Non-linear pipeline system state estimation method and system based on unscented Kalman filtering and storage medium

Abstract

The invention discloses a nonlinear pipeline system state estimation method and system based on unscented Kalman filtering and a storage medium, relates to the field of network control, and aims to solve the problems of limited state estimation precision and high power consumption caused by signal attenuation, nonlinear characteristics and limited communication resources in the existing method. Comprising the following steps: 1, establishing an oil pipeline system dynamic model, and discretizing to obtain a state space model; 2, designing a dynamic event triggering mechanism, and constructing an amplification forwarding relay system to obtain amplification measurement signal data; step 3, setting initial values of state estimation and a covariance matrix; step 4, weighting 2n + 1 Sigma sampling points obtained by unscented transformation according to a weight coefficient to obtain one-step prediction and a one-step prediction error covariance Pj + 1j; 5, designing a filter according to a system phenomenon, calculating an estimation error, calculating an estimation error covariance upper bound matrix by using a matrix inequality, and solving a filtering gain matrix Kj + 1; and step 6, executing the step 4 and the step 5 until the total duration is reached.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network control, in particular to a nonlinear pipeline system state estimation method and system based on unscented Kalman filtering and a storage medium. BACKGROUND

[0002] With the continuous improvement of industrial automation level, the operation safety and state monitoring of nonlinear pipeline system as the core infrastructure of energy transportation such as oil and natural gas become increasingly prominent. The traditional periodic inspection and manual monitoring method has been difficult to meet the real-time and high efficiency requirements of modern industry. Therefore, the online monitoring technology based on state estimation becomes a key solution.

[0003] In recent years, the online monitoring technology based on state estimation has become a key technology to solve such problems, but the existing methods still have significant limitations: the traditional extended Kalman filter is prone to large estimation error due to the strong nonlinear characteristics of the pipeline system; although particle filter can handle nonlinear problems, its high computational complexity is difficult to adapt to real-time monitoring requirements; compared with the above, unscented Kalman filter (UKF) is very efficient and reliable in state estimation of nonlinear systems through unscented transformation; in addition, the existing researches often ignore the dynamic resource scheduling and channel attenuation problems in wireless communication network, and frequent data transmission leads to high energy consumption. The dynamic event triggering mechanism transmits data only when necessary by setting dynamic triggering conditions, thereby reducing redundant communication and network load. Long-distance wireless transmission of oil pipelines will cause signal attenuation, affecting data reliability, and amplify-and-forward relay improves transmission distance and quality by amplifying signals through relay nodes, but the noise and delay introduced by it need to be compensated in state estimation. Therefore, there is an urgent need for a UKF state estimation method that combines dynamic event triggering mechanism and amplify-and-forward relay to improve the monitoring reliability and energy efficiency of oil pipeline system in complex environments. SUMMARY

[0004] The technical problem to be solved by the present application is:

[0005] The existing method often leads to limited state estimation accuracy and high power consumption in long-distance wireless monitoring due to signal attenuation, nonlinear characteristics and limited communication resources.

[0006] The technical scheme adopted by the present application to solve the above technical problems is:

[0007] The present application provides a nonlinear pipeline system state estimation method based on unscented Kalman filtering, comprising the following steps:

[0008] Step one, establish a dynamic model of the oil pipeline system, and discretize to obtain a state space model;

[0009] Step two, design dynamic event triggering mechanism, build amplification forwarding relay system, get amplification measurement signal expression form;

[0010] Step three, set state estimation initial value and covariance matrix initial value

[0011] Step four, calculate 2n+1 Sigma sample points obtained by unscented transformation at j+1 time, wherein n is the dimension of state vector, calculate the transformed Sigma sample points after nonlinear transformation, and obtain one step prediction by weighting according to weight coefficient and one step prediction error covariance P j+1|j ;

[0012] Step five, design filter according to system phenomenon, calculate estimation error Calculate the upper bound matrix of estimation error covariance matrix by using matrix inequality, and solve the filter gain matrix K j+1 ;

[0013] Step six, judge whether the value of j+1 exceeds the total time length N, if not, execute step four and step five at next time, otherwise end.

[0014] Further, the process of establishing the dynamic model of the oil pipeline system in step one is as follows:

[0015] The momentum equation of the oil pipeline system is constructed as:

[0016]

[0017] Based on the momentum equation of the oil pipeline system, the continuity equation is constructed as:

[0018]

[0019] Wherein, H is the pressure head, Q is the flow, t represents time, z represents the position on the pipeline, D' is the inner diameter of the pipeline, A' is the cross-sectional area, g is the acceleration of gravity, f is the friction coefficient, and a is the pressure wave velocity;

[0020] The pipeline is divided into m segments, and the momentum equation and continuity equation of the oil pipeline system are converted into the following difference equation:

[0021]

[0022] Wherein

[0023]

[0024] Wherein, subscript i represents the cross-sectional position of the pipeline, subscript j represents the time, Q 1，j Is the inlet flow value at j time, Qi，j is the transient flow rate value of the pipeline at the i-th section and the j-th moment, Q m+1，j is the outlet flow value at time j, H i，j It is the transient pressure head value of the pipeline at the i-th section and the j-th moment.

[0025] Furthermore, in step 1, the oil pipeline system is discretized to obtain a state space model. The specific process is as follows:

[0026] Define the state variable x j =[H 2，j , H 3，j ,...,H m-1，j , H m，j , Q 1，j , Q 2，j ,...Q m，j , Q m+1，j ] T , input vector u j =[H 1，j , H m+1，j ] T , measurement output vector y j =[Q 1，j , Q m，j ] T , construct the state space model as:

[0027]

[0028] Among them, matrices A, B, C, and D are all known matrices with appropriate dimensions, and the process noise ω j and measurement noise v j is uncorrelated Gaussian white noise with zero mean, and the process noise ω j The variance of Q′ is j >0, measurement noise v j The variance is R j >0, l(x j ) is a nonlinear function, specifically:

[0029] Furthermore, in step 2, a dynamic event triggering mechanism is designed, including triggering conditions and triggering time; the triggering time is denoted as j t , define the event trigger function as:

[0030]

[0031] Define the trigger conditions:

[0032] t(d j , η j ,θ,ε)>0

[0033] wherein, ||·|| represents the Euclidean norm of a vector, η j is an internal dynamic variable, whose initial value is set as η0≥0; its updating rule is:

[0034] η j = λη j-1 -||d j ||+ε,

[0035] wherein ε and 0<λ<1 are given positive scalars, θ is also a given positive scalar, satisfying

[0036] The sequence of triggering instants 0≤j1<j2<…<j t <j t+1 <… is determined by:

[0037]

[0038] An amplify-and-forward relay system is constructed, in which the measurement signal sent by the sensor is transmitted to the estimator through the amplify-and-forward relay. Firstly, the signal received by the relay is :

[0039]

[0040] wherein subscript s represents the channel from the sensor to the relay, p s is the transmission energy, H s is a known diagonal matrix representing the channel coefficient from the sensor to the relay; is the channel transmission noise from the sensor to the relay, with a mean of zero and a variance of

[0041] After the signal is received by the relay, the signal is amplified and transmitted to the estimator again. The signal finally received by the estimator is :

[0042]

[0043] wherein subscript r represents the channel from the relay to the estimator, p is the amplification coefficient, p r is the transmission energy, H r is a known diagonal matrix representing the channel coefficient from the relay to the estimator, is the signal received by the relay, represents the channel transmission noise from the relay to the estimator, with a variance of

[0044] According to formula (4) and (5), Reconstructing as:

[0045]

[0046] wherein, is the measurement value transmitted at the triggering time,

[0047] Further, the step four comprises the following steps:

[0048] At the time j+1, the estimated value and the upper bound of the estimated error covariance Select 2n+1 Sigma points, denoted as

[0049]

[0050] wherein, n is the dimension of the random variable x, ρ is the proportion factor between the estimated value and the Sigma sampling point, defined as: ι is still the proportion factor, and ι≥0, α is the proportion scaling factor, and 0≤α≤1, defined is the matrix the κth column of the square root matrix obtained through Cholesky decomposition;

[0051] By using the nonlinear mapping l(·), the selected Sigma points are mapped to the one-step prediction

[0052]

[0053] Set the mean weight coefficient and the variance weight coefficient as:

[0054]

[0055] wherein β reflects the high-order characteristics of the state history information;

[0056] The weighted average operation is performed on the above transformed Sigma points, and the mean value and the variance of the one-step state prediction are respectively:

[0057]

[0058] Further, in the step five, the filter is designed according to the system phenomenon as:

[0059]

[0060] wherein is the state estimation value, j+1 ∈ [j t , j t+1 ), K j+1 is the estimator gain to be determined;

[0061] The one-step prediction error and the estimation error are calculated as:

[0062]

[0063] where x j+1 is the actual state value of the system at j+1 time;

[0064] According to formula (12), the estimation error is specifically:

[0065]

[0066] where I represents the unit matrix;

[0067] According to formula (13) and the definition of the estimation error covariance matrix, we get:

[0068]

[0069] where the subscript T represents the transpose of the matrix, P j+1|j is the one-step prediction error covariance matrix, K j+1 is the estimator gain to be determined, R j+1 is the variance of the measurement noise, and the following definitions are made:

[0070]

[0071] where is the variance of , and is the variance of ;

[0072] Based on the matrix inequality, the upper bound matrix of the estimation error covariance is calculated, and the filter gain matrix is optimized by minimizing the trace of the upper bound matrix;

[0073] Given three positive scalars γ j+1 , ξ1 and ξ2, let:

[0074]

[0075] satisfy is the upper bound of the estimation error covariance matrix, and P0 is the initial variance;

[0076] Further, through formula (15), we can get:

[0077]

[0078] In order to minimize the upper bound of the estimation error covariance matrix, based on the derivation result of formula (16), the estimator filtering gain matrix is constructed as:

[0079]

[0080] Wherein

[0081]

[0082] The application provides a nonlinear pipeline system state estimation system based on unscented Kalman filtering, which has program modules corresponding to the steps of the method described in any of the above technical solutions, and executes the steps in the nonlinear pipeline system state estimation method based on unscented Kalman filtering described above.

[0083] The application also provides a computer readable storage medium, which stores a computer program configured to realize the steps in the nonlinear pipeline system state estimation method based on unscented Kalman filtering described in any of the above technical solutions when called by a processor.

[0084] Compared with the prior art, the application has the following beneficial effects:

[0085] The application provides a nonlinear pipeline system state estimation method and system based on unscented Kalman filtering and a storage medium, which combines a dynamic event triggering mechanism with an amplification forwarding relay technology and embeds an unscented Kalman filtering framework, and by designing a suitable filtering gain, the trace of the upper bound thereof is minimized to solve the filtering gain. Compared with the existing state estimator design method, the application can reflect the actual working mode of the nonlinear system remote state estimator, can solve the dual challenges of real-time performance and reliability under complex working conditions, and can consider more realistic phenomena and has higher estimation accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0086] Figure 1 The nonlinear pipeline system state estimation flowchart based on unscented Kalman filtering in the embodiment of the application;

[0087] Figure 2 The state change curve x of the nonlinear oil pipeline system in the embodiment of the application 2，j and x 3，j ;

[0088] Figure 3 The upper bound curve of the estimation error covariance matrix in the embodiment of the application;

[0089] Figure 4The trigger time record under the dynamic time trigger mechanism in the embodiment of the application. DETAILED DESCRIPTION

[0090] In order to make the personnel in the technical field better understand the application scheme, the exemplary embodiments or examples of the application will be described in the following with reference to the drawings. Obviously, the described embodiments or examples are only a part of the embodiments or examples of the application, not all. Based on the embodiments or examples in the application, all other embodiments or examples obtained by the ordinary skilled in the art without making creative efforts should belong to the protection scope of the application.

[0091] In order to make the above-mentioned purposes, features and advantages of the application more obvious and easy to understand, the specific embodiments of the application will be described in detail below with reference to the drawings.

[0092] Specific implementation scheme one: combined with Figures 1 to 4 As shown in the figure, the application provides a nonlinear pipeline system state estimation method based on unscented Kalman filter, comprising the following steps:

[0093] Step one, establish the dynamic model of the oil pipeline system, and obtain the state space model after discretization;

[0094] Step two, design a dynamic event trigger mechanism, build an amplification forwarding relay system, and obtain an amplification measurement signal expression form;

[0095] Step three, set the initial value of state estimation and the initial value of the covariance matrix

[0096] Step four, calculate 2n+1 Sigma sampling points obtained by unscented transformation at j+1 time, wherein n is the dimension of the state vector, calculate the transformed Sigma sampling points after nonlinear transformation, and obtain one-step prediction and one-step prediction error covariance

[0097] Step five, design a filter according to the system phenomenon, and calculate the estimation error Calculate the estimation error covariance upper bound matrix using matrix inequality, and solve the filter gain matrix K j+1 ;

[0098] Step six, judge whether the j+1 value exceeds the total time length N, if not, execute step four and step five at the next time, otherwise end.

[0099] Specific implementation scheme two: the specific process of establishing the dynamic model of the oil pipeline system in step one is as follows:

[0100] The momentum equation of the oil pipeline system is constructed as follows:

[0101]

[0102] The continuity equation is constructed based on the momentum equation of the oil pipeline system as follows:

[0103]

[0104] wherein H is the pressure head, Q is the flow rate, t represents time, z represents the position on the pipeline, D' is the inner diameter of the pipeline, A' is the cross-sectional area, g is the acceleration of gravity, f is the friction coefficient, and a is the pressure wave velocity;

[0105] The momentum equation and the continuity equation of the oil pipeline system are converted into the following difference equations by dividing the pipeline into m segments:

[0106]

[0107] wherein

[0108]

[0109] wherein subscript i represents the cross-sectional position of the pipeline, subscript j represents the time, Q 1，j is the inlet flow rate value at time j, Q i，j is the flow rate value of the transient change of the pipeline at the i-th cross section and the j-th time, Q m+1，j is the outlet flow rate value at time j, H i，j is the pressure head value of the transient change of the pipeline at the i-th cross section and the j-th time. The other aspects of the embodiment are the same as those of Embodiment One.

[0110] Embodiment Three: The specific process of discretizing the oil pipeline system in Step One to obtain the state space model is as follows:

[0111] The state variable x is defined as j = [H 2，j , H 3，j ,..., H m-1 , j, H m，j , Q 1，j , Q 2，j ,...Q m，j , Q m+1，j ] T The input vector u is defined as j = [H 1，j , H m+1，j ] T The measurement output vector y is defined as j = [Q 1，j , Q m,j ] T The state space model is constructed as follows:

[0112]

[0113] Among them, matrices A, B, C, and D are all known matrices with appropriate dimensions, and the process noise ω j and measurement noise v j is uncorrelated Gaussian white noise with zero mean, and the process noise ω j The variance of Q′ is j >0, measurement noise v j The variance is R j >0, l(x j ) is a nonlinear function, specifically: The rest of this implementation plan is the same as the second specific implementation plan.

[0114] Specific implementation plan 4: In step 2, design a dynamic event trigger mechanism, including trigger conditions and trigger time; the trigger time is denoted as j t , define the event trigger function as:

[0115]

[0116] Define the trigger conditions:

[0117] t(d j , η j ,θ,ε)>0

[0118] in ||·|| represents the Euclidean norm of the vector; η j It is an internal dynamic variable, the initial value is set to η0≥0, and its update rule is:

[0119] η j =λη j-1 -||d j ||+ε,

[0120] Where ε and 0<λ<1 are given positive scalars; θ is also a given positive scalar, satisfying

[0121] Trigger time sequence 0≤j1<j2<…<j t <j t+1 <…is determined by the following formula:

[0122]

[0123] Construct an amplification and forwarding relay system. The measurement signal sent by the sensor is transmitted to the estimator through the amplification and forwarding relay. First, the signal received by the repeater for:

[0124]

[0125] where subscript s denotes the channel from sensor to repeater; p s is the transmitted energy; H s is a known diagonal matrix representing the channel coefficients from sensor to repeater; is the channel transmission noise from sensor to repeater with zero mean and variance

[0126] After the signal is received by the repeater, the signal is amplified and transmitted again to the estimator, the final received signal by the estimator is:

[0127]

[0128] where subscript r denotes the channel from repeater to estimator; is the amplification coefficient; p r is the transmitted energy; H r is a known diagonal matrix representing the channel coefficients from repeater to estimator; is the signal received by the repeater, denotes the channel transmission noise from repeater to estimator with variance

[0129] According to equations (4) and (5), reconstructed as:

[0130]

[0131] where, is the measurement transmitted at the triggering moment, The other aspects of the embodiment are the same as specific embodiment three.

[0132] Specific embodiment five: step four includes the following steps:

[0133] At j+1 moment, the estimated value and the upper bound of the estimated error covariance Select 2n+1 Sigma points recorded as

[0134]

[0135] where n is the dimension of the random variable x, p is the proportion factor between the estimated value and the Sigma sampling point, defined as: is a scaling factor, which only needs to be chosen to ensure the covariance is positive definite, and is a scaling factor, which controls the range of the Sigma point set, and is a matrix is the k-th column of the square root matrix obtained by Cholesky decomposition;

[0136] The selected Sigma points are mapped to the one-step prediction by using the nonlinear mapping l(·) according to the following rule

[0137]

[0138] The mean weight coefficient and the variance weight coefficient are set as

[0139]

[0140] where β reflects the high-order characteristics of the state history information, and adjusting β can improve the approximation accuracy of the covariance.

[0141] The weighted average operation is performed on the transformed Sigma points to obtain the mean and variance of the one-step state prediction, respectively

[0142]

[0143] The other aspects of the embodiment are the same as those of Embodiment Four.

[0144] Embodiment Six: In step five, the filter is designed according to the system phenomenon, and is

[0145]

[0146] where is the state estimate value, j+1 ∈ [j t , j t+1 ), K j+1 is the estimator gain to be determined;

[0147] The one-step prediction error and the estimation error are calculated as

[0148]

[0149] where x j+1 is the actual state value of the system at j+1 time;

[0150] The estimation error is obtained according to formula (12) as

[0151]

[0152] where I denotes the identity matrix;

[0153] According to formula (13) and the definition of the estimation error covariance matrix,

[0154]

[0155] where the superscript T represents the transpose of a matrix, P j+1|j is the one-step prediction error covariance matrix, K j+1 is the estimator gain to be determined, R j+1 is the variance of the measurement noise, and for ease of representation, and are intermediate variables, and are defined as follows:

[0156]

[0157] where is the variance of is the variance of

[0158] Based on the matrix inequality, the estimation error covariance upper bound matrix is calculated, and the filter gain matrix is optimized by minimizing the trace of the upper bound matrix;

[0159] Given three positive scalars γ j+1 , ξ1, and ξ2, let:

[0160]

[0161] satisfy is the upper bound of the estimation error covariance matrix, and P0 is the initial variance;

[0162] Further, by formula (15), we have:

[0163]

[0164] To minimize the upper bound of the estimation error covariance matrix, based on the derivation result of formula (16), the estimator filter gain matrix is constructed as:

[0165]

[0166] For ease of representation, Ω j+1 and Δ j+1 are intermediate variables, and are defined as:

[0167]

[0168] ​​​The other aspects of the embodiment are the same as those of embodiment five.

[0169] The nonlinear pipeline system state estimation method (algorithm) based on the unscented Kalman filter is the underlying technical core of the present application, and various products can be derived based on the algorithm.

[0170] The method proposed in the present application develops a nonlinear pipeline system state estimation system based on the unscented Kalman filter using a programming language, which has program modules corresponding to the steps of the above technical solutions, and executes the steps in the above nonlinear pipeline system state estimation method based on the unscented Kalman filter when running.

[0171] The computer program of the developed system (software) is stored on a computer readable storage medium, and the computer program is configured to realize the steps of the above nonlinear pipeline system state estimation method based on the unscented Kalman filter when called by a processor. That is, the present application is materialized on a carrier to become a computer program product.

[0172] Various embodiments of the systems and techniques described herein can be realized in digital electronic circuitry, integrated circuitry, specially designed ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0173] The computing programs (also referred to as programs, software, software applications, or code) in the present application include machine instructions of a programmable processor, and can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., magnetic disks, optical disks, memories, programmable logic devices PLD) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0174] The advantages of the present application will be described below in conjunction with specific embodiments.

[0175] Example 1

[0176] The parameters of the oil pipeline system are selected as follows:

[0177] x j+1 =Ax j +Bu j +Dl(x j )+ω j ,

[0178] y j =Cx j +v j ,

[0179] in

[0180]

[0181] In addition, the initial value and initial variance of the state are x 0|0 =[535 0.8 0.8 0.8] T and The parameters of the event trigger part are selected as ε=0.06,θ=6,and the internal dynamic variable η j The initial value of is η0 = 1.5, and the parameter of the amplify-and-forward relay part is H s =0.5, H r =0.15, p s =1.5, p r =1.5, a=1. The remaining parameters are selected as Q j =0.1, R j =0.3. The root mean square error (MSE) is defined as: M is the number of independent experiments and is set to 300.

[0182] The simulation effect is as follows Figures 2 to 4 As shown, Figure 2 It shows the status of the oil pipeline system and the corresponding estimated value. It can be seen from the figure that the estimation method proposed by the present invention is relatively effective. In addition, due to the existence of dynamic event triggering mechanism and amplification and forwarding relay in the system, the estimated value may fluctuate slightly at certain time points, but the overall stability is good. Figure 3 Indicates the minimum upper bound of the estimation error covariance and the trajectory of the root mean square error of the estimation scheme. It can be seen that the estimation error covariance matrix There is an upper bound. Figure 4 As shown in Figure 2, the triggering time of dynamic events is recorded. The triggering time density is high at first and then low, which shows that the dynamic event triggering mechanism effectively reduces unnecessary communication burden. In summary, for the oil pipeline system with nonlinear characteristics, the invented state estimation method is effective and feasible.

[0183] Although the present application has been disclosed with reference to the above embodiments, the scope of the present application is not limited to the above. Various changes and modifications can be made to the present application without departing from the spirit and scope of the present application, and such changes and modifications are intended to fall within the scope of the present application.

Claims

1. A nonlinear pipeline system state estimation method based on unscented Kalman filtering, characterized in that: The following steps are involved: Step 1: Establish a dynamic model of the oil pipeline system and discretize it to obtain a state space model; Step 2: Design a dynamic event trigger mechanism, build an amplification and forwarding relay system, and obtain the expression form of the amplified measurement signal; Step 3: Set the initial value of state estimation and initial values ​​of the covariance matrix Step 4: Calculate the 2n+1 Sigma sampling points obtained by the untraceable transformation at time j+1, where n is the dimension of the state vector. Calculate the transformed Sigma sampling points through nonlinear transformation and weight them according to the weight coefficient to obtain the one-step prediction. and the one-step forecast error covariance P j+1|j ; Step 5: Design the filter based on the system phenomenon and calculate the estimated error Use matrix inequality to calculate the upper bound matrix of the estimation error covariance and solve the filter gain matrix K j+1 ; Step 6: Determine whether the value of j+1 exceeds the total duration N. If not, execute steps 4 and 5 at the next moment; otherwise, end.

2. The nonlinear pipeline system state estimation method based on unscented Kalman filtering according to claim 1 is characterized in that: The dynamic model of the oil pipeline system is established in step 1. The specific process is as follows: The momentum equation of the oil pipeline system is constructed as follows: Based on the momentum equation of the oil pipeline system, the continuity equation is constructed: Where H is the pressure head, Q is the flow rate, t is the time, z is the position on the pipe, D′ is the inner diameter of the pipe, A′ is the cross-sectional area, g is the acceleration due to gravity, f is the friction coefficient, and a is the pressure wave velocity; Divide the pipeline into m segments and transform the momentum equation and continuity equation of the oil pipeline system into the following difference equations: in Among them, the subscript i represents the cross-sectional position of the pipeline, the subscript j represents the time, and Q 1，j is the inlet flow value at time j, Q i，j is the transient flow rate value of the pipeline at the i-th section and the j-th moment, Q m+1，j is the outlet flow value at time j, H i，j It is the transient pressure head value of the pipeline at the i-th section and the j-th moment.

3. The nonlinear pipeline system state estimation method based on unscented Kalman filtering according to claim 2 is characterized in that: In step 1, the oil pipeline system is discretized to obtain a state space model. The specific process is as follows: Define the state variable x j =[H 2，j , H 3，j ,...,H m-1 ,j,H m，j , Q 1，j , Q 2，j ,...Q m，j , Q m+1，j ] T , input vector u j =[H 1，j , H m+1，j ] T , measurement output vector y j =[Q 1，j , Q m，j ] T , construct the state space model as: Among them, matrices A, B, C, and D are all known matrices with appropriate dimensions, and the process noise ω j and measurement noise v j is uncorrelated Gaussian white noise with zero mean, and the process noise ω j The variance of Q′ is j >0, measurement noise v j The variance is R j >0, l(x j ) is a nonlinear function, specifically:

4. The nonlinear pipeline system state estimation method based on unscented Kalman filtering according to claim 3 is characterized in that: In step 2, a dynamic event triggering mechanism is designed, including triggering conditions and triggering time; the triggering time is denoted as j t , define the event trigger function as: Define the trigger conditions: t(d j ,or j ,θ,ε)>0 in, ||·|| represents the Euclidean norm of the vector, η j is an internal dynamic variable, and its initial value is set to η0 ≥ 0; The update rule is: or j =all j-1 -||d j ||+e, Where ε and 0<λ<1 are given positive scalars, and θ is also a given positive scalar, satisfying Trigger time sequence 0≤j1<j2<…<j t <j t+1 <…is determined by the following formula: Construct an amplification and forwarding relay system. The measurement signal sent by the sensor is transmitted to the estimator through the amplification and forwarding relay. First, the signal received by the repeater for: Where, the subscript s represents the channel from the sensor to the repeater, and p s is the transferred energy, H s is a known diagonal matrix representing the channel coefficients from the sensor to the repeater; is the channel transmission noise from the sensor to the repeater, with a mean of zero and a variance of After the repeater receives the signal, it is amplified and transmitted again to the estimator, which ultimately receives the signal for: Where, the subscript r represents the channel from the repeater to the estimator, is the amplification factor, p r is the transferred energy, H r is a known diagonal matrix representing the channel coefficients from the repeater to the estimator channel, is the signal received by the repeater. represents the channel transmission noise from the repeater to the estimator, and its variance is According to formulas (4) and (5), Restructured to: in, is the measured value transmitted at the trigger moment, 5. The nonlinear pipeline system state estimation method based on unscented Kalman filtering according to claim 4 is characterized in that: Step 4 includes the following steps: At time j+1, the estimated value and an upper bound on the estimated error covariance Select 2n+1 Sigma points and denote them as Where n is the dimension of the random variable x, and ρ is the estimated value of the random variable x. The proportional factor of the distance between the sampling point and the Sigma sampling point is defined as: ι is still the scale factor, and ι ≥ 0, α is the scaling factor, and 0 ≤ Q ≤ 1, define is a matrix The k-th column of the square root matrix obtained by Cholesky decomposition; By using the nonlinear mapping l(·), the selected Sigma points are mapped to the one-step prediction according to the following rules Set the mean weight coefficient and variance weight coefficient for: Among them, β reflects the high-order characteristics of state history information; Performing weighted average calculation on the transformed Sigma points above, the mean and variance of the one-step state prediction are:

6. The nonlinear pipeline system state estimation method based on unscented Kalman filtering according to claim 5 is characterized in that: In step 5, the filter is designed according to the system phenomenon as follows: in is the state estimate, j+1∈[j t ,j t+1 ), K j+1 is the estimator gain that needs to be determined; The one-step prediction error and estimation error are calculated as follows: where x j+1 is the actual state value of the system at time j+1; According to formula (12), the estimated error is: Where I represents the identity matrix; According to formula (13) and the definition of the estimation error covariance matrix, get: Among them, the subscript T represents the transpose of the matrix, P j+1|j is the one-step forecast error covariance matrix, K j+1 is the estimator gain that needs to be determined, R j+1 is the variance of the measurement noise and is defined as follows: in yes The variance of yes variance; Calculate the upper bound matrix of the estimation error covariance based on matrix inequality, and optimize the filter gain matrix by minimizing the trace of the upper bound matrix; Given γ j+1 , ξ1 and ξ2 are three positive scalars, let: satisfy is the upper bound of the estimated error covariance matrix, P0 is the initial variance; Furthermore, through formula (15), we can get: In order to minimize the upper bound of the estimation error covariance matrix, based on the derivation of formula (16), the estimator filter gain matrix is ​​constructed as: in 7. A nonlinear pipeline system state estimation system based on unscented Kalman filtering, characterized in that: The system has a program module corresponding to the steps of the method described in any one of claims 1 to 6, and executes the steps of the nonlinear pipeline system state estimation method based on unscented Kalman filtering when running.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of the nonlinear pipeline system state estimation method based on unscented Kalman filtering according to any one of claims 1 to 6 when called by a processor.

Images