Aircraft fuel oil measuring method based on multi-sensor multi-measurement-value dynamic fusion

Abstract

The invention relates to the technical field of aircraft fuel measurement, and discloses a multi-sensor multi-measurement-value dynamic fusion aircraft fuel measurement method, which comprises the following steps of: acquiring a fuel quantity measurement value falling into a confidence interval from original fuel quantity measurement data acquired by a plurality of sensors through a software filter; grouping the multiple oil mass measurement values according to the sensors to which the multiple oil mass measurement values belong, sequentially numbering the multiple oil mass measurement values in the groups, and sequentially numbering the multiple sensors; a fusion filtering algorithm is optimized in a mode of combining a weighted average method and a Lagrange interpolation method, and a multi-sensor multi-measurement-value dynamic fusion model is constructed; and then, iterating the input fuel quantity measured value through the multi-sensor multi-measured-value dynamic fusion model to obtain a multi-sensor multi-measured-value fusion result which is used for representing a predicted value of the aircraft fuel quantity. According to the method, a fusion filtering algorithm is optimized in a mode of combining a weighted average method and a Lagrange interpolation method, a multi-sensor multi-measurement-value dynamic fusion model is constructed, a model output result is closer to a real state, and the measurement precision is improved.

Description

Technical Field

[0001] The invention relates to the technical field of aircraft fuel measurement, in particular to an aircraft fuel measurement method with dynamic fusion of multiple sensors and multiple measurement values. Background Art

[0002] At present, the fuel measurement system based on computer technology has greatly improved the accuracy of fuel measurement, but it has brought about the problem of poor measurement stability. In view of the core problem of multi-source data fusion in multi-sensor networks, it is not universal in practical applications and lacks fault tolerance. In the case of erroneous output of fuel data, it is possible to obtain erroneous fusion results, or even output fuel data that is far from the actual situation, which has seriously affected the pilot's correct judgment of the aircraft's working status. Moreover, in the process of fuel data fusion, if the waiting time is too short, it will cause too many invalid fusion times, data discreteness and low value density. If the waiting time is too long, the transmission delay will increase, which will seriously affect the reliability and effectiveness of the analysis results. Due to the uncertainty, incompleteness and instability of sensor information, the error correction ability of the fuel data fusion process is also very low, resulting in a slow convergence speed of the software filter, reduced estimation accuracy, and poor accuracy of the original data obtained, which in turn leads to inherent defects such as suboptimal filter estimation or filter divergence, inaccurate initial value estimation, and hysteresis of fuel measurement data.

[0003] The Chinese invention patent with the patent publication number CN116776285A discloses a weighted fusion method for multi-sensor fuel measurement data based on confidence intervals, and performs weighted processing on each sensor through a general weighted average data fusion filter algorithm. In order to further reduce the error, the present invention optimizes the design based on the above scheme, improves the previous weighted average data fusion filter algorithm, and then performs Lagrangian interpolation on the fuel measurement value of a single sensor to establish a multi-sensor multi-measurement dynamic fusion model based on Lagrangian interpolation, so that the model output result is closer to the real state. Summary of the invention

[0004] The present invention provides an aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion on the basis of the prior art. The fusion filtering algorithm is optimized by combining the weighted average method with the Lagrange interpolation method, and a multi-sensor multi-measurement dynamic fusion model is constructed. The model output result is closer to the real state, and can overcome the measurement error caused by the multi-sensor multi-measurement fusion during maneuvering flight, further improve the fuel measurement accuracy, and have a more superior prediction performance.

[0005] The present invention is implemented through the following technical scheme: a multi-sensor multi-measurement dynamic fusion aircraft fuel measurement method, a software filter for post-processing fuel quantity signals is set, and the fuel quantity measurement values ​​falling within the confidence interval are obtained from the original fuel quantity measurement data collected by multiple sensors through the software filter; after obtaining the fuel quantity measurement values ​​falling within the confidence interval, the multiple fuel quantity measurement values ​​are grouped according to the sensors to which they belong and are numbered in sequence within the group, and the multiple sensors are numbered in sequence at the same time; then, the fusion filtering algorithm is optimized by combining the weighted average method with the Lagrange interpolation method, and a multi-sensor multi-measurement dynamic fusion model is constructed; then, the input fuel quantity measurement values ​​are iterated through the multi-sensor multi-measurement dynamic fusion model, and the fusion results of the multi-sensor multi-measurement are obtained, which are used to characterize the predicted value of the aircraft fuel quantity.

[0006] Further, in order to better illustrate the present invention, when constructing a multi-sensor multi-measurement dynamic fusion model, an expression for the oil quantity measurement value of a single sensor and an a-order Lagrangian interpolation expression for the single oil quantity measurement value are constructed respectively, and the measurement variance expression of the single sensor is calculated by the Lagrangian interpolation method, which is used to calculate the measurement variance of each sensor; then, based on the measurement variance of the single sensor, a weight expression of the single sensor is constructed, which is used to calculate the fusion weight of each sensor; then, based on the fusion weight of the single sensor and the oil quantity measurement value of the single sensor, a fusion result expression is constructed by the weighted average method, which is used to calculate the fusion result of the multi-sensor multi-measurement, and obtain the predicted value of the aircraft fuel quantity.

[0007] Compared with the prior art, the present invention has the following advantages and beneficial effects.

[0008] (1) The present invention uses the output value of the software filter to post-process the fuel quantity signal. On the basis of the general weighted average data fusion filtering algorithm, the calculation method of the weight factor of each sensor is improved, and a multi-sensor multi-measurement dynamic fusion mathematical model based on Lagrange interpolation is established. This can overcome the measurement error caused by the fusion of multi-sensor multi-measurement during maneuvering flight and further improve the fuel measurement accuracy.

[0009] (2) The present invention is a new fuel measurement method based on weighted fusion of multiple measurements of multiple sensors of an aircraft under a confidence interval, which can better estimate the fuel data change signal and is suitable for situations where higher estimation accuracy of multiple sensors is required.

[0010] (3) The aircraft fuel measurement method disclosed in the present invention can eliminate the error of oil level fluctuation, overcome the limitation that general software filters do not consider the situation of aircraft attitude changes and the stability of fuel measurement data, further improve the fuel measurement accuracy, and have more superior performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1The main flow chart of the aircraft fuel measurement method with dynamic fusion of multiple sensors and multiple measurements provided by the present invention. DETAILED DESCRIPTION

[0012] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0013] Embodiment 1:

[0014] In the prior art, data redundancy in multi-sensor networks includes four situations:

[0015] (1) When the measurement value appears, multiple sensors within the confidence interval sample the same data;

[0016] (2) The same sensor samples the same data in different time periods, that is, the sampling target does not change, and the oil volume sampling data and the sent oil volume data overlap;

[0017] (3) In different time periods, the data sampled by the same sensor changes, but the data changes slightly and the target features do not change;

[0018] (4) Multiple sensors sample different data on the target in different time periods.

[0019] First, for the first two cases, in order to reduce the delay caused by complex calculations, when the same data is sampled for the same target, only one sensor needs to send data, and the other nodes do not send data. The algorithm is as follows: After the sensor collects the detection data, it checks the neighbor node information and compares the sampled data. If the data is the same, one sensor sends a data packet; the other sensors do not send; after the new sampled data enters, it overwrites the original sampled data; if the data collected by the sensor at different times is the same, the data packet only changes the timestamp, and the data is not sent.

[0020] Secondly, for the latter two cases, the technical solution disclosed in the Chinese invention patent with patent publication number CN116776285A, on the one hand, uses a software filter to post-process the fuel quantity signal, and on the other hand, calculates the fuel quantity through a weighted average data fusion algorithm to reduce data redundancy in the multi-sensor network.

[0021] However, compared with the prior art which only uses the weighted average data fusion algorithm, the system has weaker followability.

[0022] Therefore, this embodiment optimizes the existing fusion filter algorithm based on the technical concept of combining the weighted average method with the Lagrange interpolation method, thereby improving the calculation method of the weight factor of each sensor, and then establishing a multi-sensor multi-measurement dynamic fusion model based on Lagrange interpolation. The optimized multi-sensor multi-measurement dynamic fusion model can use the model algorithm to iteratively obtain the result of multi-sensor multi-measurement fusion according to the input of each sensor measurement value, realize the trend estimation of the aircraft fuel quantity measurement value, make it closer to the real state, and thus improve the aircraft fuel measurement accuracy.

[0023] Specifically, this embodiment provides an aircraft fuel measurement method with dynamic fusion of multiple sensors and multiple measurements, and sets a software filter for post-processing fuel quantity signals, and obtains fuel quantity measurement values ​​falling within a confidence interval from original fuel quantity measurement data collected by multiple sensors through the software filter; after obtaining the fuel quantity measurement values ​​falling within the confidence interval, the multiple fuel quantity measurement values ​​are grouped according to the sensors to which they belong and are numbered in sequence within the group, and the multiple sensors are numbered in sequence at the same time; the fusion filtering algorithm is then optimized by combining the weighted average method with the Lagrange interpolation method, and a dynamic fusion model of multiple sensors and multiple measurements is constructed; then, the input fuel quantity measurement values ​​are iterated through the dynamic fusion model of multiple sensors and multiple measurements, and a fusion result of the multiple sensors and multiple measurements is obtained, which is used to characterize the predicted value of the aircraft fuel quantity.

[0024] Compared with the technical solution disclosed in the Chinese invention patent with patent publication number CN116776285A, this embodiment is a new fuel measurement method based on weighted fusion of multi-sensor multi-measurement values ​​of aircraft under confidence interval. The optimized algorithm model can better estimate the fuel data change signal and is suitable for occasions with higher estimation accuracy of multiple sensors.

[0025] Embodiment 2:

[0026] This embodiment is described in detail based on Embodiment 1.

[0027] The aircraft fuel measurement method, such as Figure 1 As shown, it includes steps S1 to S3.

[0028] Step S1, obtaining the oil quantity measurement value;

[0029] Specifically, a software filter is set to post-process the fuel quantity signal, and the fuel quantity measurement value falling within the confidence interval is obtained from the original fuel quantity measurement data collected by multiple sensors through the software filter.

[0030] Step S2, constructing a multi-sensor multi-measurement dynamic fusion model;

[0031] Specifically, after obtaining the oil quantity measurement value that falls within the confidence interval, multiple oil quantity measurement values ​​are grouped according to the sensors they belong to and numbered in sequence within the group, and multiple sensors are numbered in sequence at the same time; then, the fusion filtering algorithm is optimized by combining the weighted average method with the Lagrange interpolation method to construct a multi-sensor multi-measurement dynamic fusion model.

[0032] Step S3, obtaining a predicted value of the amount of aircraft fuel;

[0033] Specifically, it means that the input fuel quantity measurement value is iterated through the multi-sensor multi-measurement dynamic fusion model to obtain the fusion result of the multi-sensor multi-measurement, which is used to characterize the predicted value of the aircraft fuel quantity.

[0034] The other parts of this embodiment are the same as those of Embodiment 1, and thus will not be described in detail.

[0035] Embodiment 3:

[0036] This embodiment describes step S1 in detail based on embodiment 2.

[0037] Step S1 specifically includes the following steps.

[0038] Step S11, setting a software filter for post-processing the fuel quantity signal.

[0039] Step S12: setting the filter queue address, the confidence interval length, the effective range of the oil quantity measurement value, the upper limit of the filter queue and the lower limit of the filter queue in the software filter.

[0040] Step S13, using the effective range of the oil quantity measurement value, the upper limit of the filter queue, and the lower limit of the filter queue to judge each fuel quantity measurement value of each sensor, analyze the credibility relationship between the oil quantity data collected by multiple sensors, and obtain the oil quantity measurement value.

[0041] The filtering queue address is used to define the storage start position of the collected data.

[0042] The confidence interval length is used to define the response speed and interval length of the filter.

[0043] The effective range of the oil quantity measurement value is also called the filtering limit Δ limit , used to define the value range of the aircraft fuel measurement data sampling value; and judge the false aircraft fuel measurement sampling data according to the value range.

[0044] The filter queue upper limit Δ upper and filter queue lower limit Δ lower , used to determine the confidence interval range of the filtering queue, and distinguish the credibility of the aircraft fuel measurement data sampling value through the confidence interval range.

[0045] Since the actual fuel consumption of the aircraft is reduced at a relatively stable fuel consumption rate, the concept of reliable measurement value is introduced here: the reliable measurement value changes within a certain range. When it changes downward, the reliable measurement value relative to the reference fuel volume is not less than Δ lower , when it changes upward, the maximum value cannot exceed Δ upper .

[0046] That is to say, if the fuel quantity sampling value of a certain aircraft falls within [Δ lower ,Δ upper ] interval, it is considered to be credible, and multiple measurement cycles are repeated until the filter queue is filled, and the aircraft fuel quantity sampling values ​​falling within the confidence interval are taken as credible measurement values, i.e., fuel quantity measurement values. If an extremely large value is sampled, it is considered to be invalid and is removed.

[0047] In summary, the working principle of the software filter is as follows: After the software filter starts running, the measured value must be filled up after n measurement cycles to serve as the basis for subsequent comparisons. On the basis of the queue being filled up, the filter limit Δ limit Remove obvious false data from the sampled data and exclude obvious false data. Then use the filter queue upper limit Δ upper and filter queue lower limit Δ lower The sampled data is distinguished for credibility to obtain a credible measurement value; finally, the software filter outputs the credible measurement value.

[0048] Therefore, the oil volume measurement values ​​input into the multi-sensor multi-measurement dynamic fusion model for post-processing all meet the following conditions: less than the filtering limit Δ limit And in [Δ lower ,Δ upper ] interval, that is, within the confidence interval.

[0049] The other parts of this embodiment are the same as those of Embodiment 2, and thus will not be described in detail.

[0050] Embodiment 4:

[0051] This embodiment describes step S2 in detail based on Embodiment 2 and Embodiment 3.

[0052] Step S2 specifically includes the following steps.

[0053] Step S21, after obtaining the oil quantity measurement values ​​falling within the confidence interval, multiple oil quantity measurement values ​​are grouped according to the sensors to which they belong and are numbered sequentially within the group, and multiple sensors are numbered sequentially at the same time.

[0054] Assumption 1: A total of n sensors are set in the fuel measurement system; the n sensors are numbered in sequence and the sensor number is recorded as k.

[0055] Assumption 2: There are m oil level measurement values ​​corresponding to a single sensor; the m oil level measurement values ​​are numbered in sequence and i, j, and s are used to represent individual oil level measurement values ​​with different single sensor number values.

[0056] Based on the above assumptions, the oil volume measurement values ​​collected by the same sensor are grouped into n groups. At this time, the oil volume measurement value of the kth sensor is recorded as d k , k=1,2,…,n.

[0057] The oil level measurement value d of the kth sensor k The mathematical expression is:

[0058] d k ={d1,...d k ,...d n},k=1,2,....n (1)

[0059] Where d is the general symbol for the set of all oil quantity measurement values ​​of a single sensor, and its subscript represents the number value of the single sensor, that is, d1 represents the oil quantity measurement value of the first sensor; d k represents the oil volume measurement value of the kth sensor; d n Indicates the oil level measurement value of the nth sensor.

[0060] The oil level measurement value d of the kth sensor k ,satisfy:

[0061] d k =D+ε k (2)

[0062] Among them, d k Satisfy E(d k )=D, and are independent of each other. E(·) represents the mean function.

[0063] D represents the true value of the oil level measurement corresponding to the kth sensor.

[0064] ε k is the random error of the kth sensor during measurement, which is independent of each other and follows a normal distribution represents the measurement variance of n sensors;

[0065] The fusion result expression is as follows:

[0066]

[0067] W k satisfy:

[0068]

[0069] in, Indicates the fusion result of multiple sensor measurements.

[0070] d k Represents the oil level measurement of a single sensor.

[0071] W k Represents the fusion weight of a single sensor.

[0072] k represents the sensor number, k = 1, 2, ..., n; n represents the number of sensors; d k , W k The value of the subscript k represents the number of the sensor.

[0073] The total mean square error expression of the fusion result is as follows:

[0074]

[0075] Among them, J represents the total mean square error of the fusion result.

[0076] E(·) represents the mean function.

[0077] D represents the true value of the oil level measurement corresponding to the kth sensor.

[0078] Indicates the fusion result of multiple sensor measurements.

[0079] d i Represents the fusion result corresponding to the i-th oil level measurement value of a single sensor.

[0080] W i Represents the fusion weight corresponding to the i-th oil level measurement value of a single sensor.

[0081] σ i Represents the measurement variance corresponding to the i-th oil level measurement value of a single sensor.

[0082] The fusion weight expression of a single sensor is as follows:

[0083]

[0084] Among them, W k Represents the fusion weight of a single sensor, specifically the fusion weight of the kth sensor.

[0085] It represents the measurement variance of a single sensor, specifically the measurement variance of the kth sensor.

[0086] k represents the number of the sensor, k=1, 2, ..., n; n represents the number of sensors.

[0087] In order to eliminate the influence of errors and maximize the estimation accuracy of the measurement value detected by a single sensor, the measurement variance of the sensor is estimated by interpolating the oil volume measurement value of a single sensor without the need to use the measurement values ​​of other sensors. According to formula (2), suppose a certain sensor measures a total of m oil volume measurement values, and the oil volume measurement value is:

[0088] d(x i )=D(x i )+ε(x i ) (7)

[0089] Among them, d(x i ) represents the oil level measurement value of a single sensor, specifically, the fusion measurement value corresponding to the i-th oil level measurement value of a single sensor. i represents the i-th oil volume measurement value corresponding to a single sensor. Here, d(x i ) and x i Although both are parameters that indicate the oil quantity measurement value, d(x i ) is a single sensor as the description object, x i The single oil volume measurement value sampled by the sensor is used as the description object, so d(x i ) is equivalent to the set, x i Equivalent to an element in a collection.

[0090] D(x i ) represents the true value of the i-th oil level measurement of a single sensor.

[0091] ε(x i ) represents the random error of the i-th oil level measurement value of a single sensor.

[0092] d(x i )、D(x i ),ε(x i ) i Is the parameter symbol representing a single oil quantity measurement value, x i The subscript i is a number used to represent a single oil quantity measurement value, i=1, 2, ..., m.

[0093] In the present invention, i 、x j 、x s Similarly, when the subscript is not assigned a specific value, it is used as a general expression of a single oil level measurement value; when the subscript is assigned a specific value, it is used as the oil level measurement value corresponding to a specific sensor number value. For example, when x i When the subscript i is not assigned a specific value, x i As a general expression of a single oil quantity measurement; when x iWhen the subscript i is assigned a specific value, x i Represents the i-th oil level measurement value of a single sensor.

[0094] According to the Lagrange interpolation theorem, the a-order Lagrange interpolation of the i-th oil volume measurement value of a single sensor is obtained, and the expression is as follows:

[0095]

[0096] Among them, L(x i ) represents the a-order Lagrangian interpolation of the i-th oil volume measurement value of a single sensor. a represents the order set in the Lagrangian interpolation method, a is an even number, and a <m。

[0097] x i 、x j It represents the i-th oil quantity measurement value and the j-th oil quantity measurement value corresponding to a single sensor. The values ​​of i and j both represent the serial values ​​of the oil quantity measurement values ​​in a single sensor, i=1, 2, ..., m, j=1, 2, ..., m, i≠j; m represents the number of oil quantity measurement values ​​collected by a single sensor.

[0098] d(x j ) represents the fused measurement value corresponding to the j-th oil level measurement value of a single sensor.

[0099] ω(x j ) represents the difference ratio coefficient between two sets of measurements within a single sensor.

[0100] The difference ratio coefficient ω(x j ) is a (a-1)th order polynomial, expressed as follows:

[0101]

[0102] Among them, x i 、x j 、x s represents the i-th oil level measurement value, j-th oil level measurement value, and s-th oil level measurement value corresponding to a single sensor. The values ​​of i, j, and s all represent the serial number of the oil level measurement values ​​in a single sensor, i = 1, 2, ..., m, j = 1, 2, ..., m, s = 1, 2, ..., m, i ≠ j ≠ s; m represents the number of oil level measurement values ​​collected by a single sensor. The difference x in formula (9) is i -x s and x j -x s Respectively represent the difference between two sets of oil volume measurement values ​​of a certain sensor.

[0103] By using equations (7) and (8), the difference between the oil volume measurement value and the interpolation value measured by a certain sensor can be calculated. At this time, the difference between the fusion measurement value and the Lagrange interpolation value of a single oil volume measurement value is expressed as follows:

[0104]

[0105] Among them, Δ(x i ) represents the difference between the fusion measurement value and the Lagrange interpolation value of the i-th oil level measurement of a single sensor.

[0106] d(x i ) represents the fused measurement value corresponding to the i-th oil level measurement value of a single sensor.

[0107] L(x i ) represents the a-order Lagrangian interpolation of the i-th oil level measurement value of a single sensor.

[0108] D(x i ) represents the true value of the i-th oil level measurement of a single sensor.

[0109] ε(x i ) represents the random error of the i-th oil level measurement value of a single sensor.

[0110] D(x j ) represents the true value of the j-th oil level measurement of a single sensor.

[0111] ε(x j ) represents the random error of the jth oil level measurement value of a single sensor.

[0112] ω(x j ) represents the difference ratio coefficient between two sets of measurements within a single sensor.

[0113] According to formula (10), Δ(x i ) is as follows:

[0114]

[0115] Among them, Var[Δ(x i )] represents Δ(x i ) is the variance of the equation; Var(·) represents the variance function.

[0116] Represents the variance of the oil volume value measured by a single sensor.

[0117] According to formula (11), starting from the definition of statistical variance, we can get Δ(x i ) is expressed as follows:

[0118]

[0119] In actual calculations, the boundary measurements of sensor sampling cannot be used to calculate the variance, and the a-order Lagrangian difference algorithm requires interpolation points on both sides. data, so the number of interpolated measurements that can be used to calculate the variance is (ma-2); and because d(x i ) and L(x i ) are both unbiased estimates of the oil volume measurement value, so the mathematical expectation of the two is equal to the oil volume measurement value, so we can get Δ(x i ) is as follows:

[0120] E[Δ(x i )]=E[d(x i )]-E[L(x i )]=D(x i )-D(x i )=0 (13)

[0121] Substituting formula (13) into formula (12), we can get:

[0122]

[0123] By comparing equation (14) and equation (11) about Δ(x i ) variance calculation, the variance of a single sensor can be obtained The values ​​are:

[0124]

[0125] Substituting formula (15) into formula (6), the weight corresponding to each sensor can be calculated.

[0126] Finally, the weighted fusion result of multi-sensor multi-measurement of the aircraft is obtained by combining formula (3) with the weights corresponding to each sensor obtained by the above derivation. The fusion result of multiple sensor measurements is the predicted value of the aircraft fuel quantity.

[0127] In this embodiment, formula (3) provides an expression for the fusion result of multiple sensor multi-measurement; formula (6) provides an expression for the weight of each sensor; formula (8) provides an expression for constructing an a-order Lagrangian interpolation of the i-th oil value of a certain sensor; the weight of each sensor can be calculated according to formula (9)-formula (15), and then it can be substituted into formula (3) to obtain the fusion result of multiple sensor multi-measurement.

[0128] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification or equivalent change made to the above embodiment based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. An aircraft fuel measurement method with dynamic fusion of multiple sensors and multiple measurements, wherein a software filter is provided to post-process the fuel quantity signal, and the fuel quantity measurement value falling within the confidence interval is obtained from the original fuel quantity measurement data collected by multiple sensors through the software filter, characterized in that: After obtaining the fuel quantity measurement values ​​that fall within the confidence interval, multiple fuel quantity measurement values ​​are grouped according to the sensors they belong to and numbered in sequence within the group, and multiple sensors are numbered in sequence at the same time; then, the fusion filtering algorithm is optimized by combining the weighted average method with the Lagrange interpolation method, and a multi-sensor multi-measurement dynamic fusion model is constructed; then, the input fuel quantity measurement values ​​are iterated through the multi-sensor multi-measurement dynamic fusion model to obtain the fusion results of the multi-sensor multi-measurement, which are used to characterize the predicted value of the aircraft fuel quantity.

2. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 1 is characterized by: When constructing a multi-sensor multi-measurement dynamic fusion model, the oil quantity measurement value expression of a single sensor and an a-order Lagrangian interpolation expression of a single oil quantity measurement value are constructed respectively. The measurement variance expression of a single sensor is calculated by the Lagrangian interpolation method, which is used to calculate the measurement variance of each sensor. Then, the weight expression of a single sensor is constructed based on the measurement variance of the single sensor to calculate the fusion weight of each sensor; Then, based on the fusion weight of a single sensor and the fuel quantity measurement value of a single sensor, a fusion result expression is constructed by the weighted average method to calculate the fusion result of multiple sensors and multiple measurements to obtain the predicted value of the aircraft fuel quantity.

3. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 2 is characterized by: The fusion result expression is as follows: W k satisfy: in, Indicates the fusion result of multi-sensor multi-measurement; d k Indicates the oil volume measurement value of a single sensor; W k represents the fusion weight of a single sensor; k represents the sensor number, k = 1, 2, ..., n; n represents the number of sensors; d k , W k The value of the subscript k represents the number of the sensor.

4. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 3 is characterized by: The oil volume measurement value of a single sensor is expressed as follows: d k =D+e k Among them, d k Represents the oil volume measurement value of a single sensor; D represents the true value of the oil volume data collected by the sensor; ε k represents the random error of a single sensor; d k , ε k The value of the subscript k represents the number of the sensor.

5. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 3 is characterized by: The fusion weight expression of a single sensor is as follows: Among them, W k Represents the fusion weight of a single sensor; represents the measurement variance of a single sensor; k represents the sensor number, k = 1, 2, ..., n; n represents the number of sensors; W k , The value of the subscript k represents the number of the sensor.

6. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 5 is characterized by: The measurement variance of a single sensor is expressed as follows: in, represents the measurement variance of a single sensor; Δ(x i ) represents the difference between the fusion value and the Lagrange interpolation value of a single oil volume measurement; ω(x j ) represents the difference ratio coefficient of two groups of measurements in a single sensor; a represents the order set in the Lagrange interpolation method, a is an even number, and a <m;x i 、x j It represents the i-th oil quantity measurement value and the j-th oil quantity measurement value corresponding to a single sensor. The values ​​of subscripts i and j represent the serial numbers of the oil quantity measurement values ​​in a single sensor, i=1, 2,…,m, j=1, 2,…,m, i≠j; m represents the number of oil quantity measurement values ​​collected by a single sensor.

7. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 6 is characterized by: The difference between the fusion value and the Lagrange interpolation value of a single oil volume measurement Δ(x i ), the expression is as follows: Among them, d(x i ) represents the fusion measurement value corresponding to the i-th oil level measurement value of a single sensor; L(x i ) represents the a-order Lagrangian interpolation of the i-th oil volume measurement value of a single sensor; D(x i ) represents the true value of the oil volume measurement of the i-th sensor; ε(x i ) represents the random error of the i-th oil level measurement value of a single sensor; D(x j ) represents the true value of the jth oil level measurement of a single sensor; ε(x j ) represents the random error of the jth oil level measurement value of a single sensor; ω(x j ) represents the difference ratio coefficient between two sets of measurements within a single sensor.

8. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to claim 7 is characterized by: The difference ratio coefficient ω(x j ), the expression is as follows: Among them, x i 、x j 、x s It represents the i-th oil quantity measurement value, j-th oil quantity measurement value, and s-th oil quantity measurement value corresponding to a single sensor. The values ​​of i, j, and s all represent the serial values ​​of the oil quantity measurement values ​​in a single sensor, i=1, 2, …, m, j=1, 2, …, m, s=1, 2, …, m, i≠j≠s; m represents the number of oil quantity measurement values ​​collected by a single sensor.

9. The aircraft fuel measurement method of multi-sensor multi-measurement dynamic fusion according to any one of claims 1 to 8, characterized in that: The oil quantity measurement value falling within the confidence interval is obtained from the original oil quantity measurement data collected by multiple sensors through a software filter, specifically: setting the filter queue address, the confidence interval length, the effective range of the oil quantity measurement value, the upper limit of the filter queue and the lower limit of the filter queue in the software filter; then, using the effective range of the oil quantity measurement value, the upper limit of the filter queue and the lower limit of the filter queue to judge each fuel quantity measurement value of each sensor, and obtain the oil quantity measurement value.

Images