Interference estimation device, interference estimation method and procedure

Abstract

Technical Problem: This invention provides an apparatus for estimating interference generated in a controlled object. Solution: The interference estimation apparatus includes: an acquisition unit that acquires measurement values ​​measured by sensors provided with the controlled object; and an estimation unit that calculates a variance-covariance matrix of a measurement vector with the measurement values ​​as elements, performs singular value decomposition on the variance-covariance matrix to calculate a singular vector with the largest singular value, and estimates interference generated in the controlled object based on the singular vector.

Description

Technical Field

[0001] This disclosure relates to interference estimation apparatus, interference estimation method, and procedure. Background Technology

[0002] Waste-to-energy, which involves installing a boiler in a waste incinerator, recovering the heat generated during waste incineration, and using the generated steam to generate electricity, not only treats waste as waste but also generates added value from it as fuel, which is economically important. To increase the added value of waste as fuel and stabilize the amount of steam generated, it is effective to generate electricity according to a plan. Patent Document 1 discloses a control method related to waste-to-energy, which focuses on the moisture content of the waste, a cause of variations in the heat output of the waste incinerator, and adjusts the amount of waste supplied to the incinerator per unit time based on changes in the moisture content of the waste. Specifically, Patent Document 1 discloses the following technique: estimating the interference with steam flow based on the moisture content of the waste, adjusting the waste supply before the interference affects the controlled quantity (steam flow), thereby achieving stable power generation.

[0003] Patent Document 2 discloses a method for controlling the combustion of a waste incinerator by estimating the calorific value per unit of waste supplied. However, to estimate the calorific value per unit of waste supplied, several hours of data are required, and the estimated value is obtained by averaging several hours. Therefore, especially when the properties of the waste change over time, it is impossible to estimate the "calorific value per unit time of the waste" at the current moment. As a result, the estimated boiler evaporation rate (steam flow rate) used to adjust the waste and combustion air supplied to the incinerator becomes unreliable, and power generation may vary. In the technology of Patent Document 2, (1) the calorific value of the waste is calculated by measuring the concentration of oxygen and moisture in the exhaust gas using sensors, (2) the boiler evaporation rate is calculated based on the calculated calorific value of the waste, and (3) the supply of waste, combustion air, etc., to the incinerator is controlled based on the calculated boiler evaporation rate. That is, in Patent Document 2, interference is estimated based on the change in the calorific value of the waste, and the combustion of the waste incinerator is controlled based on the boiler evaporation rate calculated based on the calorific value as a manifestation of the interference.

[0004] Existing technical documents

[0005] Patent documents

[0006] Patent Document 1: Japanese Patent Application Publication No. 2019-178850

[0007] Patent Document 2: Japanese Patent No. 5996762 Summary of the Invention

[0008] The problem the invention aims to solve

[0009] The control methods disclosed in Patent Documents 1 and 2 are based on the characteristics of waste incinerators and employ a technique of controlling variables in a stable manner relative to changes in control variables such as steam flow rate. This is an inherent technology of waste incinerators. Recent machine learning methods have not demonstrated the generality of applying the same techniques to other objects. A method is required to presume a general approach to disturbances generated in the controlled object.

[0010] This disclosure provides an interference estimation device, interference estimation method, and procedure that can solve the above problems.

[0011] Technical solution

[0012] The interference estimation apparatus disclosed herein includes: an acquisition unit that acquires measurement values ​​measured by sensors of a controlled object; and an estimation unit that calculates a variance-covariance matrix of a measurement vector with the measurement values ​​as elements, performs singular value decomposition on the variance-covariance matrix to calculate a singular vector with the maximum singular value, and estimates interference generated in the controlled object based on the singular vector and the measurement vector.

[0013] Furthermore, the interference estimation method disclosed herein includes the following steps: obtaining measurement values ​​measured by sensors of the controlled object; calculating the variance-covariance matrix of the measurement vector with the measurement values ​​as elements; performing singular value decomposition on the variance-covariance matrix to calculate the singular vector with the maximum singular value; and estimating the interference generated in the controlled object based on the singular vector and the measurement vector.

[0014] Furthermore, the program disclosed herein enables a computer to perform the following steps: acquiring measurement values ​​measured by sensors possessed by the controlled object; calculating a variance-covariance matrix of a measurement vector with the measurement values ​​as elements; performing singular value decomposition on the variance-covariance matrix to calculate a singular vector with the maximum singular value; and estimating the disturbance generated in the controlled object based on the singular vector and the measurement vector.

[0015] Invention Effects

[0016] Based on the interference estimation device, interference estimation method and procedure described above, interference can be estimated. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the control system for each implementation method.

[0018] Figure 2 This is a diagram illustrating an example of the functional configuration of the main parts of the interference estimation device in the first embodiment.

[0019] Figure 3This is a diagram illustrating an example of the interference estimation processing in the first embodiment.

[0020] Figure 4 This is a diagram illustrating an example of the update timing of the processing in the first embodiment.

[0021] Figure 5 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device in the second embodiment.

[0022] Figure 6 This is a diagram illustrating an example of the interference estimation processing in the second embodiment.

[0023] Figure 7 This is a diagram illustrating an example of the functional configuration of the main components of the interference estimation device in the third embodiment.

[0024] Figure 8 This is a diagram illustrating an example of the interference estimation processing in the third embodiment.

[0025] Figure 9 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device according to the fourth embodiment.

[0026] Figure 10 This is a diagram illustrating an example of the time difference between the occurrence of the interference in the fourth embodiment and the time difference before its effect is reflected in the measured value.

[0027] Figure 11 This is a diagram illustrating an example of the interference estimation process in the fourth embodiment.

[0028] Figure 12 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device according to the fifth embodiment.

[0029] Figure 13 This is a diagram illustrating an example of the interference estimation process in the fifth embodiment.

[0030] Figure 14 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device according to the sixth embodiment.

[0031] Figure 15 This is a diagram illustrating an example of the hardware configuration of the interference estimation device in each embodiment. Detailed Implementation

[0032] The interference estimation device according to the embodiments will now be described with reference to the accompanying drawings. In the following description, components having the same or similar functions will be labeled with the same reference numerals. Furthermore, repeated descriptions of these components will sometimes be omitted. "XX or YY" is not limited to either XX or YY, but may include both XX and YY. This also applies when there are three or more selective elements. "XX" and "YY" are arbitrary elements (e.g., arbitrary information).

[0033] <First Implementation Method>

[0034] (constitute)

[0035] Figure 1 This is a schematic diagram of the control system for each implementation method.

[0036] The control system 100 includes an interference estimation device 10, a control device 20, and a controlled object 30. When controlling the controlled object 30, the interference estimation device 10 estimates interference that may affect its control quantity. The interference estimation device 10 includes an acquisition unit 11, an estimation unit 12, and an output unit 13. The acquisition unit 11 acquires measurement values ​​measured by sensors installed on the controlled object 30. When interference occurs in the controlled object 30, the interference manifests as a change in the measurement values ​​acquired by the acquisition unit 11. The estimation unit 12 uses the measurement values ​​acquired by the acquisition unit 11 to estimate the magnitude of the interference q generated by the control quantity. The output unit 13 outputs an estimated value of the interference q (hereinafter referred to as interference q or estimated value q) to the control device 20.

[0037] The control device 20 acquires the interference q from the interference estimation device 10 and acquires the measurement value measured by the sensor of the controlled object 30. The control device 20 controls the controlled object 30 based on the interference q and the measurement value. The controlled object 30 refers to various equipment such as waste incinerators, power generation equipment, chemical equipment, ships, gas turbines, steam turbines, compressors, etc. Hereinafter, a waste incinerator is used as an example of the controlled object 30 to illustrate the interference estimation processing of this disclosure, but the application of each embodiment is not limited to waste incinerators.

[0038] The measured values ​​acquired by the control device 20 include control quantities. The control device 20 controls the controlled object 30, for example, in a manner where the control quantity is fixed. Ideally, in a waste incinerator, the operation would be such that the flow rate of the generated steam is fixed. If the generated steam flow rate is fixed, the incinerator can continuously generate steam at its maximum capacity, thus maximizing the amount of waste incinerated, i.e., the processing volume, and the revenue from electricity sales based on power generation. However, waste recovered from the city is diverse, and even if waste is supplied to the furnace at the same rate over time, it is impossible to fix the steam flow rate. In the technology of Patent Document 1, the moisture content of the waste is measured; in the technology of Patent Document 2, the calorific value per unit mass of waste is estimated, and the cause of variations in the furnace's calorific value is estimated for the adjustment of waste and combustion air. Variables such as steam flow rate that are used for control purposes are generally referred to as control quantities. Furthermore, situations where the measured values ​​of the controlled object, including the control quantity, change contrary to the intended purpose are generally referred to as disturbances. In waste incinerators, variations in the moisture content of the waste and variations in the calorific value per unit mass of waste are representative disturbances. Furthermore, if these disturbances are understood macroscopically, changes in the overall heat output of the furnace also constitute disturbances. The techniques described in Patent Documents 1 and 2 are both techniques for estimating specific disturbances related to changes in combustion rate based on thermal knowledge associated with waste incinerators. In contrast, with the technology of this embodiment, disturbance q can be estimated even without existing knowledge about the controlled object, such as thermal properties. Specifically, the dominant disturbance is estimated based on measurements obtained from sensors installed on the controlled object. Hereinafter, the estimation process of disturbance q in the estimation unit 12 will be described using a waste incinerator as an example.

[0039] Figure 2 This is a diagram illustrating an example of the functional configuration of the main parts of the interference estimation device in the first embodiment.

[0040] Figure 2 The diagram shows the main components of the estimation unit 12 in the interference estimation device 10. The estimation unit 12 includes: a unit 121 that constructs an m-row, 1-column measurement vector y with the measured values ​​measured by the sensors of the controlled object 30 as elements; a unit 122 that calculates the variance-covariance matrix of the measurement vector y; a unit 123 that performs singular value decomposition on the variance-covariance matrix to calculate the singular vector with the maximum singular value; and a unit 124 that estimates the interference q of the controlled object 30 based on the singular vector with the maximum singular value.

[0041] (Interference estimation process)

[0042] Let q∈R1 represent the disturbance. In a waste incinerator, the dominant disturbance q is the combustion rate. As the disturbance q changes, the measured value y∈R1... m The changes also vary. As shown in equation (1), the changes of both can be approximated by a linear expression.

[0043] y=c1×q……(1)

[0044] In equation (1), c1 is an m-row, 1-column coefficient vector. c1 represents the response of the measurement vector y when the disturbance q increases. In a waste incinerator, if the combustion rate increases, the steam flow rate increases, the combustion chamber temperature rises, and the oxygen concentration in the exhaust gas decreases. c1 is a coefficient vector that quantifies the increase and decrease, and its value is determined as follows. First, calculate the variance-covariance matrix Q0∈R of the measurement vector y with the measured values ​​as column elements, as in equation (2). m×m In formula (2), Var is the notation for variance.

[0045] Q0=Var(y)……(2)

[0046] Next, singular value decomposition (SVD) is performed on the variance-covariance matrix Q0 to obtain the singular vector u of equation (3). i (i = 1, 2, ..., m) ∈ R m and singular value σ 2 i (i = 1, 2, ..., m) ∈ R + The singular values ​​are ordered here according to their magnitude, following the conventions of singular value decomposition. That is, σ 2 1 is the maximum singular value, σ 2 m It is the smallest singular value. The notation T on the right shoulder denotes the transpose of the matrix.

[0047] [Formula 1]

[0048]

[0049] Next, assume there is interference ρ. i (i = 1, 2, ..., m) ∈ R 1 As in equation (4), the variation of the measurement vector y is represented by the singular vector u and the unknown disturbance ρ. The elements of the unknown disturbance ρ are linearly independent, that is, if i ≠ j, then Cov(ρ) = ... i , ρ j = 0. Cov is the notation for covariance. The value of u can be obtained by performing singular value decomposition on the variance-covariance matrix Q0 of the measurement vector y. Therefore, the value of the unknown disturbance ρ can be calculated from the measurement vector y.

[0050] [Formula 2]

[0051]

[0052] Due to the symmetry of the variance-covariance matrix Q0, the singular vector u has the property of formula (5).

[0053] [Formula 3]

[0054]

[0055] Therefore, we can multiply u from left to right on both sides of equation (4). T The interference ρ is defined explicitly.

[0056] [Formula 4]

[0057]

[0058] Formula (7) is obtained when calculating the variance-covariance matrix of the disturbance ρ.

[0059] [Formula 5]

[0060]

[0061] The variance of ρ1, the first element of the unknown disturbance, is shown in equation (7), and is the maximum singular value σ1. 2 Therefore, it can be said that the variance of the measurement vector y is caused by the maximum value of the ρ1 component. This is because, according to the properties of singular values, the following equation (8) holds.

[0062] Var(y1) + Var(y2) + ... + Var(y m )=

[0063] σ1 2 +σ2 2 +……+σ m 2 ……(8),

[0064] Especially when σ1 2 >>σ2 2 +σ3 2 +……+σ m 2 When approximated as in the following formula (8A), the variation of the measurement vector y is dominated by ρ1.

[0065] [Formula 6]

[0066]

[0067] Take the part related to ρ1 from formula (6) to obtain formula (9) as the presupposition of interference q.

[0068] [Formula 7]

[0069]

[0070] In the example of a waste incinerator, it is known that the variation in combustion rate is the dominant disturbance q. When the acquisition unit 11 acquires the measured value y, the estimation unit 12 estimates the dominant disturbance q using formula (9) according to the above process. This disturbance q is the variation in combustion rate. The output unit 13 outputs the disturbance q to the control device 20. The control device 20 treats the disturbance q as a variation in combustion rate and adjusts the supply of combustion air and waste to offset it. As a result, the waste incinerator can be operated with a fixed steam flow rate. Not limited to waste incinerators, the dominant disturbance can be estimated for any controlled object. For example, in the automatic steering of a ship, the current is the dominant disturbance, and in the speed control of a car, the slope of the road surface is the dominant disturbance. These are insights derived from experience, rather than analytical units such as equations of motion.

[0071] (action)

[0072] The above process is shown in Figure 3 . Figure 3 This diagram illustrates an example of the interference estimation process in the first embodiment. First, the acquisition unit 11 acquires measurement values ​​such as steam flow rate, combustion chamber temperature, and exhaust oxygen concentration measured by sensors installed in the waste incinerator (step S1). The estimation unit 121 uses the measurement values ​​acquired by the acquisition unit 11 to construct a measurement vector y (step S2). For example, the unit 121 constructs a measurement vector y with each measurement value of steam flow rate, combustion chamber temperature, and exhaust oxygen concentration as elements. Next, the estimation unit 122 calculates the variance-covariance matrix Q0 according to formula (2) (step S3). Next, the estimation unit 123 performs singular value decomposition on the variance-covariance matrix according to formula (3) to calculate the singular vector with the largest singular value (step S4). Next, the estimation unit 124 estimates the interference q according to formula (9) (step S5). The estimation unit 12 outputs the interference q to the control device 20.

[0073] The relationship between the update timing of the measurement vector y, interference q, Q0, and u1 in the above interference estimation process is shown in the figure. Figure 4 For example, the measurement vector y arrives at the period T of the interference estimation device 10 with a new measurement value. A To update, the interference q is also updated periodically with time T. A Updates are performed. Furthermore, updates are performed on the variance-covariance matrix Q0 based on equation (2) and u1 based on the singular value decomposition of equation (3) with a period T. B Update the vector. The singular value vector u1 is determined by the variance-covariance matrix, therefore it is updated with the period T of the variance-covariance matrix Q0. B Perform calculations. The update cycle is determined based on the object's characteristics, but is typically T. A< <T B As a special case, if the characteristics of an object are fixed, the singular value vector u1 can also be fixed to a predetermined value.

[0074] According to this embodiment, the disturbance q affecting the control quantity can be estimated based on the measured values ​​obtained in the controlled object 30. In most cases, the sensors in the controlled object 30 are installed to measure the control quantity and physical quantities affecting it; therefore, there is no need to add new sensors, and the disturbance q can be estimated using the measured values ​​of the existing sensors. Furthermore, as... Figure 4 As illustrated in the example, the period T of the measurement vector y can be obtained. A The interference q is estimated based on the latest measurement value. Furthermore, since the process of acquiring the measurement value by the sensor installed on the controlled object 30 and the above procedure can be performed, it can be widely applied to interference estimation for various controlled objects 30 regardless of their characteristics.

[0075] <Second Implementation Method>

[0076] use Figure 5 , Figure 6 The interference estimation device of the second embodiment will be described.

[0077] In the first embodiment, the dominant disturbance (e.g., combustion rate) is known in advance. In a waste incinerator, the combustion rate is known to be the dominant disturbance, so the disturbance q calculated by formula (9) is the change in combustion rate, and the supply of combustion air and waste is adjusted based on the change in combustion rate. In the second embodiment, the disturbance is estimated by converting it into a change in the control quantity. Thus, even if the dominant disturbance is unknown in advance, the same method as in the first embodiment can be applied. For example, in the case of a waste incinerator, the control quantity is the steam flow rate, and estimating the disturbance by converting it into a change in the control quantity means estimating the change in combustion rate by converting it into a change in the steam flow rate caused by it. By estimating by converting it into a change in the control quantity, the following advantages can be obtained: the disturbance q can be estimated even without the knowledge that the change in combustion rate is the dominant disturbance; moreover, based on the control of the waste incinerator, it is not necessary to convert the magnitude of the disturbance q (combustion rate) into the steam flow rate as the control quantity, and it can be processed and used for control based on the state of the steam flow rate.

[0078] (constitute)

[0079] Figure 5 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device in the second embodiment.

[0080] The interference estimation device 10A of the second embodiment includes an estimation unit 12A instead of an estimation unit 12. The measurement value acquired by the acquisition unit 11 of the second embodiment includes a control quantity. Hereinafter, the control quantity will be described as a first element of the measurement vector y.

[0081] The estimation unit 12A in the second embodiment replaces unit 124 with unit 125, which estimates the disturbance applied to the object as a change in the control quantity based on the singular vector with the largest singular value. Similar to the first embodiment, the estimation unit 12A calculates the variance-covariance matrix Q0 of the measurement vector y using formula (2), and performs singular value decomposition on the variance-covariance matrix Q0 as in formula (3). Then, the estimation unit 12A calculates the singular value vector u. i (i = 1, 2, ..., m) and singular value σ 2 i (i = 1, 2, ..., m). Taking out the first row of equation (6) yields equation (6A).

[0082] [Formula 8]

[0083]

[0084] Here, u 1，j (j = 1, 2, ..., m) represents the j-th element of the singular value vector u1 with respect to the maximum singular value. If the maximum singular value corresponding to the first element (control variable) is dominant, i.e., if σ 2 1>>σ 2 2+σ 2 3+……+σ 2 m Then the measurement vector y is approximated as in formula (10A).

[0085] [Formula 9]

[0086]

[0087] The ξ of equation (6A) is obtained through the dominant ρ1 as shown in equation (11).

[0088] [Formula 10]

[0089]

[0090] Here, ξ|ρ1 represents ξ when ρ1 is used as the input condition.

[0091] Similarly, the control quantity is expressed through ρ1 as shown below.

[0092] y1|ρ1=u 11 ρ1……(12)

[0093] Here, y1|ρ1 represents y1 when ρ1 is used as the input condition. According to formula (12), when the value of ξ is known, the disturbance converted into the change of the control quantity is denoted as q. y1 |ξ, which is shown by the following formula (13).

[0094] [Formula 11]

[0095]

[0096] (action)

[0097] The above process is shown in Figure 6 . Figure 6 This diagram illustrates an example of the interference estimation processing in the second embodiment. First, the acquisition unit 11 acquires the measurement values ​​measured by the sensor installed in the waste incinerator (step S1). The measurement values ​​include the control quantity. Next, unit 121 of the estimation unit 12A constructs the measurement vector y (step S2). Unit 121 constructs the measurement vector y using the control quantity as the first element. Next, unit 122 of the estimation unit 12A calculates the variance covariance matrix Q0 according to formula (2) (step S3). Next, unit 123 of the estimation unit 12A performs singular value decomposition on the variance covariance matrix according to formula (3) to calculate the singular vector with the largest singular value (step S4). Next, unit 125 of the estimation unit 12A estimates the interference q converted into the variation of the control quantity according to formula (13). y1 |ξ(Step S6). The estimation unit 12A outputs interference q to the control device 20. y1 |ξ.

[0098] According to this embodiment, in addition to the effects of the first embodiment, even without knowledge of the disturbance, it is possible to calculate the estimated value q obtained by converting the disturbance q into a change in the control quantity based on the measured values ​​measured in the controlled object 30. y1 |ξ. For example, similar to the first embodiment, the measurement vector y and the disturbance q y1 |ξ with period T A Updates are performed. On the other hand, the singular value vector u1 determines its value based on the variance-covariance matrix Q0, therefore, the period T for updating the variance-covariance matrix Q0 is used. B Calculations are performed. Furthermore, if the properties of the object are fixed, the singular value vector u1 can also be fixed to a predetermined value.

[0099] <Third Implementation Method>

[0100] use Figure 7 , Figure 8 The interference estimation device of the third embodiment will be described.

[0101] In the third embodiment, the estimated value q, which is presumed to be a change in the control quantity, is compared.y1 The accuracy of the disturbance estimation is determined by comparing ξ with the measured value y1 of the control quantity. If the accuracy is poor, the control device 20 cancels the adjustment to counteract the disturbance. In the example of a waste incinerator, if the difference between the estimated steam flow rate variation and the actual steam flow rate is small, the combustion air and waste supply are adjusted based on the estimated steam flow rate variation to counteract the variation. On the other hand, if the difference between the estimated steam flow rate variation and the actual steam flow rate is large, the adjustment is canceled.

[0102] (constitute)

[0103] Figure 7 This is a diagram illustrating an example of the functional configuration of the main components of the interference estimation device in the third embodiment.

[0104] The interference estimation device 10B in the third embodiment includes an estimation unit 12B instead of an estimation unit 12. The estimation unit 12B calculates the estimation value q using the process described in the second embodiment. y1 |ξ, the presumed value q y1 The accuracy of |ξ. In addition to the configuration of the estimation unit 12A in the second embodiment, the estimation unit 12B also includes: unit 126, which calculates the error between the estimated control quantity variation and the actual control quantity; unit 127, which calculates the variance of the calculated error; and unit 128, which determines the reliability of the interference estimation based on the variance of the error. The measurement value acquired by the acquisition unit 11 in the second embodiment includes the control quantity. Hereinafter, the control quantity will be described as a first element configured as the measurement vector y. The output unit 13 in the third embodiment includes the estimated value q. y1 In addition to |ξ, it also outputs the judgment result of the interference estimation accuracy (adjustment limit instruction).

[0105] The estimation unit 12B acquires the estimated value q of the variation of the estimated control quantity (e.g., steam flow rate). y1 |ξ and the actual control quantity y1 are used to calculate the estimated value q of the variation using the following formula (14). y1 |The variance of the difference between ξ and the control variable y1.

[0106] J = Var(y1-q) y1 |ξ)……(14)

[0107] If the variance J is smaller than a preset threshold, the estimation unit 12B sets the adjustment limit command to off; if the variance J is larger than the preset threshold, the estimation unit 12B sets the adjustment limit command to on. The output unit 13 outputs the estimated value q calculated by the estimation unit 12B to the control device 20. y1 |ξ and adjustment limit command. If the adjustment limit command is off (variance J is smaller than the threshold), the control device 20 implements regulation to counteract disturbances. For example, in the case of a waste incinerator, this is to suppress fluctuations in steam flow (estimated value q).y1 The waste supply and combustion air supply are adjusted using the method of |ξ). If the adjustment restriction command is on (variance J is greater than the threshold), no adjustment to counteract the interference is implemented.

[0108] (action)

[0109] The above process is shown in Figure 8 . Figure 8 This diagram illustrates an example of the interference estimation processing in the third embodiment. First, using the processing described in the second embodiment, the estimation unit 12B estimates the interference q, which is converted into a change in the control quantity. y1 |ξ(Step S10). Next, unit 126 calculates the estimated change in the control quantity, i.e., the disturbance q. y1 |The error between ξ and the control quantity y1 (step S11). Next, unit 127 calculates the variance J of the error calculated in step S11 (step S12). Next, unit 128 determines the disturbance q based on the variance J of the error calculated in step S12. y1 The reliability of the presumed |ξ (step S13). If the variance J is greater than the specified threshold, unit 128 determines the presumed estimate to be unreliable; if the variance J is less than the specified threshold, unit 128 determines the presumed estimate to be reliable. For example... Figure 7 As shown, a hysteresis width can also be provided in this determination. By setting the hysteresis width, measurement errors and variations in the control quantity y1 can be absorbed, resulting in stable control. When the determination is deemed unreliable, unit 128 sets the adjustment limit command to open; when the determination is deemed reliable, unit 128 sets the adjustment limit command to close. The estimation unit 12B outputs the estimated value q of the disturbance to the control device 20. y1 |ξ and adjust the restriction command (on or off) (step S14).

[0110] In step S13, if the adjustment restriction command remains in the open state, the estimation unit 12B can also be improved. Figure 4 The update frequency of Q0 and u1 is explained to try to improve accuracy. If the accuracy does not improve even after this, a new value can be selected as the estimated value q. y1 The second element of the measurement vector y used in the calculation of |ξ is the measurement value used below.

[0111] According to this embodiment, in addition to the effects of the second embodiment, it can also confirm the interference estimation value q. y1 While maintaining the accuracy of |ξ, control of the controlled object 30 is performed. Furthermore, the value based on the adjustment limit command is adjusted to an estimated value q based on the disturbance. y1 The embedded control device 20 has a function that automatically switches between the execution and cessation of the adjustment of |ξ, which ensures the control accuracy of the control device 20.

[0112] <Fourth Implementation Method>

[0113] use Figures 9-11 The interference estimation device 10C of the fourth embodiment will be described.

[0114] There is typically a time lag between the occurrence of an disturbance and its manifestation as a change in a control variable or measured value. For example, in a waste incinerator, after a change in combustion rate as a disturbance, its effect manifests in the furnace temperature over a period of, for example, 10 seconds, and in the form of a change in steam flow rate over a period of, for example, 300 seconds. That is, for example, even if the combustion rate changes at time t, its effect manifests in the furnace temperature at time t+10 and in the steam flow rate at time t+300. Therefore, in this case, there is a 290-second time difference in response to the change in combustion rate for both furnace temperature and steam flow rate. If this is known, then the measurement vector y should be constructed by adding a 290-second time difference to both. For example, if the measurement vector y is set to include steam flow rate and furnace temperature, then the elements of the measurement vector at time t use the steam flow rate at time t and the furnace temperature at time t-290. When the value of the lag time is uncertain, multiple values ​​of the lag time can be set by changing the lag time. For example, in the aforementioned example, the elements of the measurement vector y are set to the steam flow rate at time t and the furnace temperature at time t-290. The furnace temperature at time t-350, time t-320, and time t-260 can also be added as elements of the measurement vector y.

[0115] Figure 9 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device according to the fourth embodiment.

[0116] Figure 9 The diagram shows the configuration of the fourth embodiment in combination with the second embodiment. In addition to the configuration of the second embodiment, the estimation unit 12C of the fourth embodiment also includes a lag time correction unit 129. For each element of the measurement vector y, the lag time correction unit 129 establishes a correspondence between previously measured values ​​and lag times and stores them. For example, if the measurement vector y includes steam flow rate and furnace temperature, the lag time correction unit 129 will establish a correspondence between the measured value of the furnace temperature 290 seconds ago, acquired by the acquisition unit 11 at time t, and the case where the furnace temperature lags behind the steam flow rate by 290 seconds. The lag time correction unit 129 acquires the measurement vector y, corrects the lag time of each element of the measurement vector based on the stored lag time values, and outputs the lag time-corrected measurement vector y. ～ In the fourth embodiment, the estimation unit 12C is based on the measurement vector y after hysteresis correction. ～The assumed disturbance is used to replace the measurement vector y.

[0117] use Figure 10 Further details on time correction are provided. Figure 10 This is a graph illustrating an example of the time difference between the occurrence of an interference and the time before its effect is reflected in the measured value. The measurement vector y has m elements; let the lag time of each element be τ. i (i = 1, 2, ..., m), then the time from the occurrence of the disturbance to the manifestation of a response in the elements of the measurement vector y is as follows: Figure 10 As shown. Here, for the sake of simplicity, let's assume it's the measurement vector y. ～ The elements are arranged in order of maximum lag time. The control variable is the final output of equipment such as waste incinerators, therefore the lag time is typically in the measurement vector y. ～ The lag time of the control variable is the largest among the elements. Since this is used to estimate disturbances to compensate for variations in the control variable, it is meaningless to use elements with a slower response than the control variable for disturbance estimation. Therefore, among the elements of the measurement vector y, the lag time of the control variable is naturally the largest. Consider calculating ξ in equation (6A) at time t. Figure 10 As shown, the information used for calculating ξ is prior to time t. On the other hand, it should be noted that the effect of the disturbance on the control variable y1 occurs after τΔ = τ2 - τ1. The measurement vector y after lag time correction... ～ The lag time can be expressed as shown in the following formula (15).

[0118] [Formula 12]

[0119]

[0120] The measurement vector y after using the hysteresis correction ～ Find the variance-covariance matrix Q0, and then calculate its singular vector u. At time t, {y} ～ 2, y ～ 3, ..., y ～ m} represents the current or past value, therefore the estimation unit 12C uses them to calculate ξ|ρ1 using formula (11), and uses formula (13) to calculate the interference q. y^1 |ξ. Interference q y^1 |ξ represents the predicted time t+τ at time t. Δ The prediction of changes in control parameters (e.g., steam flow rate). Since future values ​​are known, displaying them on the operation control panel of the waste incinerator can facilitate operation.

[0121] (action)

[0122] The above process is shown in Figure 11 . Figure 11 This diagram illustrates an example of the interference estimation processing in the fourth embodiment. First, the acquisition unit 11 acquires measurements such as steam flow rate, combustion chamber temperature, and exhaust oxygen concentration from sensors installed in the waste incinerator (step S1). Next, unit 121 of the estimation unit 12C constructs a measurement vector y (step S2). Then, the lag time correction unit 129 of the estimation unit 12C acquires the measurement vector y and outputs a lag time-corrected measurement vector y, which corrects for the lag time of each element. ～ (Step S7). The following steps are the same as in the second embodiment. That is, unit 122 of the estimation unit 12C calculates the variance-covariance matrix Q0 according to formula (2) (step S3). Next, unit 123 of the estimation unit 12C performs singular value decomposition on the variance-covariance matrix according to formula (3) to calculate the singular vector of the maximum singular value (step S4). Next, unit 125 of the estimation unit 12C estimates the disturbance q converted into a change in the control quantity according to formula (13). y^1 |ξ(Step S6). The estimation unit 12C will estimate the interference value (predicted value) q. y^1 |ξ outputs to control device 20.

[0123] According to the fourth embodiment, in addition to the effects of the second embodiment, the accuracy of interference estimation can be improved by correcting for the lag time in each measurement until the influence of the interference becomes apparent. The fourth embodiment can be combined not only with the second embodiment, but also with the first and third embodiments. Furthermore, when combined with the second and third embodiments, the fourth embodiment can predict future control quantities.

[0124] <Fifth Implementation Method>

[0125] use Figures 12-13 The interference estimation device 10D of the fifth embodiment will be described.

[0126] In the fourth embodiment, a predicted value for the change in control quantity caused by the disturbance is estimated. In the fifth embodiment, the fourth and third embodiments are combined to determine the accuracy of the prediction. If the prediction accuracy is known to be poor, the control device 20 cancels the adjustment to counteract the disturbance so that the error does not cause adverse effects. For example, in the case of a waste incinerator, if the difference between the predicted change in steam flow and the actual steam flow is small, the control device 20 adjusts the combustion air and waste supply based on the predicted change in steam flow to counteract the change. On the other hand, if the difference between the predicted change in steam flow and the actual steam flow is large, the control device 20 cancels the adjustment.

[0127] (constitute)

[0128] Figure 12This diagram illustrates an example of the functional configuration of the main components of the interference estimation device according to the fifth embodiment.

[0129] The interference estimation device 10D of the fifth embodiment includes an estimation unit 12D instead of an estimation unit 12. The estimation unit 12D calculates the estimation value q through a process similar to that described in the fourth embodiment. y^1 |ξ, the variance estimation value q based on the difference from the actual control quantity (e.g., steam flow rate) y1. y^1 The accuracy of |ξ. In addition to the configuration of the estimation unit 12C in the fourth embodiment and the units 126-128 in the third embodiment, the estimation unit 12D also includes a unit 130 that estimates the control quantity based on the variation of the control quantity. This variation of the control quantity is based on the lag time until the influence of the disturbance manifests in the control quantity, and the variation of the predicted value of the disturbance. The measurement value acquired by the acquisition unit 11 in the second embodiment includes the control quantity. The output unit 13 in the fifth embodiment outputs the predicted value y of the control quantity. ^ 1. The result of determining the prediction accuracy of the predicted value of the control quantity (adjustment limit instruction).

[0130] In the fifth embodiment, the measurement vector after lag correction is constructed as described in the fourth embodiment, as in formula (16).

[0131] [Formula 13]

[0132]

[0133] Measurement vector z ～ For the measurement vector y ～ The second element is the addition of the control variable y1(t). The measurement vector z is then calculated. ～ The variance-covariance matrix Q0 is obtained, and then its singular vector u is calculated. At time t, {z ～ 2, z ～ 3, ..., z ～ m+1} represents the current or past value, so we use them and use formula (17) to calculate ξ.

[0134] [Formula 14]

[0135]

[0136] Based on ξ, z is obtained as in equation (18). ～ 1 is the predicted value at time t+y1.

[0137] [Formula 15]

[0138]

[0139] The predicted value of the interference qz～1 |ξ(t) is q y^1 |ξ(t), is the control quantity y1(t+τ) at time t. Δ The predicted value of y1(t+τ). Δ The value is a predicted value, and is therefore denoted as y to distinguish it from the actual measured value. ^ 1(t+τ Δ If τ is used Δ × The time derivative of the predicted value with respect to time t and time t+τ Δ By approximating the increments of the predicted values, formula (19) is derived.

[0140] [Formula 16]

[0141]

[0142] According to equation (19), a differential equation representing the time variation of the predicted value is obtained as in equation (20).

[0143] [Formula 17]

[0144]

[0145] By numerically integrating equation (20) over time, the predicted value q based on the disturbance is obtained. y^1 |ξ(t) gives the estimated value of the control quantity y^1(t) at the current time t. In actual calculations, equation (20) can be obtained through a time constant τ, as shown in equation (21). Δ A first-order hysteresis filter with a gain of 1 can be used for simple calculation.

[0146] [Formula 18]

[0147]

[0148] Unit 130 obtains the value q based on the disturbance at time t+τΔ predicted at time t using formula (21). y^1 |ξ(t) is used to estimate the control quantity at time t. Unit 127 calculates the variance of the difference between the estimated value and the measured value (actual value) using the following formula (22) in the same manner as in the third embodiment, based on y^1(t) calculated by formula (21) and the measured value y1(t) of the actual control quantity.

[0149] J=Var(y1(t)-y^1(t))……(22)

[0150] (action)

[0151] Figure 13 This is a diagram illustrating an example of the interference estimation process in the fifth embodiment.

[0152] Through the processing described in the fourth embodiment, the estimation unit 12D predicts the disturbance q of the control quantity converted to time t+τΔ. y^1 |ξ (Step S20). Next, unit 130 estimates the control quantity at time t (Step S21). As described above, the predicted value of the disturbance q y^1 |ξ(t) represents the control quantity y^1(t+τΔ) at time t+τΔ. Unit 130 performs a time reversal calculation using formula (21) and estimates the control quantity y^1(t) at time t based on the control quantity y^1(t+τΔ). Next, unit 126 calculates the error between the estimated control quantity at time t and the actual control quantity (step S22). Next, unit 127 calculates the variance J of the error calculated in step S22 (step S23). Unit 127 calculates the variance J according to formula (22). Next, unit 128 determines the reliability of the estimation of the control quantity y^1(t) based on the variance J of the error calculated in step S23 (step S24). If the variance J is greater than a specified threshold, unit 128 determines that the estimation is unreliable; if the variance J is less than the specified threshold, unit 128 determines that the estimation is reliable. Figure 12 As shown, a hysteresis width can also be provided in this determination. By setting the hysteresis width, measurement errors and variations in the control quantity y1 can be absorbed, resulting in stable control. When the determination is deemed unreliable, unit 128 sets the adjustment limit command to open; when the determination is deemed reliable, unit 128 sets the adjustment limit command to close. The estimation unit 12D outputs the predicted value y^1 of the control quantity and the adjustment limit command (open or close) to the control device 20 (step S25).

[0153] In step S24, if the adjustment restriction command remains in the open state, the estimation unit 12D can also be improved. Figure 4 The update frequency of Q0 and u1 is explained to try to improve accuracy. If the accuracy does not improve even after this, a new value can be selected as the estimated value q. y^1 The measurement vector z used in the calculation of |ξ ～ The measurement value used below the second element, or its lag time reset.

[0154] According to this embodiment, in addition to the effects of the fourth embodiment, it can also confirm the interference prediction value q. y^1 While ensuring the prediction accuracy of |ξ, control of the controlled object 30 is performed. Furthermore, the value of the adjustment limit instruction is used to control the predicted value q based on the disturbance. y^1 The embedded control device 20 has a function that automatically switches between the execution and cessation of the adjustment of |ξ, which ensures the control accuracy of the control device 20.

[0155] <Sixth Implementation Method>

[0156] use Figure 14 The interference estimation device 10E of the sixth embodiment will be described.

[0157] In the presupposition of the disturbance in this disclosure, the singular vector u1∈R m It is important. The singular vector u1 is processed according to a pre-specified lag time {τ1, τ2, ..., τ...} m Every period T B Updates are performed. The values ​​of the singular vectors change with each update. Although the changed values ​​can be used directly, for example, if multiple singular vectors are calculated and the most preferred singular vector is used by majority vote, the reliability of the disturbance estimation is expected to be higher than when there is only one. The same applies to the lag time. Furthermore, for example, consider that the types of measured values ​​reflected by the disturbance may change depending on the operating mode of the controlled object 30 (start-up, rated operation, low output operation). Therefore, in the sixth embodiment, the update timing of the singular vectors, the lag time, and the measured values ​​constituting the measurement vector y are made different, and the prediction accuracy of the control quantity is determined by the method of the fifth embodiment. The control quantity with the highest accuracy is used to control the controlled object 30.

[0158] (constitute)

[0159] Figure 14 This diagram illustrates an example of the functional configuration of the main components of the interference estimation device according to the sixth embodiment.

[0160] The interference estimation device 10E of the sixth embodiment includes: a plurality of estimation units 12D of the fifth embodiment; a selection unit 131, which selects the smallest variance J from the variances J calculated by the plurality of estimation units 12D; a selection unit 132, which selects a predicted value y^1 of a control quantity corresponding to the variance J selected by the selection unit 131; and a selection unit 133, which selects an adjustment restriction command corresponding to the variance J selected by the selection unit 131. The output unit 13 of the sixth embodiment outputs the predicted value y^1 of the control quantity selected by the selection unit 132 and the adjustment restriction command selected by the selection unit 133.

[0161] For example, such as Figure 14 As shown, the system includes estimation units 12D-1 to 12D-2. The selection unit 131 selects the smallest variance among the variances [J]1 and [J]2 of the difference between the predicted value of the control quantity considering the variation of the control quantity caused by the disturbance and the actual control quantity. The smallest variance is then selected and designated as i. ＊ (Equation (23)).

[0162] [Formula 19]

[0163]

[0164] Then, select units 132 and 133 respectively and select number i. ＊ The output from the estimation unit 12D is used as the estimated value y^ for the change in the control quantity. i＊ Adjustment restriction instructions ＊ The output unit 13 outputs them to the control device 20.

[0165] According to this embodiment, the reliability of interference estimation can be improved. The estimation unit 12E may have a period T for the singular vector of the variance-covariance matrix Q0 of the updated measurement vector y. B Multiple estimation units 12D with different values ​​can be provided, or multiple estimation units 12D with different values ​​set for the lag time of the control quantity until the disturbance appears in the form of a change in the value of the control quantity can be provided, or multiple estimation units 12D with disturbance estimation based on different types of measurement vectors y can be provided, or multiple estimation units 12D with two or three of the three parameters being different.

[0166] Figure 15 This is a diagram illustrating an example of the hardware configuration of the interference estimation device in each embodiment.

[0167] The computer 900 has a CPU 901, a main storage device 902, an auxiliary storage device 903, an input / output interface 904, and a communication interface 905.

[0168] The aforementioned interference estimation devices 10-10E are installed in the computer 900. Furthermore, each of the aforementioned functions is stored as a program in the auxiliary storage device 903. The CPU 901 reads the program from the auxiliary storage device 903 and extends the program to the main storage device 902, executing the aforementioned processing according to the program. Additionally, the CPU 901 secures a storage area in the main storage device 902 according to the program. Furthermore, the CPU 901 secures a storage area in the auxiliary storage device 903 according to the program for storing data during the processing.

[0169] It should be noted that programs for implementing all or part of the functions of interference estimation devices 10-20E can also be stored in a computer-readable storage medium, and the program stored in the storage medium can be read into the computer system and executed to perform processing of each functional unit. Here, "computer system" refers to hardware including an operating system (OS) and peripheral devices. Furthermore, if the "computer system" utilizes a WWW system, it also includes a homepage providing environment (or display environment). "Computer-readable storage medium" refers to removable media such as CDs, DVDs, and USB drives, and storage devices such as hard drives built into the computer system. Alternatively, when the program is distributed to computer 900 via a communication line, the receiving computer 900 can expand the program in main storage device 902 and execute the aforementioned processing. Furthermore, the program can be a program for implementing a portion of the aforementioned functions, or it can be a program that can be implemented in combination with a program that has already stored the aforementioned functions in the computer system.

[0170] As described above, some embodiments of this disclosure have been illustrated. All of these embodiments are provided by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other ways, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included within the scope and spirit of the invention, as well as within the scope of the invention as described in the claims and its equivalents.

[0171] <Postscript>

[0172] The interference estimation devices 10 to 10E, interference estimation methods, and procedures described in each embodiment can be understood, for example, in the following manner.

[0173] (1) The interference estimation device 10-10E of the first scheme includes: an acquisition unit 11, which acquires the measurement values ​​measured by the sensors of the controlled object; and an estimation unit 12, which calculates the variance-covariance matrix of the measurement vector with the measurement values ​​as elements, performs singular value decomposition on the variance-covariance matrix to calculate the singular vector with the maximum singular value, and estimates the interference generated in the controlled object based on the singular vector and the measurement vector.

[0174] Therefore, the interference generated in the controlled object can be estimated based on the measurement values ​​measured by the sensors of the controlled object (first embodiment).

[0175] (2) The interference estimation device 10A to 10E of the second scheme is the interference estimation device 10A to 10E of (1). The acquisition unit acquires the measurement value of the target variable of the control that becomes the control object, i.e., the control quantity. The estimation unit estimates the interference as the change of the control quantity based on the singular vector (formula (13)) of the maximum singular value in the measurement vector.

[0176] Therefore, the disturbance can be converted into a change in the control quantity for estimation, and the magnitude of the disturbance can be estimated even when the dominant disturbance is unknown in advance (second embodiment).

[0177] (3) The interference estimation device 10B to 10C of the third scheme is the interference estimation device 10B to 10C of (1) to (2). The estimation unit determines the reliability of the estimated interference based on the magnitude of the variance of the estimated interference, and outputs the result of the determination together with the estimated interference.

[0178] Therefore, the reliability of the interference estimation can be determined. For example, in cases of low reliability, the accuracy of control can be ensured by controlling the object without using the estimation result (third embodiment).

[0179] (4) The interference estimation device 10C to 10D of the fourth scheme is the interference estimation device 10C to 10D of (1) to (3). The estimation unit has a correction unit. The correction unit corrects the lag time until the interference manifests as a change in the measured value. The estimation unit corrects the lag time of the measured value obtained by the acquisition unit through the correction unit and uses the measurement vector with the corrected measured value as an element to estimate the interference.

[0180] There is usually a time lag between the generation of a disturbance in the controlled object and its manifestation as a change in the control variable or measured value. According to the fourth embodiment, by taking into account the time lag when estimating the disturbance, the estimation accuracy of the disturbance can be improved (fourth implementation).

[0181] (5) The interference estimation device 10D of the fifth scheme is the interference estimation device 10D of (4). The estimation unit estimates the estimated value of the measured value with the longest lag time based on the estimated interference. Based on the variance of the difference between the estimated value of the measured value and the measured value of the measured value, the reliability of the estimated interference is determined, and the result of the determination is output together with the estimated value of the measured value.

[0182] Therefore, the disturbance is converted into a change in the control quantity, and compared with the measured value of the control quantity to determine the accuracy of the disturbance estimation. This ensures the accuracy of control using disturbances (Fifth Implementation).

[0183] (6) The interference estimation device 10E of the sixth scheme has a plurality of (5) interference estimation devices 10D estimation units, and selects the interference estimated by the estimation unit with the highest reliability of the interference.

[0184] Therefore, the disturbances estimated with the highest accuracy can be used for control (sixth embodiment).

[0185] (7) The interference estimation device 10E of the seventh scheme is the interference estimation device 10E of (6), wherein the plurality of estimation parts respectively estimate the interference based on the measurement value that is different from the other estimation parts, or estimate the interference by performing a correction of the lag time that is different from the other estimation parts, or estimate the interference by updating the variance covariance matrix and the singular vector at a period that is different from the other estimation parts.

[0186] By providing various conditions to estimate interference, the likelihood of being able to estimate interference with high accuracy can be increased.

[0187] (8) The interference estimation method of the eighth scheme has the following steps: obtaining the measurement values ​​measured by the sensors of the controlled object; calculating the variance-covariance matrix of the measurement vector with the measurement values ​​as elements; performing singular value decomposition on the variance-covariance matrix to calculate the singular vector with the maximum singular value; and estimating the interference generated in the controlled object based on the singular vector and the measurement vector.

[0188] (9) The program of the ninth scheme enables the computer to perform the following steps: acquire the measurement values ​​measured by the sensors of the controlled object; calculate the variance-covariance matrix of the measurement vector with the measurement values ​​as elements, perform singular value decomposition on the variance-covariance matrix to calculate the singular vector with the maximum singular value, and estimate the disturbance generated in the controlled object based on the singular vector and the measurement vector.

[0189] Explanation of reference numerals in the attached figures

[0190] 100: Control system;

[0191] 10~10E: Interference estimation device;

[0192] 11: Acquisition Department;

[0193] 12~12E: Estimated part;

[0194] 13: Output section;

[0195] 20: Control device;

[0196] 30: Controlled object;

[0197] 900: Computer;

[0198] 901: CPU;

[0199] 902: Main storage device;

[0200] 903: Auxiliary storage device;

[0201] 904: Input / output interface;

[0202] 905: Communication interface.

Claims

1. An interference estimation device, the interference estimation device comprising: The acquisition unit acquires measurement values ​​obtained by the sensors of the controlled object; and The estimation unit calculates the variance-covariance matrix of the measurement vector with the measured values ​​as elements, performs singular value decomposition on the variance-covariance matrix to calculate the singular vector with the maximum singular value, and determines a linear expression that approximately represents the relationship between the measured values ​​and the interference by estimating the product of the transpose of the singular vector with the maximum singular value calculated by the singular value decomposition and the measurement vector as the interference generated in the controlled object, and uses the measurement vector with the measured values ​​acquired by the acquisition unit as elements to estimate the interference generated in the controlled object.

2. The interference estimation device according to claim 1, wherein, The acquisition unit acquires the measured value of the variable, i.e., the control quantity, which is the purpose of controlling the controlled object. The estimation unit estimates the disturbance as a change in the control quantity based on the singular vector of the maximum singular value of the measurement vector.

3. The interference estimation device according to claim 1, wherein, The estimation unit determines the reliability of the estimated interference based on the magnitude of the variance of the estimated interference, and outputs the determination result together with the estimated interference.

4. The interference estimation device according to claim 2, wherein, The estimation unit determines the reliability of the estimated interference based on the magnitude of the variance of the estimated interference, and outputs the determination result together with the estimated interference.

5. The interference estimation device according to any one of claims 1 to 4, wherein, The estimation unit includes a correction unit that corrects for the hysteresis time until the interference manifests as a change in the measured value. The estimation unit corrects the lag time of the measurement value acquired by the acquisition unit through the correction unit, and uses the measurement vector with the corrected measurement value as an element to estimate the interference.

6. The interference estimation device according to claim 5, wherein, Based on the estimated interference, the estimation unit estimates the estimated value of the measurement value with the longest lag time among the measured values, determines the reliability of the estimated interference based on the variance of the difference between the estimated value and the measured value, and outputs the determination result together with the estimated value of the measured value.

7. The interference estimation device according to claim 6, wherein, The system has multiple estimation units, and the interference estimated by the estimation unit with the highest reliability is selected.

8. The interference estimation device according to claim 7, wherein, The multiple estimation parts estimate the interference based on measurements that are different from those of the other estimation parts, or estimate the interference by correcting for lag times that are different from those of the other estimation parts, or estimate the interference by updating the variance-covariance matrix and the singular vector at a different period than those of the other estimation parts.

9. An interference estimation method, the interference estimation method comprising the following steps: Acquire the measured values ​​obtained by the sensors of the controlled object; and Calculate the variance-covariance matrix of the measurement vector with the measured values ​​as elements, perform singular value decomposition on the variance-covariance matrix to calculate the singular vector of the maximum singular value, determine a linear expression that approximately represents the relationship between the measured values ​​and the interference by estimating the product of the transpose of the singular vector of the maximum singular value calculated by the singular value decomposition and the measurement vector as the disturbance generated in the controlled object, and use the measurement vector with the acquired measured values ​​as elements to estimate the disturbance generated in the controlled object.

10. A non-transitory computer-readable storage medium storing a program that causes a computer to perform the following steps: Acquire the measured values ​​obtained by the sensors of the controlled object; and Calculate the variance-covariance matrix of the measurement vector with the measured values ​​as elements, perform singular value decomposition on the variance-covariance matrix to calculate the singular vector of the maximum singular value, determine a linear expression that approximately represents the relationship between the measured values ​​and the interference by estimating the product of the transpose of the singular vector of the maximum singular value calculated by the singular value decomposition and the measurement vector as the disturbance generated in the controlled object, and use the measurement vector with the acquired measured values ​​as elements to estimate the disturbance generated in the controlled object.

Images