calculate the input energy equivalent amount and the supply energy equivalent amount at a certain point in time indicates the efficient operating state of the target facility 101 at that point in time. It is possible to identify whether it is a result of sampling or a result of sampling a state temporarily inefficient due to equipment switching or the like. The mapping load equivalent does not have to be independent of efficiency. When humidity and temperature cannot be measured, for example, when a target system including a plurality of target facilities 101 is a diagnosis target, the efficiency mapping unit 115 is based on efficiency such as the energy consumption cost of the entire target system. A determined value may be used as the mapping load equivalent amount.

  The efficiency mapping unit 115 generates a set of mapping load equivalent amount and converted efficiency equivalent amount as samples based on measurement data for each sampling period, and generates a plurality of samples based on measurement data within a preset monitoring period. Generate as map data.

  FIG. 2 shows map data.

  The efficiency mapping unit 115 of the present embodiment uses the power consumption per unit time as the mapping load equivalent and uses the inverse efficiency equivalent as the post-conversion efficiency equivalent. The inverse efficiency equivalent is the reciprocal of efficiency. This map data is plotted in a map space using the power consumption per unit time on the horizontal axis and the reverse efficiency equivalent on the vertical axis.

  The efficiency mapping unit 115 converts the efficiency into a post-conversion efficiency equivalent so that the function representing the post-conversion efficiency equivalent amount with respect to the efficiency becomes an upward convex monotone increasing function or a downward convex monotone decreasing function. When such a conversion is performed, the higher the efficiency, the smaller the change in the equivalent amount after conversion with respect to the change in efficiency there, and the higher the density of samples near the boundary showing the highest efficiency in the map space, It becomes easy to extract a characteristic curve which will be described later. When map data is displayed, the operator (user) can easily determine the maximum efficiency of the facility. The efficiency mapping unit 115 may convert the efficiency into a post-conversion efficiency equivalent amount using, for example, a logarithm or an exponential function. The mapping load equivalent amount may be an outside air enthalpy, an outside air specific enthalpy, or the total input energy per unit time.

  FIG. 3 shows a characteristic curve extraction process.

  The characteristic curve extraction unit 116 executes a characteristic curve extraction process for extracting a characteristic curve indicating the maximum efficiency of the target facility from the map data.

  In STEP 401, the characteristic curve extraction unit 116 extracts the range of the mapping load equivalent amount in the map data.

  In STEP 402, the characteristic curve extraction unit 116 divides the range extracted in STEP 401 into N sections using a predetermined number of sections N. The characteristic curve extraction unit 116 determines feature points of each section, determines a feature curve representing the entire range using them, and determines a feature vector based on the feature curve. The characteristic curve extraction unit 116 of this embodiment uses a linear function as the characteristic curve, fits feature points of each of the N divided sections to a linear function by the least square method, and uses the intercept and slope of the linear function as a feature vector. . In this case, each section may change every time the characteristic curve extraction process is performed.

  Note that the characteristic curve extraction unit 116 may use feature points in each section as feature vectors. In this case, if the section changes each time the characteristic curve extraction process is performed, the feature vectors cannot be compared with a simple process such as an inner product. When the feature vector is expressed in such a form, the characteristic curve extracting unit 116 does not depend only on the range extracted in STEP 401 so that the section does not change each time the characteristic curve extraction process is executed, and the characteristic curve extracting unit 116 does not depend on the range. You may divide into sections using a range with a sufficient margin.

  In STEP 403, the characteristic curve extraction unit 116 performs the following processing 404 to 409 on each section of the mapping load equivalent amount. Here, the characteristic curve extraction unit 116 first extracts a value range corresponding to the post-conversion efficiency amount of the samples included in the section. The post-conversion efficiency equivalent amount is an amount that should take only a positive value, but may appear negative due to sampling as described above. For this reason, in the range extraction in this step, the range may be extracted from a limited value such as limiting to a positive value.

  In STEP 405, the characteristic curve extraction means 116 divides the value range of the post-conversion efficiency equivalent amount extracted in STEP 404 into M internal sections using a predetermined number of internal sections M. In STEP 406, the characteristic curve extraction unit 116 adds up the number of samples included in each internal section. As a result, a histogram of the converted efficiency equivalent amount is obtained for the section of the mapping load equivalent amount determined in STEP 403. In STEP 408, the characteristic curve extraction unit 116 fits the histogram thus obtained with a probability density function in which the random variable x satisfies 0 ≦ x <∞. As such a probability density function, for example, a gamma distribution, a Weibull distribution, an Erlang distribution, or the like can be considered.

  For example, when considering a gamma distribution having two parameters of shape and scale, using the method of moments, the scale parameter can be obtained by sample variance / sample average, and the shape parameter can be obtained by sample mean ^ 2 / sample variance.

  In STEP 409, the characteristic curve extraction unit 116 extracts feature points from the probability density function for the mapping load equivalent amount in each section. For example, the characteristic curve extraction unit 116 uses, for each section, a set of a post-conversion efficiency equivalent amount at which the probability density has a peak value and a representative value of the mapping load equivalent amount in the section as a feature point. The representative value is, for example, the central value of the section. Instead of the post-conversion efficiency equivalent amount at which the probability density takes a peak value, an expected value of post-conversion efficiency by a probability density function may be used.

  The characteristic curve extraction unit 116 extracts characteristic points for all sections, and then determines a characteristic curve indicating the entire range of the mapping load equivalent amount in STEP 410 using the characteristic points. Here, the characteristic curve extracting means 116 may apply a polynomial to a plurality of feature points obtained in STEP 409 of each section, determine parameters by the least square method, etc., and use these as characteristic curves. The characteristic curve extraction unit 116 of the present embodiment acquires feature points of each section, sets the characteristic curve as a linear function, applies a plurality of feature points to the linear function, and calculates the intercept and slope of the linear function as a feature vector. .

  In this embodiment, by representing the characteristic curve with a linear function, the feature of the map data can be represented by the intercept and the slope of the linear function as the feature vector.

  FIG. 4 shows feature vector information.

  The feature vector DB 117 records feature vector information 801 indicating the overall features of the map data. The feature vector information 801 includes each component of the feature vector. The feature vector information 801 may further include period information indicating a period during which the measurement data used for calculating the feature vector is measured. The period information includes at least one of a period start time and a period end time. The feature vector information 801 may further include operating facility information indicating whether or not each target facility is operating. The operating facility information may be a bit indicating whether or not each target facility is operating. It may represent the identification number of the target facility that is operating. An embodiment using the operating facility information will be described later. The feature vector DB 117 may record a set of mapped mapping load equivalent amount and converted efficiency equivalent amount.

  The characteristic change detection unit 118 compares the latest feature vector extracted by the characteristic curve extraction unit 116 with the past feature vector, displays a comparison screen showing the comparison result on the output device 119, and allows the operator of the equipment performance diagnosis device 110 to Encourage confirmation of change detection. The characteristic change detection unit 118 may display a comparison screen on the output device 119 when the comparison result satisfies a predetermined determination condition, or may display information indicating deterioration of the target facility on the output device 119. Good.

  FIG. 5 shows a comparison screen.

  The comparison screen of the present embodiment shows the latest data 602 that is map data calculated by the latest diagnosis processing and the comparison target data 603 that is map data calculated in the past. The characteristic change detecting unit 118 selects the comparison target data 603 by a method as described later. The operator may select the comparison target data 603 using the combo box 604. In order to examine the deterioration factor of the equipment, the operating equipment information 605 indicating the operating equipment in the target equipment is also necessary, so the comparison screen also displays the operating equipment information 605. The characteristic change detection unit 118 may calculate and display a comparison result 606 between the feature vector (past feature amount) of the comparison target data 603 and the feature vector of the latest data 602. Here, the characteristic change detection unit 118 calculates the past feature vector by normalizing the feature vector of the comparison target data 603 by the magnitude thereof, and the past feature vector and the feature vector of the latest data 602 are compared as the comparison result 606. May be calculated, or the correlation between the feature vector of the comparison target data 603 and the feature vector of the latest data 602 may be calculated. The aforementioned determination condition is, for example, that the comparison result 606 falls below a predetermined determination threshold.

  For example, when the comparison result 606 falls outside a predetermined range, the characteristic change detection unit 118 recognizes that the change is significant. When it is recognized that the change is significant and the feature vector is associated with the operating equipment information, the characteristic change detection unit 118 may narrow down the equipment considered to be a deterioration factor or prioritize the equipment. Good.

  The characteristic change detection unit 118 selects comparison target data 603 that satisfies a preset acquisition condition from the feature vector DB 117. The acquisition condition may be that a predetermined past time is a monitoring period. The timing of the map data of the comparison target data 603 differs for each type of equipment. In the case of air conditioning of a building or the like, the comparison target data 603 is preferably map data of one year or more. For example, the characteristic change detection unit 118 acquires comparison target data of a plurality of periods, such as one year ago, two years ago, three years ago, four years ago, and five years ago. You may compare. Further, the acquisition condition may be map data corresponding to the same operating equipment information as the latest operating equipment information among the map data at a predetermined time. Thereby, the state of the equipment of the newest data and comparison object data can be matched, and the precision of comparison can be improved.

  FIG. 6 shows the diagnostic process.

  The equipment performance diagnostic apparatus 110 performs the diagnostic processing 212 of 201 to 207 at a preset diagnostic cycle (for example, a one-month cycle or a ten-day cycle).

  The collection means 103 collects measurement data necessary for calculation of the input energy equivalent amount, the supplied energy equivalent amount, and the mapping load equivalent amount from the target equipment 101 through the sensor 102 at a predetermined collection period (for example, one hour period). Collect (201). The collection unit 103 collects them over the monitoring period (202), and transmits the measurement data collected when the collection is completed to the facility performance diagnosis apparatus 110 via the communication network 105 (210). The length of the monitoring period 202 may be equal to the diagnosis period, for example. The equipment performance diagnosis apparatus 110 that has received this measurement data calculates the input energy equivalent amount by the input energy equivalent amount calculation means 112 (203), and calculates the supply energy equivalent amount by the supply energy equivalent amount calculation means 113 (204). . The efficiency mapping means 115 performs mapping processing upon completion of these calculations (205). At this time, the efficiency mapping unit 115 also uses the measurement data transmitted from the collection unit 103 as information necessary for calculating the mapping load equivalent amount (210). The variable allocation definition data 114 is information that defines how measurement data should be used in the calculation of the input energy equivalent amount, the supply energy equivalent amount, and the mapping load equivalent amount. The variable assignment definition data 114 is used as an input of the calculation processes 203, 204, and 205 (210). Upon completion of the process 205 of the efficiency mapping means 115, the characteristic curve extraction means 116 performs a characteristic curve extraction process (206). When this process ends, the characteristic change detection means 118 stores the extraction result in the feature vector DB 117 and implements a change detection process (207). At this time, past feature vector calculation results are acquired from the feature vector DB 117. If a significant characteristic change is detected, the characteristic change detection unit 118 causes the output device 119 to display a comparison screen.

  FIG. 7 shows an input screen for the variable assignment definition data 114.

  The output device 119 displays an input screen for the variable assignment definition data 114. A part 1001 for inputting the correspondence of variables, a part 1002 for inputting a calculation expression for the input energy equivalent, a part 1003 for inputting a calculation expression for the supply energy equivalent, and a part for inputting the calculation expression for the mapping load equivalent 1004. The operator inputs the correspondence between the identifier (TAG) of the measurement data of the sensor 102 collected by the collection unit 103 and the internal variable name (ID) used inside the equipment performance diagnosis apparatus 110 to the part 1001. The operator may input an internal variable comment for the portion 1001. The operator inputs a calculation formula as a character string representing a formula including an internal variable name for the portions 1002, 1003, and 1004. The variable assignment definition data 114 is saved in a format such as text tagged with information input on the input screen.

  The input energy equivalent amount calculation means 112 calculates the input energy equivalent amount from the measurement data using the definition formula input to the portion 1002. The delivered energy equivalent amount calculation means 113 calculates the delivered energy equivalent amount from the measurement data using the definition formula input to the portion 1003. The efficiency mapping unit 115 calculates the mapping load equivalent amount from the measurement data using the definition formula input to the portion 1004.

  The efficiency mapping means 115 of this embodiment uses the outside air ratio enthalpy as the mapping load equivalent amount, and uses the inverse efficiency equivalent amount as the post-conversion efficiency equivalent amount. The reverse efficiency equivalent amount is input energy for supplying a unit amount of energy. Since the meaning of the inverse efficiency equivalent amount is easy to understand, the operator can easily evaluate and confirm the result.

  FIG. 8 shows map data of the second embodiment.

  The characteristic curve extraction means 116 extracts the lower envelope of this map data. The vertical axis represents the equivalent amount of reverse efficiency, which is calculated by the equivalent amount of input energy / equivalent amount of supplied energy. The amount of energy delivered is the product of the heat medium flow rate and the return temperature difference. Since the heat medium flow rate does not normally flow backward, it does not take a negative value. Also, the amount of input energy is the amount of electric power and the like, so it does not become a negative value. On the other hand, when the pump stops and the flow rate is zero or close to it, the return temperature is lower than the forward temperature even in the case of cooling, or the return temperature is higher than the forward temperature even in the case of heating. There is. For this reason, since samples are widely distributed outside the region shown in this figure, simply drawing the lower envelope corresponding to the reverse efficiency can extract the measurement data of the operating state corresponding to the maximum efficiency operation, Properties (characteristics) common to them cannot always be derived.

  The equipment performance diagnosis apparatus 110 can perform mapping that concentrates the samples of the timing at which the operation is performed at the highest efficiency by using the reciprocal of the efficiency as the post-conversion efficiency equivalent amount.

  In the present embodiment, a method of extracting feature vectors from feature points of N sections in STEP 410 of the characteristic curve extraction process will be described.

  FIG. 9 shows feature vectors of the third embodiment.

  The characteristic curve extraction unit 116 of this embodiment sorts the N feature points indicating the mapping load equivalent amount and the converted efficiency equivalent amount by the mapping load equivalent amount, and among the sorted feature points 802, after the conversion The vector corresponding to the efficiency is defined as a feature vector. The feature vector of the present embodiment can express a finer feature than the feature vector indicating the linear function of the first and second embodiments.

  In this embodiment, characteristic curve extraction processing will be described. The characteristic curve extraction means 116 of this embodiment calculates a characteristic curve where map data is concentrated using SFA (Stochastic Frontier Analysis). By applying such a method, the characteristic curve can be calculated as a linear model by a more general method.

  In this embodiment, characteristic curve extraction processing will be described.

  FIG. 10 shows a characteristic curve extraction process according to the fifth embodiment.

  Consider a straight line L perpendicular to the unit vector w at a distance b from the origin for an appropriate unit vector w and scalar quantity b in the matching result space. If the inner product of each sample x (t) and w of the matching result is <w, x>, a value 1101 obtained by subtracting b from <w, x> represents the distance from the line L to the sample x (t). The characteristic curve extraction unit 116 has a positive inner product <w, x> and a distance for all samples x (t) in which both the converted efficiency equivalent amount and the mapping load equivalent amount have positive values. Determine w and b that minimize the square sum of 1101. Samples for which either the post-conversion efficiency equivalent amount or the mapping load equivalent amount is negative are caused by the measurement timing and are considered to be samples that do not correspond to the optimal operating state. do not do. In this way, for the vector X (t) from the origin to each sample, the distribution of components in the direction perpendicular to the line L is 1102. The characteristic curve extraction unit 116 assumes that the distribution 1102 is, for example, a γ distribution, fits parameters of the γ distribution, and obtains the mode value δ of the γ distribution. As a result, the characteristic curve extracting unit 116 can obtain a straight line parallel to the line L and having the most dense samples as a line M perpendicular to w at a distance b + δ from the origin.

  The characteristic curve extraction means 116 can determine the line with the highest sample density for the matching result, although it is not completely strict by this method. An Erlang distribution, a Weibull distribution, or the like can be considered as a distribution function to be used. This can be determined based on the distribution shape of the sample.

  The facility performance diagnosis apparatus 110 according to the present embodiment was calculated from measurement data related to input energy equivalent amount or supply energy equivalent amount, operation facility information for identifying facilities operating during the measured period, and measurement data. Record in combination with feature vectors. In this way, since feature vectors can be stored for each combination of equipment that has been operating, it is possible to narrow down the equipment that causes performance degradation.

  FIG. 11 shows the deterioration factor determination process.

  First, in STEP 501, when the characteristic change detection unit 118 detects deterioration based on the comparison between the feature vector of the comparison target data and the feature vector of the latest data, it acquires operating facility information corresponding to the feature vector. Next, in STEP 502, the characteristic change detecting unit 118 selects only one unit from among the target facilities indicated in the operating facility information as a selected facility in order. In STEP 521, the characteristic change detecting unit 118 extracts the latest feature vector corresponding to the operating facility information in a state where all of the target facilities except the selected facility are operating as the latest stop vector from the feature vector DB 117. In STEP 522, the characteristic change detecting unit 118 extracts a feature vector corresponding to the same operating equipment information as the operating equipment information of the latest stop vector at the time closest to the comparison target data from the feature vector DB 117 as a comparison target stop vector. In STEP 523, the characteristic change detection unit 118 compares the latest stop vector with the comparison target stop vector. Since this comparison result is a scalar value such as an inner product or correlation as described above, a pair of the comparison result and the identification number of the selected facility is stored. At this time, the characteristic change detecting means 118 stores the pairs so that the comparison results are arranged in descending order from the larger comparison result (from the inner product closer to 1). The comparison result is an inner product of the feature vector normalized as described above and another feature vector. The characteristic change detection unit 118 may normalize either the latest stop vector or the comparison target stop vector. In the case of such a comparison result, the fact that the inner product is close to 1 corresponds to the fact that the change is small. The comparison result may be a correlation between the latest stop vector and the comparison target stop vector. The characteristic change detecting unit 118 repeats STEP 502 for all target facilities, and then outputs a list of identification numbers of the selected facilities in the order of the stored pairs. If the change in the feature vector when all the equipment other than the selected equipment is operating is small, the change in the feature vector when the selected equipment is operating will be large, so the selected equipment may contribute to deterioration. Becomes higher. This list shows the ranking of equipment that is likely to contribute to degradation. The characteristic change detection unit 118 may display the list on the output device 119. The characteristic change detecting means 118 may output the equipment with the highest rank, or may output a list of equipment corresponding to the comparison result exceeding the preset deterioration threshold in the order of the stored pairs. Good. Thereby, the operator can specify the deteriorated equipment.

  Note that a plurality of collection means 103 may be provided for each of the plurality of target facilities 101. In this case, the characteristic change detection unit 118 may sequentially perform the degradation factor determination process on the measurement data collected by using each collection unit 103 separately. When the facility performance diagnosis apparatus 110 monitors a plurality of customer facilities as a service, such a configuration method is standard.

  In this embodiment, a map screen that displays map data and feature vectors will be described.

  FIG. 12 shows a map screen.

  The characteristic change detection unit 118 causes the output device 119 to display a map screen in which the characteristic curve 702 is superimposed on the map data 701. As other display formats, the characteristic curve of the comparison target data and the characteristic curve of the latest data may be displayed in an overlapping manner, a table showing feature vector components, a sample distribution graph for each section, etc. May be. By displaying in this way, the operator can easily determine whether the characteristic curve is appropriate as a characteristic of the actual sample group.

  In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. It is also possible to replace part of the configuration of one embodiment with the configuration of another embodiment with the configuration of another embodiment, and to add the configuration of another embodiment to the configuration of one embodiment. Is possible. In addition, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

  Moreover, the facility performance diagnosis apparatus 110 may include a storage device, a communication device, and a processing device. In this case, each unit of the equipment performance diagnosis apparatus 110 may be realized as one or a plurality of programs. The storage device is a memory, a hard disk, an SSD (Solid State Drive), or the like, and stores information such as programs, tables, and files that realize each function. The processing device is, for example, a CPU (Central Processing Unit), and executes the above-described diagnostic processing according to this program. The communication device is, for example, a NIC (Network Interface Card), is connected to the collection unit 103 via the communication network 105, and acquires measurement data from the collection unit 103.

  Further, the diagnosis process by the facility performance diagnosis apparatus 110 can be considered as a performance diagnosis method for detecting deterioration or change in the performance of the target facility.

  According to each of the above embodiments, even when the integrated value and the instantaneous value are mixed in the measurement data, the equipment performance diagnostic device 110 maps the mapping load equivalent amount and the converted efficiency equivalent amount, The highest efficiency characteristics of the target equipment can be detected. Thereby, even when the integrated value and the instantaneous value are mixed in the measurement data, or even when the measurement data contains noise, a change in the performance of the target facility can be detected. Thereby, the equipment performance diagnostic device 110 can utilize information of sensors and forms originally installed in the equipment when providing equipment monitoring and diagnostic services to the existing equipment. In addition, the equipment performance diagnostic device 110 calculates the converted efficiency equivalent amount by converting the ratio of the supplied energy to the input energy by nonlinear conversion, and uses the relationship between the mapping load equivalent amount and the converted efficiency equivalent amount, It is possible to detect an amount corresponding to the performance of the target equipment in an ideal operation state. Since the upper limit of the efficiency of the target equipment is determined by the performance of each device and the control system that operates them, the equipment performance diagnostic device 110 can efficiently extract a sample of the timing at which it is operating at the highest efficiency, Degradation can be detected by comparing changes in the characteristics.

DESCRIPTION OF SYMBOLS 101 ... Target equipment, 102 ... Sensor, 103 ... Collecting means, 104 ... Information processing device, 105 ... Communication network, 110 ... Equipment performance diagnostic device, 112 ... Input energy equivalent amount calculating means, 113 ... Delivered energy equivalent amount calculating means, 114: Variable assignment definition data, 115: Efficiency mapping means, 116 ... Characteristic curve extracting means, 118 ... Characteristic change detecting means, 119 ... Output device

Claims (11)

A storage device;
A communication device connected to the communication network;
A processing device connected to the storage device and the communication device;
With
The communication device acquires a plurality of measurement data respectively measured at a plurality of time points within a monitoring period by a sensor provided in the facility via the communication network,
The processing device is
Based on each measurement data, calculate the input equivalent amount indicating the amount of energy input to the facility,
Based on each measurement data, calculate a delivery equivalent amount indicating the amount of energy delivered from the facility,
Based on each measurement data, calculate the load-related amount related to the load of the equipment,
Based on the input equivalent amount and the supply equivalent amount corresponding to each measurement data, an efficiency related amount related to the efficiency of the facility is calculated,
Calculating a feature amount indicating a relationship between the load-related amount and the efficiency-related amount corresponding to the plurality of time points;
Storing the calculated feature amount in the storage device;
Based on preset acquisition conditions, acquire a past feature quantity that is a feature quantity of the past period,
Comparing the calculated feature quantity with the past feature quantity;
Equipment performance diagnostic equipment. The processing device calculates the reciprocal of the efficiency of the equipment as the efficiency-related amount based on the input equivalent amount and the supply equivalent amount obtained from each measurement data.
The equipment performance diagnosis apparatus according to claim 1. The processing device, based on each measurement data, calculates either the outdoor enthalpy of the facility and the energy input to the facility per unit time as the load-related amount,
The equipment performance diagnostic device according to claim 2. The processing device causes the display device to display information indicating the result of the comparison when the result of the comparison satisfies a predetermined determination condition.
The equipment performance diagnosis apparatus according to claim 1. The feature amount is a vector,
The processing device calculates a correlation between the calculated feature quantity vector and the past feature quantity vector, and when the correlation falls below a predetermined determination threshold, information indicating deterioration of the equipment is displayed. Display on the device,
The equipment performance diagnostic device according to claim 4. The processing device causes a display device to display a plot of the load-related amount and the efficiency-related amount corresponding to the plurality of time points.
The equipment performance diagnosis apparatus according to claim 1. The processing device calculates a curve indicating the relationship, and causes the display device to display the curve.
The equipment performance diagnosis apparatus according to claim 1. The processing device calculates a straight line indicating the relationship, and calculates information defining the straight line as the feature amount.
The equipment performance diagnosis apparatus according to claim 1. The plurality of measurement data includes operation facility information that is measured during the monitoring period by sensors provided in each of the plurality of facilities and indicates facilities that are operating within the monitoring period among the plurality of facilities. ,
The processing device associates the operating facility information and the feature quantity and stores them in the storage device,
The processing device acquires, as the past feature amount, a feature amount stored at a preset time and associated with the same operation facility information as the operation facility information from the storage device.
The equipment performance diagnosis apparatus according to claim 1. The processing device acquires, from the storage device, a plurality of past feature amounts corresponding to a plurality of pieces of operating facility information different from each other when the comparison result satisfies a predetermined determination condition, and the plurality of pieces of operating facility information and on the basis of the plurality of past feature quantity to determine the likelihood of a particular facility degradation of the plurality of facilities, and displays the result of the determination in Viewing device,
The equipment performance diagnostic device according to claim 9. A plurality of measurement data respectively measured at multiple time points within the monitoring period by a sensor provided at the facility, obtained via the communications network,
Based on each measurement data, calculate the input equivalent amount indicating the amount of energy input to the facility,
Based on each measurement data, calculate a delivery equivalent amount indicating the amount of energy delivered from the facility,
Based on each measurement data, calculate the load-related amount related to the load of the equipment,
Based on the input equivalent amount and the supply equivalent amount corresponding to each measurement data, an efficiency related amount related to the efficiency of the facility is calculated,
Calculating a feature amount indicating a relationship between the load-related amount and the efficiency-related amount corresponding to the plurality of time points;
Storing the calculated feature amount in a storage device;
Based on preset acquisition conditions, acquire a past feature quantity that is a feature quantity of the past period,
Comparing the calculated feature quantity with the past feature quantity;
Equipment performance diagnostic method using equipment performance diagnostic equipment.

Images