JP7467876B2 - PERFORMANCE CHANGE DETECTION DEVICE, PERFORMANCE CHANGE DETECTION METHOD, AND PROGRAM - Google Patents

Description

The present invention relates to a performance change detection device, a performance change detection method, and a program.

In recent years, the performance of manufacturing plants and various facilities and equipment has become important. Performance refers to, for example, the ratio of output to input of manufacturing plants and various facilities and equipment. Specifically, for example, in the case of a power generation plant, it refers to the ratio of power generation to the amount of fuel input (i.e., the rate of power generation for a certain amount of fuel), and in the case of manufacturing equipment, it refers to the ratio of production volume to input raw materials (i.e., the amount of product for a certain amount of raw materials). These performances are widely recognized as important indicators not only in the industrial sector but also in the general consumer sector from the standpoint of energy conservation, resource conservation, and cost conservation.

When the performance of manufacturing plants and various facilities and equipment declines, costs increase and productivity declines, so changes in performance are a major concern for managers and operators of manufacturing plants and various facilities and equipment. For this reason, there is a demand to monitor changes in performance, and if a performance decline occurs, to carry out inspections and repairs early on in order to improve and restore performance.

Generally, performance changes due to various internal and external factors. For example, in the case of a power plant, performance changes due to the surrounding temperature and outside air temperature. For this reason, in order to correctly evaluate changes in performance of manufacturing plants, various facilities, equipment, etc., it is desirable to remove internal and external factors. In response to this, there is known technology that removes external factors such as outside air temperature and seawater temperature and evaluates changes in the performance values of the equipment that constitutes a power plant (see, for example, Patent Document 1).

JP 2012-21487 A

However, for example, in the technology described in Patent Document 1 above, the performance values after removing external factors are analyzed manually to capture any trends of decline, etc. As a result, even if a change in performance has occurred, for example, the change may be overlooked or it may not be possible to correctly determine whether a change has actually occurred, and the change in performance may not be captured.

The embodiment of the present invention was made in consideration of the above points, and aims to detect changes in performance.

To achieve the above object, the performance change detection device according to this embodiment includes an estimation means for estimating a model expressing the relationship between a state quantity and a performance value, using state data including a state quantity of an object for which a change in performance is to be detected, and performance data including a performance value of the object; a calculation means for calculating a distribution of the performance values of the object under predetermined standard conditions for each predetermined period, using the model estimated by the estimation means and the state data; and a detection means for detecting the change in performance, using the distribution for each period calculated by the calculation means.

Changes in performance can be detected.

10A and 10B are diagrams for explaining an example of state data and performance data; FIG. 2 is a diagram illustrating an example of a functional configuration of a performance change detection device according to the present embodiment. 11 is a flowchart illustrating an example of a performance change detection process according to the embodiment. FIG. 1 is a diagram for illustrating an embodiment of the present invention. FIG. 2 is a diagram illustrating an example of a hardware configuration of a performance change detection device according to the present embodiment.

Embodiments of the present invention will be described in detail below with reference to the drawings. In this embodiment, a performance change detection device 10 capable of detecting performance changes in manufacturing plants, various facilities, equipment, etc. will be described. Hereinafter, manufacturing plants, various facilities, equipment, etc. will be collectively referred to as "target facilities."

As described above, performance refers to the ratio of output (e.g., finished products, semi-finished products, electricity, etc.) to input (e.g., electricity, raw materials, fuel, etc.) of the target equipment. For example, if the target equipment is a power generation plant, the performance may be the ratio of the amount of electricity generated to the amount of fuel input. Also, if the target equipment is a manufacturing equipment, the performance may be the ratio of the amount of production to the amount of input raw materials. However, these performances are merely examples, and in this embodiment, any performance may be used as the performance of the target equipment.

In this embodiment, the state of the target equipment (e.g., temperature, pressure, flow rate, etc.) is measured by one or more sensors at equal sampling intervals, and these measurement results are obtained as two-dimensional state data. In addition, the performance of the target equipment is also calculated at the same intervals as the above measurement results, and these performance results are obtained as performance data. Note that the performance of the target equipment is calculated by calculating the ratio of the output to the input of the target equipment, etc.

The above-mentioned state data is expressed as two-dimensional data made up of multiple samples, with one sample being data represented by one or more state variables. This one sample is data that represents the measurement results when the target equipment is measured by one or more sensors at one sampling timing. Furthermore, the above-mentioned performance data is expressed as one-dimensional data made up of values (performance values) that represent specified performance.

An example of state data is shown in Fig. 1(a). The state data shown in Fig. 1(a) is two-dimensional data composed of K samples, with data represented by state variables _X1 , _X2 , ..., _XN being one sample. Each state variable _X1 , _X2 , ..., _XN is a variable representing a sensing result such as temperature, pressure, flow rate, etc. of the target equipment. Also, the sample number is a number representing each sample.

Hereinafter, the sample number will be represented by the variable k, and the sample with sample number k will also be represented as state vector _xk . Also, hereinafter, the value (state quantity) of state variable _Xn (n=1, ..., N) in state quantity vector _xk (k=1, ..., K) will also be represented as _xkn . In other words, each state quantity vector _xk is one-dimensional data (vector data) composed of state quantities _xkn (n=1, ..., N).

An example of the performance data is shown in Fig. 1(b). The performance data shown in Fig. 1(b) is one-dimensional data (vector data) composed of K performance values. The kth performance value is the performance value calculated when the state quantity vector _xk of the state data is obtained. Hereinafter, the performance value of sample number k will also be represented as _yk .

<Functional configuration>
First, the functional configuration of the performance change detection device 10 according to the present embodiment will be described with reference to Fig. 2. Fig. 2 is a diagram showing an example of the functional configuration of the performance change detection device 10 according to the present embodiment.

As shown in FIG. 2, the performance change detection device 10 according to this embodiment has a performance change detection processing unit 110 and a memory unit 120.

The storage unit 120 stores state data and performance data corresponding to the state data for each piece of target equipment. Here, the performance data corresponding to the state data refers to performance data composed of performance values _yk calculated when each row k of the state data of the K×N matrix (i.e., each state quantity vector _xk ) is obtained. In other words, the state data and the performance data correspond to each other when the state data and the performance data are composed of state quantities _xkn (n=1,...,N) and performance values _yk related to the same sample number k (k=1,...,K), respectively.

The performance change detection processing unit 110 executes a performance change detection process to detect a change in the performance of the target equipment. Here, the performance change detection processing unit 110 includes an input unit 111, a model estimation unit 112, a distribution calculation unit 113, a statistical test unit 114, and a change detection unit 115.

The input unit 111 reads and inputs status data and performance data for multiple time periods from the storage unit 120. Here, the width of the period is set arbitrarily in advance, but it is possible to set, for example, any time width (e.g., one hour width), one day width, one month width, etc.

The model estimation unit 112 uses the state data and performance data input by the input unit 111 to estimate a performance model that models the relationship between the state quantity and the performance value.

The distribution calculation unit 113 uses the performance model estimated by the model estimation unit 112 to calculate the distribution of performance values under standard conditions for each period. Here, the standard conditions refer to the standard state of the target equipment, and are, for example, state quantities when the internal and external factors take standard or representative values. In this case, the standard or representative values of the internal and external factors are any values empirically determined to be standard or representative by the user, etc., and a specific example is the average annual temperature when the external factor is the outside temperature. Note that the internal factors are factors that affect the performance of the internal structure, characteristics, etc. of the target equipment. Also, the external factors are factors that affect the performance other than the internal factors of the target equipment, and are, for example, the outside temperature and seawater temperature if the target equipment is a power generation plant.

The statistical testing unit 114 performs a statistical test between adjacent periods using the distribution of performance values for each period calculated by the distribution calculation unit 113. In other words, the statistical testing unit 114 tests whether or not there is a significant difference between the performance value in a certain period and the performance value in the period adjacent to this period.

The change detection unit 115 detects a change in performance using the results of the statistical testing performed by the statistical testing unit 114. For example, if there is a significant difference between the performance value in a certain period and the performance value in an adjacent period, the change detection unit 115 detects that the performance has changed between these periods. On the other hand, if this is not the case, the change detection unit 115 determines that the performance has not changed between these periods.

The performance change detection device 10 may be, for example, a device or system that functions as a PC (personal computer), a smartphone, a tablet terminal, a PLC (Programmable Logic Controller), or an MES (Manufacturing Execution System).

<Performance change detection process>
Next, a performance change detection process for detecting a performance change of a target facility by the performance change detection device 10 according to the present embodiment will be described with reference to Fig. 3. Fig. 3 is a flowchart showing an example of the performance change detection process according to the present embodiment.

First, the input unit 111 of the performance change detection processing unit 110 reads and inputs the state data and performance data for a plurality of periods from the storage unit 120 (step S101). Hereinafter, each of the plurality of periods is assumed to be P _j (j=1, ..., J), and the state data and performance data for the period P _j (j=1, ..., J) are assumed to have been input. In addition, the period P _j and the period P _j+1 are assumed to be adjacent periods to each other. For example, if the width of each period is one day, the period P _j is the jth day, and the period P _j+1 is the j+1th day. Or, for example, if the width of each period is one month, the period P _j is the jth month, and the period P _j+1 is the j+1th month.

The status data and performance data in the period _Pj refer to the status data obtained during the period _Pj (i.e., data obtained by measuring the status of the target equipment during the period _Pj ) and the performance data corresponding to this status data.

Next, the model estimation unit 112 of the performance change detection processing unit 110 estimates a performance model using the state data and performance data input in the above step S101 (step S102). Here, the performance model may be estimated for each period _Pj , or may be estimated for all periods _Pj . When estimating a performance model for each period _Pj , a performance model _fj is estimated for each period _Pj using the state data and performance data in that period _Pj . On the other hand, when estimating a performance model common to all periods _Pj , a performance model f is estimated using the state data and performance data in all periods _P1 to _PJ .

Hereinafter, as an example, a case where _{a performance model fj for} a period Pj is estimated will be described. The performance model _fj can be estimated from a physical causal relationship between a state and performance, or can be estimated by a statistical method from state data and performance data. In this embodiment, as an example, a case where a performance model _fj is estimated by a statistical method will be described.

As described above, in the period _Pj , the performance value _yk and the state vector _xk are acquired at the same timing. Therefore, assuming that the variable representing the performance value is y and the variable representing the state vector is x, a function y= _fj (x)+ _ej that approximates y with x is estimated as the performance model _fj . Here, _ej is an estimation error. This estimation can be performed using linear regression (single regression or multiple regression) or a nonlinear machine learning method. More specifically, for example, _fj (parameters) can be estimated by solving an optimization problem that minimizes the estimation error E shown in the following formula (1).

ここで、Ｋ_ｊは期間Ｐ_ｊにおける性能値ｙ_ｋ及び状態量ベクトルｘ_ｋのサンプル番号（インデックス）の集合である。

Here, K _j is a set of sample numbers (indexes) of the performance value y _k and the state vector x _k in the period P _j .

When estimating a performance model f common to all periods P ₁ to P _J , f _j in the above formula (1) should be replaced with f, and K _j should be replaced with the union of all K _j .

Next, the distribution calculation unit 113 of the performance change detection processing unit 110 calculates the distribution of the performance value under the standard conditions for each period using the performance model estimated in step S102 above (step S103). Here, the performance estimate value f _j (x _k ) obtained from the state quantity vector x _k and the performance model f _j in the period P _j can be regarded as the expected value (average value) of the performance in the state quantity vector x _k . On the other hand, since the actual performance value y _k is data obtained as a result having a variation (statistical variation) from the expected value f _j (x _k ), the variation (i.e., the degree of deviation from the expected value) can be expressed as y _k -f _j (x _k ).

Moreover, when the standard condition is _x0 , the expected value of the performance under the standard condition _x0 is expressed as _fj ( _x0 ). On the other hand, the variation of the actual performance value _y0 under the standard condition _x0 can be expressed as follows, for example. That is, if it is assumed that the distribution (variation) of the corresponding performance value is constant even if the state quantity changes, the difference _yk - _fj ( _xk ) between each performance value _yk and the performance estimate value _fj ( _xk ) is considered to follow the distribution (variation) of the actual performance under the standard condition _x0 . Therefore, the actual performance value under the standard condition _x0 in the period _Pj can be obtained by adding the amount _yk - _fj ( _xk ) corresponding to the distribution of the performance value in the period _Pj to the expected value of the performance under the standard condition _x0 , _fj ( _x0 ).

From the above, the distribution calculation section 113 can calculate the distribution of the actual performance values under the standard condition x ₀ in the period P _j as {y _0k = f _j (x ₀ ) + y _k - f _j (x _k )|kεK _j }.

Next, the statistical test unit 114 of the performance change detection processing unit 110 performs a statistical test between adjacent periods (i.e., between periods _Pj and Pj ₊₁ ) using the distribution of performance values under standard conditions _x0 for each period calculated in step S103 above (step S104). Hereinafter, the distribution of performance values under standard conditions _x0 in period _Pj is represented as _Yj = { _yk ^(j) | _k∈Kj }, and the distribution of performance values under standard conditions _x0 in period Pj ₊₁ is represented as Yj ₊₁ = { _yk ^(j+1) |k∈Kj ₊₁ }. Furthermore, the average value of _yk ^(j) is denoted by _μj , the average value of _yk ^(j+1) by μj ₊₁ , the standard deviation of _yk ^(j) by _sj , the standard deviation of _yk ^(j+1) by _sj+1 , the number of elements of _Kj by _Nj , and the number of elements of Kj ₊₁ by _Nj+1 .

At this time, the statistical testing unit 114 tests whether or not there is a significant difference between _Yj and Yj ₊₁ by testing the difference in the population means of the two groups. The null hypothesis is that "there is no difference between the population means of the two groups," and the alternative hypothesis is that "there is a difference between the population means of the two groups."

For example, when the population variance is known, the statistical testing unit 114 can test whether or not a significant difference exists using a normal distribution. That is, the statistical testing unit 114 first calculates the statistic u using the following formula (2).

そして、統計検定部１１４は、例えば危険率を５％として正規分布の累積分布関数から得られる値ｕ（５）＝１．９６０と、上記の式（２）により算出された統計量ｕとを比較し、ｕ＞ｕ（５）の場合にはＹ_ｊとＹ_ｊ＋１との間に有意差があるとし、そうでない場合にはＹ_ｊとＹ_ｊ＋１との間に有意差がないとする。

Then, the statistical test unit 114 compares the value u(5) = 1.960 obtained from the cumulative distribution function of the normal distribution with a risk rate of 5%, for example, with the statistical amount u calculated by the above formula (2), and if u > u(5), it determines that there is a significant difference between _Yj and Yj ₊₁ , and if not, it determines that there is no significant difference between _Yj and Yj ₊₁ .

For example, when the population variance is unknown, the statistical testing unit 114 can test whether or not a significant difference exists by a t-test. That is, the statistical testing unit 114 first calculates t using the following formula (3).

そして、統計検定部１１４は、例えば危険率を５％としてｔ分布の累積分布関数とサンプル数とに応じた閾値と、上記の式（３）により算出された統計量ｔとを比較し、ｔ＞閾値の場合にはＹ_ｊとＹ_ｊ＋１との間に有意差があるとし、そうでない場合にはＹ_ｊとＹ_ｊ＋１との間に有意差がないとする。

Then, the statistical test unit 114 compares the statistical amount t calculated by the above formula (3) with a threshold value according to the cumulative distribution function of the t-distribution and the number of samples, with a risk rate of, for example, 5%, and determines that there is a significant difference between _Yj and Yj ₊₁ if t > the threshold value, and determines that there is no significant difference between _Yj and Yj ₊₁ if not.

Next, if it is determined in step S104 that there is a significant difference between _Yj and Yj ₊₁ , the change detection unit 115 of the performance change detection processing unit 110 detects that the performance has changed between periods _Pj and Pj ₊₁ (step S105). Note that this detection result may be displayed on a display or the like of the performance change detection device 10, or may be notified to a terminal device or the like connected to the performance change detection device 10 via a communication network.

Here, in the above formulas (2) and (3), the numerator is an absolute value, but by replacing this numerator with μ _j -μ _j+1 , for example, it is possible to detect that performance has deteriorated in period P _j+1 compared to period P _j when there is a significant difference between Y _j and Y _j+1 . On the other hand, by replacing the numerator with μ _j+1 -μ _j , for example, it is possible to detect that performance has improved in period P _j+1 compared to period P _j when there is a significant difference between Y _j and Y _j+1 . Therefore, for example, the change detection unit 115 may display on a display or notify a terminal device connected via a communication network only when performance has deteriorated in period P _j+1 compared to period P _j .

As described above, the performance change detection device 10 according to this embodiment uses state data and performance data (i.e., a set of multiple samples and a set of performance values when these samples are obtained) to calculate the distribution of performance values of the target equipment under standard conditions, and then tests for significant differences in this distribution between different time periods, thereby detecting changes in performance. In particular, the performance change detection device 10 according to this embodiment is able to capture performance excluding internal and external factors by calculating the distribution of performance values of the target equipment under standard conditions, and is able to quantitatively and automatically capture significant differences in the distribution of performance between different time periods. This allows for rapid detection of performance changes, making it possible, for example, to rapidly take various measures such as maintenance inspection and repair of the target equipment in response to changes in performance.

<Example>
In the following, as an example, a case where a performance change of a certain plant is detected by the performance change detection device 10 according to the present embodiment will be described. In this example, one-dimensional state data including temperature as a state variable is used. In addition, assuming that two years' worth of state data has been obtained, the period is set in one-year units.

The state data and performance data used in the embodiment are shown in Fig. 4(a). As shown in Fig. 4(a), the performance value y is affected by the temperature x, and there is a tendency for the performance to be low when the temperature is high and high when the temperature is low. Fig. 4(b) shows a distribution diagram in which the state quantity (temperature x) included in the state data shown in Fig. 4(a) and the points indicated by the performance value y corresponding to this state quantity are plotted. As shown in Fig. 4(b), it can be seen that there is a negative correlation between the performance value y and the temperature x.

Therefore, in this embodiment, the performance model is estimated using the function y = ax + b, which estimates the performance value from temperature in common for all periods (i.e., the 2012 period and the 2013 period).

At this time, the above steps S101 to S103 were executed by the performance change detection device 10 according to the present embodiment, and the distribution of performance values under the standard condition x ₀ was calculated, as shown in FIG. 4(c). Note that the standard condition was x ₀ = 15°C.

When step S104 is executed by the performance change detection device 10 according to this embodiment, the statistic calculated by the above formula (2) is u = 3.8. Therefore, by comparing this with u(5) = 1.960, it is found that there is a significant difference, and it is possible to detect in step S105 that a performance change has occurred between the period of 2012 and the period of 2013.

As shown in FIG. 4(c), the distribution of performance values under standard conditions does not differ significantly between the years 2012 and 2013, so it is difficult to detect changes in performance, for example, by visual inspection. In contrast, the performance change detection device 10 according to this embodiment makes it possible to detect changes in performance as described above.

<Hardware Configuration>
Finally, the hardware configuration of the performance change detection device 10 according to the present embodiment will be described with reference to Fig. 5. Fig. 5 is a diagram showing an example of the hardware configuration of the performance change detection device 10 according to the present embodiment.

As shown in FIG. 5, the performance change detection device 10 according to this embodiment has an input device 201, a display device 202, an external I/F 203, a communication I/F 204, a ROM (Read Only Memory) 205, a RAM (Random Access Memory) 206, a processor 207, and an auxiliary storage device 208. Each of these pieces of hardware are connected to each other via a bus 209 so that they can communicate with each other.

The input device 201 is, for example, various buttons, a touch panel, a keyboard, a mouse, etc., and is used to input various operations to the performance change detection device 10. The display device 202 is, for example, a display, etc., and displays the processing results of the performance change detection device 10. Note that the performance change detection device 10 does not necessarily have to have at least one of the input device 201 and the display device 202.

The external I/F 203 is an interface with an external device. The external device may be a recording medium 203a or the like. The performance change detection device 10 can read and write data from and to the recording medium 203a or the like via the external I/F 203. The recording medium 203a may be, for example, an SD memory card, a USB memory, a CD (Compact Disk), or a DVD (Digital Versatile Disk). Note that one or more programs that realize the performance change detection processing unit 110 may be stored in the recording medium 203a.

The communication I/F 204 is an interface for connecting the performance change detection device 10 to a communication network. The performance change detection device 10 can perform data communication with other devices, equipment, etc. via the communication I/F 204. Note that one or more programs that realize the performance change detection processing unit 110 may be acquired (downloaded) from a specified server, etc. via the communication I/F 204.

ROM 205 is a non-volatile semiconductor memory that can retain data even when the power is turned off. RAM 206 is a volatile semiconductor memory that temporarily retains programs and data.

The processor 207 is, for example, a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), and is an arithmetic device that reads programs and data from the ROM 205, the auxiliary storage device 208, etc. onto the RAM 206 and executes various processes. The performance change detection processing unit 110 is realized, for example, by a process in which one or more programs stored in the auxiliary storage device 208, etc. are executed by the processor 207.

The auxiliary storage device 208 is, for example, a hard disk drive (HDD) or a solid state drive (SSD), and is a non-volatile memory that stores programs and data. The storage unit 120 is realized, for example, using the auxiliary storage device 208. However, the storage unit 120 may also be realized, for example, using a storage device connected to the performance change detection device 10 via a communication network.

The performance change detection device 10 according to this embodiment has the hardware configuration shown in FIG. 5, and is therefore capable of implementing the various processes described above. Note that the hardware configuration shown in FIG. 5 is merely an example, and the performance change detection device 10 may have other hardware configurations. For example, the performance change detection device 10 may have multiple processors 207, multiple auxiliary storage devices 208, multiple ROMs 205, multiple RAMs 206, etc.

The present invention is not limited to the specifically disclosed embodiments above, and various modifications and variations are possible without departing from the scope of the claims.

REFERENCE SIGNS LIST 10 Performance change detection device 110 Performance change detection processing unit 111 Input unit 112 Model estimation unit 113 Distribution calculation unit 114 Statistical test unit 115 Change detection unit 120 Storage unit

Claims (8)

an estimation means for estimating a model representing a relationship between a state quantity and a performance value, using state data including a state quantity of an object whose change in performance is to be detected and performance data including a performance value of the object;
a calculation means for calculating a distribution of performance values of the target under predetermined standard conditions for each predetermined period of time using the model estimated by the estimation means, the state data, and the performance data ;
a detection means for detecting a change in the performance by using the distribution for each period calculated by the calculation means;
having
The performance value is a value representing a ratio of an output to an input to the target,
The detection means is
A performance change detection device that, when a change in the performance is detected, further detects that the performance has deteriorated or improved, and, when no change in the performance is detected, detects that the performance has not changed. The detection means is
The performance change detection device according to claim 1 , further comprising: a statistical test being performed between distributions for different periods, and a change in the performance is detected when a significant difference exists between the distributions. The performance change detection device according to claim 2, wherein the different periods are a first period and a second period that is continuous with the first period and follows the first period. The detection means is
4. The performance change detection device according to claim 2, further comprising: a test for a normal distribution or a t-test to determine whether or not a significant difference exists between the distributions, depending on whether the variance of the distributions is known. The calculation means is
5. The performance change detection device according to claim 1, wherein a distribution of the performance values is calculated by adding, for each sampling timing at which the state quantity of the target is sampled, a difference between a first performance estimate value estimated by the model from the state quantity represented by the standard condition, and a second performance estimate value estimated by the model from one or more state quantities sampled at the same sampling timing, and the performance value corresponding to one or more state quantities. The performance change detection device according to any one of claims 1 to 5, wherein the standard conditions are the state quantities of the object when there are no internal and external factors that affect the performance of the object, or when the internal and external factors are constant and do not change. an estimation procedure for estimating a model expressing a relationship between a state quantity and a performance value of an object for which a change in performance is to be detected, using state data including a state quantity of the object and performance data including a performance value of the object;
a calculation step of calculating a distribution of performance values of the target under predetermined standard conditions for each predetermined period of time using the model estimated in the estimation step, the state data, and the performance data ;
a detection step of detecting a change in the performance by using the distribution for each period calculated in the calculation step;
The computer executes
The performance value is a value representing a ratio of an output to an input to the target,
The detection step includes:
A performance change detection method, further detecting that the performance has deteriorated or improved when a change in the performance is detected, and detecting that the performance has not changed when no change in the performance is detected. an estimation procedure for estimating a model expressing a relationship between a state quantity and a performance value of an object for which a change in performance is to be detected, using state data including a state quantity of the object and performance data including a performance value of the object;
a calculation step of calculating a distribution of performance values of the target under predetermined standard conditions for each predetermined period of time using the model estimated in the estimation step, the state data, and the performance data ;
a detection step of detecting a change in the performance by using the distribution for each period calculated in the calculation step;
Run the following on your computer:
The performance value is a value representing a ratio of an output to an input to the target,
The detection step includes:
If a change in performance is detected, the program further detects that the performance has decreased or improved, and if no change in performance is detected, the program detects that the performance has not changed.

Images