A determination method for abnormal power consumption of a meter

Abstract

The present invention relates to a method for determining abnormal power consumption of a meter, which includes the steps of: collecting voltage data of the meter at a fixed time. If there is a steep increase in the voltage data, the monotonic stack algorithm is used to determine whether the reason for the steep increase in voltage is abnormal AD sampling or battery replacement operation. If it is a battery replacement operation, record the time of battery replacement; perform isotonic regression processing on the voltage data within a battery replacement cycle to obtain an isotonic voltage data sequence; based on the isotonic voltage data sequence, determine whether there is an abnormal situation of high power consumption in the meter. If so, construct a dictionary of the number of voltage occurrences; based on the dictionary of the number of voltage occurrences, determine the time point when high power consumption occurs. The purpose of the present invention is to accurately determine whether there is an abnormal situation of high power consumption when the voltage of the meter suddenly increases, and locate the time point of high power consumption abnormality.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy detection, and particularly relates to a method for determining abnormal power consumption of a meter. Background Art

[0002] With the continuous popularization of Internet of Things communication technology, intelligent metering meters with remote transmission functions have gradually replaced the original ordinary meters. They can regularly report metering data and the operating status of the meters (such as battery voltage, valve status, signal strength, etc.) to the management center, and remote meter reading can be achieved without disturbing households. Intelligent metering meters usually consist of a main control circuit board, a basic meter, a housing, etc. Since the components on the circuit board (such as resistors, capacitors, etc.) have a certain service life, corrosion or short - circuit phenomena may occur under relatively bad working conditions, and the most serious impact is that the meter may have the phenomenon of abnormally high power consumption.

[0003] A typical time - battery voltage data of a meter is as Figure 1 shown, and there are 3 key mutation points, which are analyzed as follows:

[0004] ① The voltage value should have decreased slowly, but at this time the voltage value decreases rapidly, which is very likely that the components are damaged and there is a phenomenon of high power consumption;

[0005] ② The voltage value has dropped to 4.7V, and it is judged that the power is almost exhausted; after replacing the battery again, the voltage returns to 6.3V;

[0006] ③ After replacing the battery for 2 - 3 months, the voltage value drops to 4.7V again, and it is judged that the power is exhausted again; after replacing the battery again, the voltage returns to 6.3V; but there is still a situation where the voltage value drops rapidly.

[0007] According to the design mode of the meter, the normal voltage of 4 dry batteries is about 6.4V, and the service life is about 12 months. If the power consumption drops too fast after replacing the new battery, it indicates that there is very likely a problem of abnormal power consumption. Summary of the Invention

[0008] The purpose of the present invention is to accurately determine whether there is an abnormal situation of high power consumption when the voltage of the meter suddenly increases, and locate the time point of abnormal high power consumption, and provide a method for determining abnormal power consumption of a meter.

[0009] In order to achieve the above - mentioned invention purpose, the embodiments of the present invention provide the following technical solutions:

[0010] A method for determining abnormal power consumption of a meter, comprising the following steps:

[0011] Step 1: Collect the voltage data of the meter at a fixed time. If there is a sudden increase in voltage in the voltage data, use the monotonic stack algorithm to determine whether the reason for the sudden increase in voltage is abnormal AD sampling or battery replacement operation. If it is a battery replacement operation, record the time of battery replacement and proceed to Step 2;

[0012] Step 2: Perform isotonic regression processing on the voltage data within a battery replacement cycle to obtain an isotonic voltage data sequence;

[0013] Step 3: Based on the isotonic voltage data sequence, determine whether there is an abnormal situation of high power consumption in the meter. If so, construct a dictionary of the number of voltage occurrences; based on the dictionary of the number of voltage occurrences, determine the time point of high power consumption.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention uses an algorithm combining state transition and monotonic stack to determine whether the reason for the sudden increase in the battery voltage of the meter is abnormal AD sampling or battery replacement operation. If it is determined to be a battery replacement operation, perform isotonic regression processing on the voltage data within a battery replacement cycle to determine whether there is an abnormal situation of high power consumption in the meter, and the time point of high power consumption can be located. And meter repair is required, rather than simply replacing the battery. The whole process of judgment does not require manual participation, reducing the manual workload, and the judgment result is accurate and efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a typical time-battery voltage data waveform diagram of the meter in the background technology;

[0017] Figure 2 It shows the situation of abnormal AD sampling when collecting the voltage data of the meter in the embodiment of the present invention;

[0018] Figure 3 It is a schematic diagram of the isotonic voltage data sequence in the embodiment of the present invention;

[0019] Figure 4 It is a state transition schematic diagram of Step 1 of the present invention;

[0020] Figure 5 It is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention described and illustrated in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents the selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0022] It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present invention, terms such as "first", "second", etc. are only used for distinguishing descriptions, and cannot be understood as indicating or implying relative importance, or implying any such actual relationship or order between these entities or operations. In addition, terms such as "connected" and "coupled" can be directly connected between components, or indirectly connected via other components.

[0023] Embodiment:

[0024] The present invention is realized through the following technical solutions. As Figure 5 shown, a method for determining abnormal power consumption of a meter includes the following steps:

[0025] Step 1, collect the voltage data of the meter at a fixed time. If the voltage data shows a sudden increase in voltage, use the monotonic stack algorithm to determine whether the reason for the sudden increase in voltage is AD sampling anomaly or battery replacement operation. If it is a battery replacement operation, record the time of battery replacement and proceed to Step 2.

[0026] During the AD sampling process of the voltage data of the meter, due to other inevitable technical reasons or external factors, the collected voltage data may occasionally show abnormal fluctuations, as Figure 2 shown in ① and ②. Such AD sampling anomalies are manifested as sudden increases or decreases in voltage, which are easily confused with the sudden increase in voltage during battery replacement. Therefore, in this step, an improved method combining the state transition algorithm and the monotonic stack algorithm is used to determine whether the reason for the sudden increase in voltage is an AD sampling anomaly or caused by battery replacement. For ease of understanding, the voltage state is recorded as three states: 1. Normal operation state; 2. Sudden increase in voltage state; 3. Battery replacement state. Regardless of the reason for the sudden increase in voltage state, the meter will necessarily first go through the normal operation state. Therefore, as Figure 4 shown, Step 1 may include the following steps:

[0027] Step 1-1, Normal operating state: After the meter is powered on, it is in the normal operating state. The voltage data of the meter is collected at fixed times. By comparing the difference between the current voltage data v i and the previous voltage data v i-1 , if it satisfies v i -v i-1 > 1V and v i-1 < 5V, it is determined that the voltage has increased steeply, and it jumps to Step 1-2; otherwise, it remains in the normal operating state.

[0028] Step 1-2, Voltage steep increase state: When in the voltage steep increase state, based on the monotonic stack algorithm, it is judged whether this voltage steep increase state is an AD sampling anomaly or a battery replacement operation; if it is judged as an AD sampling anomaly, it returns to Step 1-1 and returns to the normal operating state; if it is judged as a battery replacement operation, it jumps to Step 1-3.

[0029] Step 1-3, Battery replacement state: When it is determined that it is in the battery replacement state, the time when the voltage has increased steeply is recorded as the time x k of the battery replacement operation. After the recording is completed, it returns to Step 1-1 and returns to the normal operating state.

[0030] Specifically, since the steep increase in voltage may correspond to a battery replacement operation or an AD sampling anomaly, Step 1-2 of this solution is based on the monotonic stack algorithm to judge. By comparing the maximum voltage drop degree within a window size of m within n sampling times, it is judged whether this voltage steep increase is a battery replacement operation. Assume that the voltage data {v1, v2, v3, v4, v5} at 5 times is collected, and m = 3 is set. Then the window sizes of {v1, v2, v3}, {v2, v3, v4}, and {v3, v4, v5} can be selected in turn. Here, it is only used as an example for easy understanding of the meanings of n and m. In fact, the values of n and m are much larger than this. In this embodiment, m = n / 2 or m = (n + 1) / 2 can be preferably used.

[0031] For each sampling time v i , the minimum value among the subsequent m sampling voltages is found in turn. The time complexity of the algorithm is O(nm). This solution designs an algorithm of a monotonic double-ended queue (that is, both the head end and the tail end), which can complete the judgment with a time complexity of O(n + m), so the calculation complexity is greatly reduced. The specific steps of Step 1-2 are as follows:

[0032] Step 1-2-1, Record the voltage data sequence collected at n + m times as V = {v i} = {v1, v2,..., v n+m}, v iDenote the \(i\)-th voltage data in sequence \(V\), where \(i = 1, 2, \cdots, n + m\), \(n\) is the number of sampling times, and \(m\) is the window size; Denote the subsequent voltage minimum value sequence of the first \(n + 1\) sampling voltages as \(Y\), and the sequence \(Y\) is empty in the initial state; Create a double-ended queue \(D\) for recording the subscripts of voltage data, and the queue \(D\) is empty in the initial state; First, select the voltage \(v_1\) with subscript \(i = 1\) in sequence \(V\).

[0033] Step 1-2-2, if the queue \(D\) is empty, jump to Step 1-2-3; Otherwise, denote the tail element of the queue \(D\) as \(a\). If \(v\) i \(< v\) a \(, v\) a is the voltage data with subscript \(a\) in sequence \(V\), then remove the tail element \(a\) of the queue \(D\) and execute this step again; If \(v\) i \(\geq v\) a \(, then jump to Step 1-2-3.\)

[0034] Step 1-2-3, if the queue \(D\) is empty, jump to Step 1-2-4; If the queue \(D\) is not empty, denote the head element of the queue \(D\) as \(b\). If \(b\leq i - m\), then remove the head element \(b\) of the queue \(D\) and execute this step again; If \(b > i - m\), then jump to Step 1-2-4.

[0035] Step 1-2-4, add \(i\) to the tail of the queue \(D\). If \(i < m\), execute \(i = i + 1\) and jump to Step 1-2-2; If \(i\geq m\), then denote the head element of the queue \(D\) as \(b\), and add \(v\) b to sequence \(Y\), \(v\) b is the voltage data with subscript \(b\) in sequence \(V\); Then, if \(i < n + m\), execute \(i = i + 1\) and jump to Step 1-2-2. If \(i\geq n + m\), then end this step and jump to Step 1-2-5.

[0036] As an example, when the voltage is in a steep increase state, the voltage \(v_1 = 6.1V\). Set \(n = 9\) and \(m = 3\). Then the voltage data sequence \(V\) is shown in Table 1.

[0037] Table 1 Parameter table of voltage data sequence \(V\)

[0038]

[0039] Initialization: \(D=\{\}\); \(Y = \{\}\);

[0040] \(i = 1\), \(v\) i \(= v_1\): At this time, \(D\) is empty. Jump from Step 1-2-2 to Step 1-2-3 and then to Step 1-2-4. Add 1 to the tail of \(D\), \(D=\{1\}\); Since \(1 < 3\), execute \(i = i + 1\) and jump to Step 1-2-2;

[0041] \(i = 2\), \(v\) i= v2: At this time, D = {1}. Denote the last element 1 in the queue as a. Since v2 = v1, jump from step 1 - 2 - 2 to step 1 - 2 - 3. Denote 1 as b. Since 1 > 2 - 3, jump from step 1 - 2 - 3 to step 1 - 2 - 4. Add 2 to the end of D, so D = {1, 2}. Since 2 < 3, execute i = i + 1 and jump to step 1 - 2 - 2;

[0042] i = 3, v i = v3: At this time, D = {1, 2}. Denote the last element 2 in the queue as a. Since v3 < v2, remove the last element 2. At this time, D = {1}. Denote the last element 1 in the queue as a. Since v3 < v1, remove the last element 1. At this time, D = {}. Jump from step 1 - 2 - 2 to step 1 - 2 - 3 and then to step 1 - 2 - 4. Add 3 to the end of D, so D = {3}. Since i = 3, denote 3 as b, and add v3 to Y. At this time, Y = {v3}. Since 3 < 9 + 3, execute i = i + 1 and jump to step 1 - 2 - 2;

[0043] i = 4, v i = v4: At this time, D = {3}. Denote the last element 3 in the queue as a. Since v4 = v3, jump from step 1 - 2 - 2 to step 1 - 2 - 3. Denote the first element 3 in the queue as b. Since 3 > 4 - 3, jump from step 1 - 2 - 3 to step 1 - 2 - 4. Add 4 to the end of D, so D = {3, 4}. Since 4 > 3, denote the first element 3 in the queue as b, and add v3 to Y. Y = {v3, v3}. Since 4 < 9 + 3, execute i = i + 1 and jump to step 1 - 2 - 2;

[0044] i = 5, v i = v5: At this time, D = {3, 4}. Denote the last element 4 in the queue as a. Since v5 = v4, jump from step 1 - 2 - 2 to step 1 - 2 - 3. Denote the first element 3 in the queue as b. Since 3 > 5 - 3, jump from step 1 - 2 - 3 to step 1 - 2 - 4. Add 5 to the end of D, so D = {3, 4, 5}. Since 5 > 3, denote the first element 3 in the queue as b, and add v3 to Y. Y = {v3, v3, v3}. Since 5 < 9 + 3, execute i = i + 1 and jump to step 1 - 2 - 2;

[0045] i = 6, v i= v6: At this time, D = {3, 4, 5}. Denote the tail element 5 as a. Since v6 > v5, jump from step 1-2-2 to step 1-2-3. Denote the head element 3 as b. Since 3 = 6 - 3, remove the head element 3. At this time, D = {4, 5}. Denote the head element 4 as b. Since 4 > 6 - 3, jump from step 1-2-3 to step 1-2-4. Add 6 to the tail of D, D = {4, 5, 6}. Since 6 > 3, denote the head element 4 as b and add v4 to Y, Y = {v3, v3, v3, v4}. Since 6 < 9 + 3, execute i = i + 1 and jump to step 1-2-2.

[0046] i = 7, v i = v7: At this time, D = {4, 5, 6}. Denote the tail element 6 as a. Since v7 < v6, remove the tail element 6. At this time, D = {4, 5}. Denote the tail element 5 as a. Since v7 < v5, remove the tail element 5. At this time, D = {4}. Denote the tail element 4 as a. Since v7 < v4, remove the tail element 4. At this time, D = {}, jump from step 1-2-2 to step 1-2-3 and then to step 1-2-4. Add 7 to the tail of D, D = {7}. Since 7 > 3, denote the head element 7 as b and add v7 to Y, Y = {v3, v3, v3, v4, v7}. Since 7 < 9 + 3, execute i = i + 1 and jump to step 1-2-2.

[0047] i = 8, v i = v8: At this time, D = {7}. Denote the tail element 7 as a. Since v8 > v7, jump from step 1-2-2 to step 1-2-3. Denote the head element 7 as b. Since 7 > 8 - 3, jump from step 1-2-3 to step 1-2-4. Add 8 to the tail of D, D = {7, 8}. Since 8 > 3, add v7 to Y, Y = {v3, v3, v3, v4, v7, v7}. Since 8 < 9 + 3, execute i = i + 1 and jump to step 1-2-2.

[0048] i = 9, v i = v9: At this time, D = {7, 8}. Denote the tail element 8 as a. Since v8 > v7, jump from step 1-2-2 to step 1-2-3. Denote the head element 7 as b. Since 7 > 9 - 3, jump from step 1-2-3 to step 1-2-4. Add 9 to the tail of D, D = {7, 8, 9}. Since 9 > 3, add v7 to Y, Y = {v3, v3, v3, v4, v7, v7, v7}. Since 9 < 9 + 3, execute i = i + 1 and jump to step 1-2-2.

[0049] i = 10, v i = v 10 : At this time, D = {7, 8, 9}. Denote the tail element 9 as a. Since v10 <If v9, then remove the element 9 at the end of the queue. At this time, D = {7, 8}; Denote the element 8 at the end of the queue as a. Since v 10 <If v8, then remove the element 8 at the end of the queue. At this time, D = {7}; Denote the element 7 at the end of the queue as a. Since v 10 = v7, jump from step 1 - 2 - 2 to step 1 - 2 - 3; Denote the element 7 at the head of the queue as b. Since 7 = 10 - 3, then remove the element 7 at the head of the queue. At this time, D = {}, jump from step 1 - 2 - 3 to step 1 - 2 - 4; Add 10 to the end of D, D = {10}. Since 10 > 3, denote the element 10 at the head of the queue as b, and add v 10 to Y, Y = {v3, v3, v3, v4, v7, v7, v7, v 10}; Since 10 < 9 + 3, execute i = i + 1 and jump to step 1 - 2 - 2;

[0050] i = 11, v i = v 11 : At this time, D = {10}. Denote the element 10 at the end of the queue as a. Since v 11 < v 10 , then remove the element 10 at the end of the queue. At this time, D = {}, jump from step 1 - 2 - 2 to step 1 - 2 - 3 and then to step 1 - 2 - 4; Add 11 to the end of D, D = {11}. Since 11 > 3, denote the element 11 at the head of the queue as b, and add v 11 to Y, Y = {v3, v3, v3, v4, v7, v7, v7, v 10 , v 11}; Since 11 < 9 + 3, execute i = i + 1 and jump to step 1 - 2 - 2;

[0051] i = 12, v i = v 12 : At this time, D = {11}. Denote the element 11 at the end of the queue as a. Since v 12 < v 11 , then remove the element 11 at the end of the queue. At this time, D = {}, jump from step 1 - 2 - 2 to step 1 - 2 - 3 and then to step 1 - 2 - 4; Add 12 to the end of D, D = {12}. Since 12 > 3, denote the element 12 at the head of the queue as b, and add v 12 to Y, Y = {v3, v3, v3, v4, v7, v7, v7, v 10 , v 11 , v 12}; Since 12 = 9 + 3, end the step.

[0052] Step 1 - 2 - 5. After the above steps 1 - 2 - 1, 1 - 2 - 2, 1 - 2 - 3, and 1 - 2 - 4, for the first n + 1 voltage data {v1, v2,..., v in the voltage data sequence Vn+1}, for each v i , calculate v i - y i respectively, where y i represents the i-th voltage data in sequence Y. If all are less than the set threshold, it indicates that this voltage steep increase is a battery replacement operation, and record the time corresponding to voltage data v1 as the battery replacement time x k .

[0053] For example, finally Y = {6.0, 6.0, 6.0, 6.0, 5.9, 5.9, 5.9, 5.9, 5.8, 5.7}, and there are 10 voltage data in Y. The first 10 voltage data in voltage data sequence V are {6.1, 6.1, 6.0, 6.0, 6.0, 6.1, 5.9, 6.0, 6.0, 5.9}. The threshold can be set to 0.3, and calculate v i - y i respectively, and get {0.1, 0.1, 0, 0, 0.1, 0.2, 0, 0.1, 0.2, 0.2}. It can be seen that all are less than the set threshold 0.3, which indicates that this voltage steep increase is a battery replacement operation, and record the time corresponding to voltage data v1 as the battery replacement time x for this time k .

[0054] In step 1, when the meter shows a steep increase in voltage during normal operation, this situation may be due to abnormal AD sampling or a battery replacement operation. To determine which one it is, through the method of this step, collect the voltage data when and after the voltage steep increase, and use the monotonic stack algorithm to judge, that is, the process from step 1-2-1 to step 1-2-5, to realize the automatic judgment of the specific factors of the voltage steep increase situation, without manual participation in the calculation, improve the judgment efficiency, and reduce the manual complexity. If it is judged as abnormal AD sampling, it is a normal situation of the voltage change of the meter. If it is judged as a battery replacement operation, go to step 2

[0055] Step 2, perform isotonic regression processing on the voltage data within a battery replacement cycle to obtain an isotonic voltage data sequence

[0056] Each time the battery replacement time x k is recorded, except for the first record, a battery replacement cycle [x k-1 , x k can be determined. Then, it is necessary to analyze the voltage data within a battery replacement cycle to judge whether there is a problem of high power consumption in the meter. If there is no problem of high power consumption, the voltage data within a battery replacement cycle should show a gradually decreasing trend. Therefore, in this step, first perform isotonic regression processing on the voltage data within a battery replacement cycle

[0057] The specific steps of step 2 are as follows

[0058] Step 2-1: Denote the battery replacement cycle as [x k-1 ,x k . There are J voltage data within it. Let x k-1 represent the (k - 1)-th battery replacement time, and x k represent the k-th battery replacement time. The voltage data sequence is VJ = {v j} = {v1, v2,..., v J}, where v j represents the j-th voltage data in the voltage data sequence VJ, and j = 1, 2,..., J. Maintain a subscript traversal index j = 2; create a subscript index set B and initialize B = {1}; create an ordered voltage data sequence U and initialize U = VJ.

[0059] Step 2-2: If j > J, jump to Step 2-3; otherwise, judge if u j ≤u j-1 , where u j and u j-1 represent the j-th voltage data and the (j - 1)-th voltage data in the sequence U respectively. Clear the elements in B, execute B = {j}, j = j + 1, and execute this step again; if u j >u j-1 , then add j to B, calculate the weighted average voltage data v0 according to the following formula, and execute u p =v0 for all elements p ∈ B in B, and execute this step again;

[0060]

[0061] where u p represents the voltage data with subscript p in the sequence U; represents the weight of u p , and the calculation method is:

[0062]

[0063] where v p and v p-1 represent the voltage data with subscripts p and p - 1 in the sequence VJ respectively;

[0064] Step 2-3: If the elements in the sequence U satisfy u1 ≥ u2 ≥... ≥ u J , where u1 is the first voltage data in the sequence U, u2 is the second voltage data in the sequence U, and u J is the J-th voltage data in the sequence U, then end this step; otherwise, execute B = {1}, j = 2, and jump to Step 2-2.

[0065] For example, assume that the voltage data sequence within a battery replacement cycle is VJ = {6.1, 6.1, 6.0, 6.0, 6.0, 6.1, 5.9, 6.0}, J = 8. In the initial state, j = 2, B = {1}, i = 2, and U = VJ = {6.1, 6.1, 6.0, 6.0, 6.0, 6.1, 5.9, 6.0};

[0066] Since 2 < 8 and u2 = u1, clear the elements in B, execute B = {2}, and j = j + 1; since 3 < 8 and u3 < u2, clear the elements in B, execute B = {3}, and j = j + 1; since 4 < 8 and u4 = u5, clear the elements in B, execute B = {4}, and j = j + 1; since 5 < 8 and u5 = u4, clear the elements in B, execute B = {5}, and j = j + 1; since 6 < 8 and u6 > u5, add 6 to B. At this time, B = {5, 6}, calculate v0 when p = 5 and p = 6 respectively to obtain u5 and u6, and replace u5 and u6 in U with the v0 values calculated here; determine whether U satisfies u1 ≥ u2 ≥... ≥ u J , if it is satisfied, end the calculation; if not, reset B = {1}, j = 2, and jump to step 2 - 2 to continue the iterative calculation based on the current U until u1 ≥ u2 ≥... ≥ u J is satisfied, so as to obtain the ordered voltage data sequence U.

[0067] Obtain the ordered voltage data sequence U = {u1, u2,..., u J}}, and minimize the following objective function through the isotonic regression algorithm:

[0068]

[0069] where u1 ≥ u2 ≥... ≥ u J ; is the weight of v j , . In this solution, is calculated as follows: when j = 1, it is assigned as ; when v j is not within the range of [4V, 6.5V], it indicates that there must be an AD sampling anomaly, and the weight is assigned as ; in other cases, it is the absolute value of the difference between its previous AD sampling voltage value and is normalized to the range of [0, 1].

[0070]

[0071] where v j , v j-1 respectively represent the j - th voltage data and the (j - 1) - th voltage data in the voltage data sequence VJ.

[0072] Step 3: Based on the order-preserving voltage data sequence, determine whether there is an abnormal situation of high power consumption in the meter. If so, construct a dictionary of the number of occurrences of voltages; based on the dictionary of the number of occurrences of voltages, determine the time points of high power consumption.

[0073] After the above processing, the voltage data becomes stepped data, as Figure 3 shown is the voltage data of a certain abnormal meter after order-preserving regression processing within a battery replacement cycle. Next, based on the order-preserving voltage data sequence U, analyze whether there is a problem of high power consumption in the meter:

[0074] Step 3-1: Determine whether it is high power consumption. According to the setting of a conventional meter, the voltage data is sampled AD once a day. If the number of voltage data within the current battery replacement cycle is less than the set value L (e.g., L = 270), it can be determined that the meter has a problem of high power consumption and proceed to Step 3-2.

[0075] Step 3-2: Determine the time points of high power consumption anomalies. If it is confirmed to be high power consumption, establish a voltage - number of occurrences dictionary, traverse the order-preserving regression voltage data sequence U, and count the number of times each voltage data is collected. Taking Figure 3 as an example, for instance, the statistical result is:

[0076] {6.2: 69, 6.1: 101, 6.0: 2, 5.9: 3, 5.8: 10, 5.7: 8, 5.6: 8, 5.5: 9, 5.4: 8, 5.3: 10, 5.2: 12, 5.1: 8, 5.0: 6, 4.9: 6, 4.8: 4, 4.7: 1}

[0077] Among them, 6.2V is collected 69 times (c1 = 69), 6.1V is collected 101 times (c2 = 101), 6.0V is collected 2 times (c3 = 2), and the same applies hereinafter. Therefore, a sequence C = {c1, c2,..., c K} of the number of occurrences of voltage data can be constructed, where c k represents the number of times the k-th voltage data is collected, c = 1, 2,..., K, and K represents the number of voltage data values.

[0078] Perform the following operations on the sequence C = {c1, c2,..., c K}:

[0079] Step 3-2-1: Initialize the traversal subscript k = 1, denote the length of sequence C as K, set the size of the judgment window as d (d is an even number), create a standard deviation statistical sequence S, and in the initial state, sequence S is empty.

[0080] Step 3-2-2: Extract the k-th to (k + d - 1)-th elements from sequence C, {c k , c k+1 ,..., c k+d-1}, and calculate the standard deviation s of this sequence using the following formula, then add s to sequence S.

[0081]

[0082] where s represents the standard deviation of the sequence {c k , c k+1 ,..., c k+d-1}; c k , c k+1 , c k+d-1 all represent elements in the sequence {c k , c k+1 ,..., c k+d-1}; represents the average value of the sequence {c k , c k+1 ,..., c k+d-1}; d represents the number of elements in the sequence {c k , c k+1 ,..., c k+d-1}.

[0083] Step 3-2-3: Execute k = k + 1. If k < K - d, jump to Step 3-2-2; otherwise, jump to 3-2-4.

[0084] Step 3-2-4: Denote the subscript corresponding to the maximum value in sequence S as r, execute r = r + d / 2, and for all k < r, execute Σc k . The obtained value is the time node most likely to have high power consumption.

[0085] For example, if voltage data is recorded once a day, and finally it is calculated that Σc k = 65, then it is determined that there is an abnormal situation of high power consumption on the 65th day.

[0086] As described above, this is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A determination method for abnormal power consumption of a meter, characterized in that: It includes the following steps: Step 1: Collect the voltage data of the meter at a fixed time. If there is a sudden increase in voltage in the voltage data, use the monotonic stack algorithm to determine whether the reason for the sudden increase in voltage is abnormal AD sampling or battery replacement operation. If it is a battery replacement operation, record the time of battery replacement and proceed to Step 2; In Step 1, when there is a sudden increase in voltage in the voltage data: Step 1-2-1, record the voltage data sequence collected at n + m moments as V = {v i} = {v1, v2,..., v n+m}, where v i represents the i-th voltage data in the sequence V, i = 1, 2,..., n + m, n is the number of sampling moments, and m is the window size; record the subsequent voltage minimum value sequence of the first n + 1 sampling voltages as Y, and the sequence Y is empty in the initial state; create a double-ended queue D for recording the subscripts of voltage data, and the queue D is empty in the initial state; first select the voltage v1 with subscript i = 1 in the sequence V; Step 1-2-2, if queue D is empty, jump to Step 1-2-3; otherwise, denote the tail element of queue D as a. If v i <v a , v a is the voltage data at index a in sequence V, then remove the tail element a of queue D and execute this step again; if v i ≥v a , then jump to Step 1-2-3; Step 1-2-3: If queue D is empty, jump to Step 1-2-4; if queue D is not empty, record the first element of queue D as b. If b ≤ i - m, remove the first element b of queue D and execute this step again; if b > i - m, jump to Step 1-2-4; Step 1-2-4, add i to the end of queue D. If i < m, execute i = i + 1 and jump to Step 1-2-2; if i ≥ m, then denote the head element of queue D as b, and add v b to sequence Y, where v b is the voltage data at index b in sequence V. Then, if i < n + m, execute i = i + 1 and jump to Step 1-2-2; if i ≥ n + m, then end this step and jump to Step 1-2-5; Step 1-2-5, after the above-mentioned Steps 1-2-1, 1-2-2, 1-2-3, and 1-2-4, for the first n + 1 voltage data {v1, v2,..., v n+1} in the voltage data sequence V, for each v i , calculate the value of v i - y i respectively, where y i represents the i-th voltage data in the sequence Y. If all are less than the set threshold, it indicates that this voltage steep increase is a battery replacement operation, and record the time x k corresponding to the voltage data v1 as the battery replacement time; Step 2: Perform isotonic regression processing on the voltage data within a battery replacement cycle to obtain an isotonic voltage data sequence; Step 3: Based on the isotonic voltage data sequence, determine whether there is an abnormal situation of high power consumption in the meter. If so, construct a dictionary of the number of voltage occurrences; Based on the dictionary of the number of voltage occurrences, determine the time point of high power consumption; Step 3 specifically includes the following steps: Step 3-1: Determine whether it is high power consumption: According to the setting, the voltage data is sampled once a day by AD. If the number of voltage data in the current battery replacement cycle is less than the set value L, it can be determined that the meter has a high power consumption problem and proceed to Step 3-2; Step 3-2: Determine the time point of high power consumption abnormality: If it is confirmed to be high power consumption, establish a dictionary of voltage - number of occurrences, traverse the isotonic regression voltage data sequence U, and count the number of times each voltage data is collected to form a voltage data number of occurrences sequence C; Step 3-2 specifically includes the following steps: Step 3-2-1: Initialize the traversal subscript k = 1, record the length of sequence C as K, set the size of the judgment window as d, d is an even number, and create a standard deviation statistical sequence as S. In the initial state, sequence S is empty; Step 3-2-2, take out the k-th element to the (k + d - 1)-th element {c k , c k+1 ,..., c k+d-1} in sequence C, calculate the standard deviation s of this sequence with the following formula, and add s to sequence S; where s represents the standard deviation of the sequence {c k , c k+1 ,..., c k+d-1}; c k , c k+1 , c k+d-1 all represent elements in the sequence {c k , c k+1 ,..., c k+d-1}; represents the average value of the sequence {c k , c k+1 ,..., c k+d-1}; d represents the number of elements in the sequence {c k , c k+1 ,..., c k+d-1}; Step 3-2-3: Execute k = k + 1. If k < K - d, jump to Step 3-2-2; otherwise, jump to 3-2-4; Step 3-2-4, denote the subscript corresponding to the maximum value in sequence S as r, execute r = r + d / 2, and for all k < r, execute Σc k , and the obtained value is the time node when high power consumption occurs.

2. The determination method for abnormal power consumption of the meter according to claim 1, characterized in that: Step 1 specifically includes the following steps: Step 1-1, normal operation state: After the meter is powered on, it is in the normal operation state. The voltage data of the meter is collected at fixed times. By comparing the difference between the current voltage data v i and the previous voltage data v i-1 , if it satisfies v i -v i-1 > 1V and v i-1 < 5V, it is determined that the voltage has increased steeply and jumps to Step 1-2; otherwise, it remains in the normal operation state; Step 1-2: Voltage sudden increase state: When in the voltage sudden increase state, based on the monotonic stack algorithm, determine whether this voltage sudden increase state is abnormal AD sampling or battery replacement operation; if it is judged to be abnormal AD sampling, return to Step 1-1 and return to the normal operation state; if it is judged to be a battery replacement operation, jump to Step 1-3; Step 1-3, Battery Replacement State: When it is determined that the device is in the battery replacement state, record the time of the steep voltage increase as the time x of the battery replacement operation k , after the recording is completed, return to Step 1-1 to return to the normal operating state.

3. The determination method for abnormal power consumption of the meter according to claim 1, characterized in that: Step 2 specifically includes the following steps: Step 2-1, record the battery replacement cycle as [x k-1 , x k , within which there are J voltage data. Let x k-1 represent the (k - 1)-th battery replacement time, and x k represent the k-th battery replacement time. The voltage data sequence is VJ = {v j} = {v1, v2,..., v J}, where v j represents the j-th voltage data in the voltage data sequence VJ, and j = 1, 2,..., J; maintain a subscript traversal serial number j = 2; create a subscript serial number set B and initialize B = {1}; create an order-preserving voltage data sequence U and initialize U = VJ; Step 2-2, if j > J, jump to Step 2-3; otherwise, judge if u j ≤ u j-1 , where u j and u j-1 represent the j-th voltage data and the (j-1)-th voltage data in sequence U respectively. Clear the elements in B, execute B = {j}, j = j + 1, and execute this step again; if u j > u j-1 , then add j to B, calculate the weighted average voltage data v0 according to the following formula, and execute u p = v0 for all elements p ∈ B in B, and execute this step again; where, u p represents the voltage data with subscript p in sequence U; represents u p 's weight, and the calculation method is: where, v p and v p-1 represent the voltage data with subscripts p and p - 1 in sequence VJ, respectively; Step 2-3, if the elements in sequence U satisfy u1≥u2≥...≥u J , where u1 is the first voltage data in sequence U, u2 is the second voltage data in sequence U, and u J is the Jth voltage data in sequence U, then end this step; otherwise, execute B={1}, j = 2, and jump to Step 2-2.

4. The determination method for abnormal power consumption of the meter according to claim 3, characterized in that: Step 2 also includes the following steps: Obtain the order-preserving voltage data sequence U = {u j} = {u1, u2,..., u J}, where u j represents the j-th voltage data in the order-preserving voltage data sequence U, j = 1, 2,... J, and J is the number of voltage data in the order-preserving voltage data sequence U. Minimize the following objective function f(x) through the order-preserving regression algorithm: where u1≥u2≥...≥u J ; is the weight of v j ; ; The calculation method is as follows: When j = 1, it is assigned as ; When v j is not in the range of [4V, 6.5V], it indicates an abnormal AD sampling, and the weight is assigned as ; In other cases, it is based on the absolute value of the difference from its previous AD sampling voltage value and normalized to the range of [0, 1]; where, v j and v j-1 respectively represent the j-th voltage data and the (j - 1)-th voltage data in the voltage data sequence VJ.

Images