JP2002174649A - Power consumption measuring system having abnormality reporting function - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the system reliability for totalizing and outputting the power consumption for every division or process from a measured value of a watt-hour meter placed anywhere in power distribution facilities. SOLUTION: The occurrence of an abnormality is detected by treating the measured value of the watt-hour meter by a plurality of discrimination logics, and the type of the discrimination logic detected the occurrence of the abnormality is outputted. In particular, a regression equation is stored which holds between the aggregated value of the power consumption for every division or process and an index showing the degree of activity for the division or the process, and a phenomenon is detected in which the power consumption for every division or process and the power consumption calculated by the regression equation largely separates. By the abnormality discrimination function, a wrong electric work or a failure of equipment is rapidly discovered, thereby preventing continuously measuring a wrong power consumption.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to a large-scale power distribution facility, for example, a power distribution facility of a large factory, which supplies power to a plurality of departments and / or processes.
The present invention relates to a technique for improving the reliability of a system for measuring which electric power amount is used by which department and / or process.

[0002]

2. Description of the Related Art In a large-scale power distribution facility, power is distributed to a plurality of departments such as a forging department, a machining department, and an assembly department, and it is required to measure the power consumption of each department. Alternatively, for example, power is distributed to a plurality of processes such as a crank process and an engine process, and it is required to measure the amount of power used in each process. By weighing by department and / or process, information such as which engine is being produced and how much power is used can be managed. In large-scale distribution facilities, electrical work is frequently performed, and the distribution system is frequently switched. For example, electrical work is frequently performed such that power is distributed from the distribution board A to the production equipment in the forging section until last week, but is distributed from the distribution board B from the following week. It is not easy to keep tabulating the power consumption by section and / or process while responding to frequently performed electrical works. The relationship between the aggregation formulas of the measurement values of each watt-hour meter for totalizing the amount '' is updated as the electrical work progresses.
An application has been filed in Japanese Patent Application No. 2000-362241 (however, this application has not been published yet). With this update technology, the accuracy and reliability of the power usage tabulated by department and / or process is greatly improved.

[0003]

However, actually, since various abnormal phenomena occur, there may still be a problem with the total value. For example, an error may occur in which accurate metering data cannot be logged due to a mistake in mounting the watt hour meter on the power distribution equipment or a trouble in the instrumentation system. If such an abnormal phenomenon occurs, the amount of electric power used for each department and / or each process will deviate from the true value. According to the present invention, an abnormal phenomenon detection function is incorporated into a system that totalizes and outputs power consumption by department and / or process from measured values of watt hour meters arranged at various places in distribution equipment, and has a function of detecting abnormal phenomena. And / or to further improve the accuracy and reliability of the amount of power consumption tabulated by process.

[0004]

Means for Solving the Problems, Action, and Effect One power usage metering system of the present invention has an abnormality notification function, and uses a power meter that is disposed at various parts of the power distribution equipment to measure the department and department. A device that totalizes and outputs the power consumption for each process, a device that detects the occurrence of an abnormality by processing the measured value of the watt-hour meter with a plurality of determination logics, and a type of the determination logic that detects the occurrence of the abnormality Output device.

According to the power consumption metering system having the abnormality notification function, the power consumption of each department and / or process is tabulated and output, and the measured value of the watt hour meter is processed by a plurality of determination logics. In order to detect the occurrence of abnormalities and to output the type of discrimination logic that has detected abnormalities when abnormalities are detected, if any abnormalities are included in the power consumption totaled by department and / or process This is immediately reported, and the accuracy and reliability of the power consumption tabulated by department or process is improved.

[0006] Another power consumption metering system having an abnormality notification function according to the present invention summarizes the power consumption of each department and / or process from the measured value of a watt hour meter arranged at various places in the power distribution equipment. A device that stores a regression equation that is established between a total value of power consumption for each department and / or process and an index indicating the activity of the department and / or process; The apparatus is provided with a device that compares the power consumption for each department and / or process and the power usage calculated by the regression equation to identify and report a department and / or process where the two are significantly different.

[0007] The power consumption of each department and / or process strongly correlates with an index indicating the activity of the department and / or process. Here, the index indicating the activity is preferably, for example, the number of manufactured acceptable products, the operation time of the equipment (in other words, the operation stop time of the equipment), the number of defective products generated, and the like. There is a relationship that the activity is higher and the power consumption increases as the number of the devices increases and the facility operation stop time is shorter. By using many past performance values, it is possible to obtain a regression equation that holds between the power consumption and the activity index, and the system of the present invention stores the regression equation thus obtained. ing. If an abnormality occurs in the weighing data, a regression equation that should be established between the power consumption and the activity index will not be established. In this system,
Comparing the total power consumption for each department and / or process with the power consumption calculated by the regression equation, and when the two are significantly different (for example, a difference of 20% or more), report the abnormality. Therefore, erroneous construction or the like can be quickly found, and erroneous power consumption can be prevented from being continuously measured.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 summarizes the power consumption by department and / or process to which power is supplied from a power distribution facility, and detects and reports any abnormal phenomenon if it has occurred. FIG. The power consumption totaling system 18 with an abnormality notification function is configured by a computer, and is operated by a divisional (process) -specific power usage totalization program 20 and an abnormality monitoring and notification program 22. The system 18 is connected to watt-hour meters 1, 2,... Arranged at various places in the power distribution equipment, and the measured value of each watt-hour meter is read into the system 18 at hourly intervals. This system 18 includes a distribution equipment management system 1
2 are connected. The power distribution equipment management system is also configured by a computer, and stores an equation used to total the power consumption of each department (process). For example, in the case of the distribution system indicated by reference numeral 10 at the lower left, the department (or process) A
In order to calculate the amount of power used for the measurement, the measured values of the watt hour meters 6, 3, and 4 may be added and the measured values of the watt hour meter 7 may be subtracted. In order to calculate the power consumption for the section (or process) B, the weighing values of the watt-hour meters 6 and the weighing values of the watt-hour meters 6 may be subtracted. The distribution formula 14 is stored in the power distribution equipment management system 12, and is used by the system 18.

The section (process) -specific power consumption totalizing program 20 is executed hourly, and as a result, the department (process) -specific power consumption is totaled every hour. Table 24 shows the contents of tallying, and illustrates an example of tallying for one department (process). Power consumption per hour (unit: KWH), daily cumulative power consumption per day, and monthly cumulative power consumption per month are obtained.

The system 18 executes the abnormality monitoring / notification program 20 every hour to detect the occurrence of an abnormal phenomenon. Various logics are prepared for detecting abnormal phenomena, and the power consumption for each process shown in Table 24 or the weighing value for each watt-hour meter before being tabulated for each process is processed. Detect the occurrence of

The abnormality determination logic is the same as that shown in FIG.
Species are used, and four types of abnormalities shown in FIG. 21 can be detected. The abnormality determination logic detecting the abnormality or the detected abnormality type is displayed on the monitor screen to call the operator's attention. Operator 28
Examines the data in detail based on the abnormal display on the monitor screen and examines the cause of the abnormality. Also, take necessary measures. When necessary, a notification is sent to the manager 30 of the department and / or process in which the abnormality has occurred by e-mail or the like.

First, a "0" value abnormality determination logic and a "0" value abnormality detected by the logic (type I abnormality)
Will be described. In this abnormality determination logic, the measured value of the amount of power consumption per hour of each watt hour meter stored in the procedure shown in FIG. 2 is to be processed. According to the processing procedure of FIG. 2, the value of the power consumption per hour of each watt hour meter is stored in the system 18 every hour. For example, in the column of the watt hour meter 1 at 11:00 on January 1 in FIG. 3, the power consumption from 00:00 to 01:00 measured by the watt hour meter 1 is stored. Stores power consumption per hour measured from 01:00 to 02:00.

The "0" value abnormality determination logic employs the processing procedure shown in FIG. This procedure is performed once a week, usually at the end of Friday. In this process, the watt hour meter whose hourly power consumption for one week before the execution timing is all “0” is extracted. The watt hour meter extracted in this manner is listed on the “0” value abnormal screen on the monitor screen. FIG. 22 exemplifies a change in the accumulated power consumption when the “0” value abnormality persists after 20 days. FIG.
An example is shown in which the measured values of the watt hour meters 4 and 6 continue to be "0" for one week. In this case, the location of the watt hour meter is shown for the convenience of the operator. FIG. 5 shows that the watt hour meter 4 is arranged on the low voltage
Exemplifies a case in which the circuit is arranged on a circuit breaker の of a distribution board xx of a low-voltage main line xx. Furthermore, the department and process in which the watt hour meter is involved are also displayed.
In this example, the watt hour meter 6 is involved in the β step of the B1 section, and the watt hour meter 6 is involved in the α step of the two sections.

When power is never used during one week, the power is not actually used and the equipment has no abnormality, and the equipment is erroneously wired (for example, load equipment that is to be distributed via the wattmeter 4). In other words, when power is distributed by another route), when one week has passed before the weighing data is read due to disconnection of the instrumentation system, etc., and when the load equipment is operating normally In some cases, such as when the watt hour meter is faulty or when the system is initialized incorrectly, the operator can immediately know that the data must be examined in detail to find out the real cause. When the "0" value abnormality determination logic indicates that an abnormality has been detected on the monitor screen, the operator examines the data. As a result, if it is confirmed that a type I abnormality ("0" value abnormality) shown in FIG. 21 has occurred, it is input to a computer to clarify the abnormality type. The specified abnormality type is displayed on the monitor screen.

It is found that it is necessary to correct the total value of the electric power consumption for each department (process) shown in Table 24 of FIG. 1 depending on the cause. As a result, the accuracy and reliability of the power consumption obtained by summing up by department (process) are improved.

Next, an out-of-appropriate range abnormality determination logic which is a second abnormality determination logic will be described. In the abnormality determination logic, the data of FIG. 3 stored in the procedure of FIG. 2, that is, the amount of power consumption per hour measured for each watt hour meter is processed to determine whether an abnormality has occurred. I do.
This process is executed once a week as shown in FIG. In this processing, for each watt-hour meter, the number of data that is equal to or greater than a predetermined maximum value (MAX), out of the 24-hour × 7-day power usage stored during one week, The number of data that is equal to or less than a predetermined minimum value (MIN) is counted, a watt-hour meter whose count number is equal to or more than 10 is extracted, and an abnormality is displayed on the watt-hour meter. Here, the maximum value (MAX) indicates that the power meter is normally used when viewed from the power distribution equipment (for example, a transformer, a low-voltage main line, an outlet from a breaker of a distribution board) in which the watt hour meter is arranged. Is defined as the amount of power per hour that should not exceed. Similarly, the minimum value (MIN)
Defines the amount of power per hour that should not fall below the normal value when viewed from the power distribution equipment where the watt-hour meter is placed. FIG. 7 exemplifies an abnormal phenomenon determined by the out-of-appropriate-range abnormality determining logic, and an anomaly is displayed on a watt-hour meter in which power usage outside the appropriate range is frequently detected. FIG. 8 shows an example of the abnormality display screen, in which the presence position in the power distribution system is shown so that the presence position and the like of the watt hour meter in which the abnormality is detected can be easily understood. In FIG. 8, a watt hour meter に placed in the high voltage feeder ○ is shown.
The mark 電力 indicates the case where the watt hour meter XX disposed on the breaker lead line XX frequently detects the power usage outside the appropriate range. When the monitor screen indicates that an abnormality has been detected by the out-of-appropriate abnormality determination logic, the operator examines the data. As a result, if it is confirmed that the type II abnormality (abnormality outside the proper range) shown in FIG. 21 has occurred, the abnormality is input to the computer to clarify the abnormality type. The specified abnormality type is displayed on the monitor screen.

Next, the balance abnormality determination logic, which is the third abnormality determination logic, will be described. The distribution system branches toward the end as illustrated in FIG. FIG. 9 illustrates a case where the low-voltage main line is branched into two by a distribution board. The number of branches is not limited to two. If the distribution system is branched, the sum of the power consumption measured before the branch and the power consumption measured after the branch should be equal. FIG.
In the case of, if the power consumption measured by the watt-hour meter X is X, the power consumption measured by the watt-hour meter A is A, and the power consumption measured by the watt-hour meter B is B, the relationship X = A + B Must hold. The third balance abnormality determination logic detects the occurrence of an abnormal phenomenon in which the above relationship is not established.
In the abnormality determination logic, the power consumption per day (exemplified in FIG. 10) obtained by summing the power consumption per hour shown in FIG. 3 over 24 hours is set as a processing target. As shown in FIG. 11, this abnormality determination logic
Executed several times. FIG. 11 exemplifies the content of the processing executed on January 28. On the left side of the discriminant, X (power consumption before branching) and A + B (power consumption after branching) for the past four weeks are shown. The sum of the squares of the differences is used. Here, a is a constant and is a value larger than 1. The right side is the square of the deviation on January 28, which is the processing target day, and it is determined that there is an abnormality when the right side is larger than the left side. At this time, January 2
On the 8th, it is displayed that an abnormal balance has occurred in the portion where X branches to A and B. The operator checks the cause of the balance abnormality according to the abnormality display. For example, if there is no abnormality in the balance at the portion further branching from A and no abnormality at the portion further branching from B, the weighing value of A and the weighing value of B are correct, and the weighing value of X may be shifted. It can be immediately guessed and checked. Alternatively, an abnormality such as a slight leakage in the wiring of the branch B (if a large leakage is detected by the second out-of-range abnormal abnormality determination logic. In this case, even a slight leakage can be detected) is detected. . FIG. 12 shows a screen displayed on the monitor screen when a balance abnormality is detected, and it is immediately understood that the balance abnormality is detected at the branches of X, A, and B. The description of X, A, and B (name,
(ID symbol, location) are also displayed. When an abnormality is detected by the balance abnormality determination logic, a “0” value abnormality (type I abnormality) or an out-of-appropriate range abnormality (type II abnormality) shown in FIG. 21 often occurs. When the balance abnormality is displayed on the monitor, the operator can analyze the data to find out if a “0” value abnormality or an abnormality outside the proper range has occurred. When the operator finds a "0" value abnormality or out-of-appropriate abnormality, it inputs it to the computer and clarifies the abnormality type. The specified abnormality type is displayed on the monitor screen.

The fourth "difference determination logic from the previous time" monitors the occurrence of a sudden change (a large change that does not occur in a normal state) in the measured value of the watt hour meter. Here, FIG.
And the hourly power consumption shown in FIG. 13 is set as a processing target, and is executed every hour as shown in FIG. The left side of the abnormality discriminant indicates the average value of the square of the difference between the hourly measured value for the past four weeks and the hourly measured value one hour ago. Here, b is a constant, and is a value larger than 1. J + 1 indicates the execution time of the abnormality determination processing. The right side is the square of the difference between the measured value Aj + 1 of the execution time of the abnormality determination process and the measured value Aj one hour before that. When the right side is larger than the left side, it is assumed that there is a difference abnormality from the previous time. FIG. 23 shows an example of a sudden abnormality detected by the “difference abnormality determination logic from the previous time”. If such a sudden abnormality occurs, it is detected by the “difference abnormality determination logic from the previous time”. In the difference abnormality determination logic from the last time, a sudden value abnormality of type III in FIG. 21 is detected. FIG. 14 shows a screen displayed on the monitor screen when an abnormality is detected by the difference abnormality determination logic with respect to the previous time, and indicates at what time and which watt hour meter the sudden value abnormality occurred. Also in this case, the description (name, ID symbol, location) of the watt hour meter and the like are also displayed.

The fifth abnormality determination logic, which is a variation abnormality determination logic for each time zone, detects an abnormality in the variation amount of the power consumption for each time zone. Here, the time zone refers to a difference between a factory operating time zone and a non-operating time zone. As shown in FIG. 15, the power consumption per hour during the operation time period is large, and the power consumption per hour during the non-operation time period is small. Naturally, the fluctuation range of the power consumption during the working hours is large, and the fluctuation width of the power consumption during the non-working hours is small. In this abnormality determination logic, the hourly measured values of the electric power consumption per hour of all wattmeters and the hourly total values of the electric power consumption per hour obtained by totaling the measured values for each department or process are processed. I do. Also, as shown in FIG.
This abnormality determination logic is executed every hour. In the abnormality determination logic, first, it is determined whether the execution time of the abnormality determination logic belongs to the operating time or the non-operating time zone. If it belongs to the operating time period, the average value and the standard deviation are obtained by statistically processing the measured value of the power consumption per hour measured in the past operating time period. Then, from the calculated average value and standard deviation, it is determined whether or not a large change that occurs only at a probability of 5% has occurred (specifically, an upper limit value obtained by adding 3σ to the average value, and a lower limit value obtained by subtracting 3σ from the average value). Compared to
It is determined whether it is between the upper limit and the lower limit). 5%
If a large change occurs only with the probability of, the fact that an abnormally large change has occurred is displayed. The same processing is performed during non-operation (however, the data to be statistically processed is limited to non-operation data). It is determined whether or not a large change that occurs only at a probability of 5% occurs during non-operation. to decide. FIG. 17 shows an example displayed on the monitor screen. If an abnormally large fluctuation occurs in the watt hour meter or the power consumption summed up by each department (process), a large fluctuation occurs in the watt meter or the department (process). And the time at which abnormally large fluctuations occur. When this abnormal display is performed, the operator examines the data in detail and pursues the cause of the abnormal display. The causes are roughly classified into two types.
It may be a sudden value abnormality shown in I, type IV
May be suddenly a "0" abnormality. An operator can determine which type of abnormality has occurred by examining the data in detail. The operator inputs the determined abnormality type to the system 18. The type of abnormality specified in this way is displayed on the monitor screen.

In the sixth abnormality determination logic, the power consumption per day of the department (and / or process) is compared with the production activity per day, and the production activity and the power consumption almost correspond. It is determined whether the error has occurred or does not correspond to the abnormality. FIG. 18 shows a correlation that is established between the amount of power consumption measured during normal times and the activity level. Here, the production quantity of the passed products produced in one day is adopted as an index indicating the activity best. Naturally, the daily power consumption and the daily production quantity are strongly correlated, and the larger the production quantity, the larger the power consumption. However, they do not completely match and vary. This is because, for example, the downtime of the equipment, the number of defective products, the temperature, and other factors affect the variation. Therefore, by performing a multivariate analysis of the relationship between the power consumption per day and the number of acceptable devices, the number of equipment downtime, the number of defective devices, the temperature, and other factors, the influence of each factor on the power consumption can be known. In the apparatus of this embodiment, the past normal data is subjected to multivariate analysis to calculate a correlation coefficient established between each factor and the power consumption. The actual abnormality determination logic is as shown in FIG.
Run once a day. Here, a comparison is made between the power consumption per day aggregated for each department (and / or process) and the power consumption estimated from the production activity per day for that sector (and / or process). It is determined whether they substantially match or greatly differ. Specifically, the index indicating the production activity per day and the power consumption per day calculated from the regression equation (predetermined) described with reference to FIG. Process) 80 to 1 of daily power consumption tabulated separately
It is determined whether it is within 20%. If it is out of the range, since both are abnormally different, the department (and / or process) where the abnormality occurred and the date are specified, and the planned power consumption (or estimated power consumption) and the actual power consumption are calculated. It indicates that an abnormality has occurred in which the amount of power consumption has greatly deviated.

When this abnormality is displayed, the operator can examine the data in detail and identify the cause of the abnormality. For example, the measured value of one watt-hour meter becomes "0" after a certain time, and it can be known that the measured value suddenly becomes "0" abnormal while the production is performed correctly. The failure of the meter can be quickly known. Type I
When the "0" value abnormality is not detected until one week has elapsed, the sixth abnormality determination logic (logic for detecting an abnormality in which the estimated power consumption and the actual power consumption greatly differ) is used. The "0" abnormality can be suddenly detected early.

FIG. 20 shows a list of the six abnormality determination logics described above, and shows the abnormality types that can be detected by each of the abnormality determination logics. Abnormality types are broadly classified from I to IV, and are listed in FIG.

In the "0" value abnormality determination logic, the type I
"0" value abnormal phenomenon is detected. The out-of-appropriate abnormality determination logic detects an out-of-appropriate abnormal phenomenon of Type II. When an abnormality is detected by the balance abnormality determination logic, the operator examines the data to find out whether a type I “0” value abnormal phenomenon or a type II abnormal phenomenon outside the proper range has occurred. Can be. The difference abnormality determination logic from the previous time detects a type III sudden abnormality phenomenon. When an abnormality is detected by the logic for determining the amount of fluctuation by time zone or the logic for detecting an abnormality that has a large difference from the estimated amount (estimated amount) by department (process), the operator examines the data to determine the type III. It can be found that a sudden abnormal phenomenon or a sudden "0" abnormal phenomenon of type IV has occurred.

The power consumption metering / counting / abnormality determination system of the present embodiment counts and outputs the power consumption of each department and / or process from the measured value of a wattmeter arranged at various places in the power distribution equipment. A device that detects the occurrence of an abnormality by processing the measured value of the watt-hour meter with six types of determination logic, and a monitor screen that outputs the type of the determination logic that has detected the occurrence of the abnormality. If any abnormality occurs, it is displayed, so that the accuracy of the power consumption by department and / or process is guaranteed, and the reliability of the total value is improved. In particular, the total value of the electric power consumption for each department and / or process, and an index indicating the activity of the department and / or process (for example, the production quantity of accepted products, equipment downtime, production quantity of rejected products, etc.) , The department storing the regression equation that is established between the departments, and the departments and / or departments in which the power consumption calculated by the regression equation is greatly different from those calculated by the regression equation. Since the apparatus is provided with a device for specifying and notifying the process, the reliability of the power consumption totaled for each department or process is ensured, and the continuation of the totalization of erroneous power usage is prohibited.

[Brief description of the drawings]

[Fig. 1] A system is constructed in which a power consumption metering / counting / abnormality determination system, a distribution equipment management system, and a watt hour meter group are connected to measure and aggregate the power consumption of distribution equipment to determine an abnormality. Indicates that it has been done.

FIG. 2 shows a process of storing hourly electric power consumption measured by each watt hour meter.

FIG. 3 schematically shows data contents stored in the processing of FIG. 2;

FIG. 4 shows a processing procedure of “0” value abnormality determination logic.

FIG. 5 shows a monitor display when an abnormality is detected by the “0” value abnormality determination logic.

FIG. 6 shows a processing procedure of an out-of-appropriate range abnormality determination logic.

FIG. 7 shows processing contents of an out-of-appropriate range abnormality determination logic.

FIG. 8 shows a monitor display when an abnormality is detected by an out-of-appropriate-range abnormality determination logic.

FIG. 9 illustrates a distribution system branch.

FIG. 10 schematically shows data processed by the balance abnormality determination logic.

FIG. 11 shows a processing procedure of a balance abnormality determination logic.

FIG. 12 shows a monitor display when an abnormality is detected by the balance abnormality determination logic.

FIG. 13 shows a processing procedure of difference abnormality determination logic from the previous time.

FIG. 14 shows a monitor display when an abnormality is detected by the difference abnormality determination logic from the previous time.

FIG. 15 shows the processing content of a time-slot-based variation amount abnormality determination logic.

FIG. 16 shows a processing procedure of a fluctuation amount abnormality determination logic for each time zone.

FIG. 17 shows a monitor display when an abnormality is detected by the time-zone-based fluctuation amount abnormality determination logic.

FIG. 18 schematically shows a relationship established between the amount of power consumption per day and the activity per day.

FIG. 19 shows a processing procedure of a logic for detecting an abnormality having a large difference from a scheduled amount (estimated amount) by department (process).

FIG. 20 shows a list of abnormality determination logic.

FIG. 21 shows a list of abnormality types.

FIG. 22 illustrates an example of a change in the accumulated power consumption measured when the “0” value abnormal phenomenon occurs.

FIG. 23 illustrates an example of a change in the accumulated power consumption measured when a sudden abnormality occurs.

Claims (2)

[Claims] An apparatus for summarizing and outputting power consumption for each department and / or process from a measured value of a watt hour meter arranged at various places in a distribution facility, and a plurality of measured values of the watt hour meter. An electric power consumption metering system with an abnormality notification function that includes a device that detects the occurrence of an abnormality by processing with the judgment logic, and a device that outputs the type of the judgment logic that detected the occurrence of the abnormality. 2. An apparatus for summarizing and outputting the electric power consumption for each department and / or process from the measured value of a watt-hour meter arranged everywhere in the distribution equipment, and an electric power consumption for each department and / or process. A device that stores a regression equation that is established between an aggregate value of the amount and an index indicating the activity of the department and / or process; an electricity usage amount for each aggregated department and / or process;
An electric power consumption metering system having an abnormality notification function including a device for comparing the electric power consumption calculated by a regression equation and specifying and notifying a department and / or a process in which both greatly differ.