A fault self-diagnosis method and system for a temperature transmitter

Abstract

The application discloses a kind of temperature transmitter's fault self-diagnosis method and system, the present application relates to the technical field of fault self-diagnosis, the present application real-time acquisition temperature time series of temperature transmitter to be measured, synchronous development temperature mutation detection and time series integrity check, and mark suspected abnormal transmitter;Again with suspected abnormal transmitter as core, in combination with thermal monitoring system and physical distance screening near neighbor candidate equipment, by temperature trend isotropicity filtration forms collaborative diagnosis cluster;Finally, time stamp global alignment is done to real-time temperature sequence, constructs global judgment matrix and completes fault determination.The present application is judged by two-dimensional preliminary judgment, and collaborative equipment is selected by combining topology and distance, and is accurately determined by matrix, effectively distinguishes hardware failure and working condition fluctuation, reduces the rate of false positive and false negative, and improves the diagnosis accuracy and anti-interference of temperature transmitter.

Description

Technical Field

[0001] This invention relates to the field of fault self-diagnosis technology, specifically to a fault self-diagnosis method and system for temperature transmitters. Background Technology

[0002] Temperature transmitters are the core sensing units for temperature monitoring and process control in industrial settings. Their operational reliability directly affects production safety and control accuracy. Fault self-diagnosis technology is a key support for ensuring the stable operation of transmitters. However, existing fault self-diagnosis schemes for temperature transmitters still have the following shortcomings: First, existing diagnostic methods mostly rely solely on the time-series data of a single transmitter as the basis for judgment, and only identify anomalies through numerical mutations and threshold exceedances. They do not introduce related equipment under the same operating conditions to carry out collaborative cross-verification, which easily leads to misjudging normal process fluctuations as equipment failures. Secondly, when screening collaborative diagnostic equipment, the existing solution only relies on physical geometric distance to delineate candidate equipment, without combining the thermal monitoring system affiliation and the process medium flow path topology. The selected equipment has no actual thermal coupling relationship with the suspected abnormal transmitter, and the collaborative reference value is extremely low. Third, the fault diagnosis process lacks standardized timing alignment and matrix-based judgment mechanisms. Furthermore, the timing of multiple devices is not synchronized, and the judgment logic is fragmented, making it impossible to form an accurate basis for fault diagnosis. Therefore, there is an urgent need for a self-diagnosis method and system for temperature transmitter faults. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a self-diagnosis method and system for temperature transmitter faults, which solves the problems of easy misjudgment, weak coordination, and insufficient judgment accuracy in existing temperature transmitter fault diagnosis.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a fault self-diagnosis method for a temperature transmitter, comprising: S1. Real-time acquisition of the temperature time-series stream sequence of the temperature transmitter under test. Simultaneous temperature change detection and timing integrity verification are performed on the temperature time-series stream sequence. If both temperature change detection and timing integrity verification pass, the temperature transmitter under test is determined to be operating normally. If either verification fails, the temperature transmitter under test is determined to have a suspected fault and is marked as a suspected abnormal transmitter. S2. Taking the suspected abnormal transmitter as the core, retrieve the physical installation coordinate data of the transmitters in the industrial field, generate a set of nearest candidate transmitters based on the actual physical distance, and for each candidate temperature transmitter in the set of nearest candidate transmitters, simultaneously obtain its historical temperature time series sequence, and perform temperature trend filtering with the historical temperature time series sequence of the suspected abnormal transmitter to form a collaborative diagnostic transmitter cluster. S3. Real-time synchronous acquisition of the real-time temperature time-series stream sequence of each temperature transmitter in the suspected abnormal transmitter and the collaborative diagnostic transmitter cluster. At the same time, perform global timestamp alignment on all real-time temperature time-series stream sequences to construct a global judgment matrix, and complete the fault determination by combining the real-time temperature time-series stream sequence of the suspected abnormal transmitter.

[0005] As a further aspect of the present invention, the specific operation for performing temperature abrupt change detection on the temperature time-series stream sequence is as follows: For a temperature time-series sequence S, for each sampling point S[i] with index i from 1 to n-1, calculate its change direction attribute relative to its preceding adjacent sampling point S[i-1]. The specific rules are as follows: if S[i] > S[i-1], then the change direction attribute corresponding to this sampling point is increasing; if S[i] = S[i-1], then the change direction attribute corresponding to this sampling point is flat; if S[i] < S[i-1], then the change direction attribute corresponding to this sampling point is decreasing, where n is the total number of sampling points in the sequence S. According to the time sequence of the temperature time series flow sequence S, all the calculated change direction attributes are combined to generate a change direction sequence D that is aligned with the time sequence of S. The index of D is j, j∈[0,n-2] and takes integer values. D[j] uniquely corresponds to the change direction attributes of S[j+1] and S[j]. For the changing direction sequence D, perform adjacent direction reversal verification and isolated direction verification respectively, and output the original sequence sampling point positions of all nodes marked as abnormal direction nodes. If there are no abnormal direction nodes, output an empty result. If any marked abnormal direction node exists, the temperature change detection is deemed to fail; otherwise, the temperature change detection is deemed to pass.

[0006] As a further aspect of the present invention, the specific rules for the adjacent direction reversal verification are as follows: traverse the changing direction sequence D, and for each element D[j] with index j from 1 to n-2, verify the direction combination of D[j] and the preceding adjacent element D[j-1]. If the two are directly adjacent combinations of "ascending and descending" or "descending and ascending", then mark the original sequence sampling point position S[j] corresponding to this group of adjacent elements as an abnormal direction node; the specific rules for the isolated direction verification are as follows: traverse the changing direction sequence D, and for each element D[j] with index j from 1 to n-2, verify the consistency of D[j] with the preceding and following adjacent elements D[j-1] and D[j+1]. If the directions of D[j] are not the same as those of D[j-1] and D[j+1], and the directions of D[j-1] and D[j+1] are completely consistent, then mark the original sequence sampling point position S[j] corresponding to this element as an abnormal direction node.

[0007] As a further aspect of the present invention, the specific operation for performing time sequence integrity verification on the temperature time sequence is as follows: Extract the sampling position IDs of the first and last sampling points in S, respectively, and denote them as ID_start and ID_end. Then, generate a set of integers that increment by 1 consecutively, denoted as Set_std, with ID_start as the starting value and ID_end as the ending value. The acquisition sequence ID is automatically generated synchronously when the temperature transmitter acquisition hardware triggers sampling, and is bound to the sampling point for output, following the principle of "the ID increments by 1 for each acquisition trigger"; Extract the acquisition sequence IDs of all sampling points in S, and form the actual acquisition sequence set according to the order of the sampling points, denoted as Set_act; Perform a full consistency comparison between Set_act and Set_std: if the elements of Set_act and Set_std are completely identical and their order is completely matched, the timing integrity check is considered to have passed; otherwise, the timing integrity check is considered to have failed.

[0008] As a further aspect of the present invention, the specific operation for generating the nearest neighbor candidate transmitter set is as follows: Extract the inherent thermal monitoring system attributes of the suspected abnormal transmitter, identify the thermal system A directly monitored by the suspected abnormal transmitter, and retrieve the fixed process medium flow path topology data of the industrial site to locate two types of related objects: first, all temperature transmitters installed inside thermal system A; second, temperature transmitters in adjacent process units that have direct medium communication and direct heat exchange with thermal system A and have direct thermal influence related to thermal system A. The thermal system is a fixed process unit in the industrial site with continuous heat conduction and heat exchange relationship. The two types of temperature transmitters mentioned above are summarized to form a basic associated region transmitter set. At the same time, the three-dimensional physical installation coordinates of each temperature transmitter in the set are extracted, the actual physical distance L between each temperature transmitter and suspected abnormal transmitters is calculated, and temperature transmitters with L > Lth are eliminated to generate the final nearest neighbor candidate transmitter set, where Lth is the distance threshold.

[0009] As a further aspect of the present invention, the specific operation for temperature trend unidirectional filtering is as follows: The sampling points of the historical sequence S_sus of the suspected abnormal transmitter and the historical sequence S_can of the candidate temperature transmitter are matched one-to-one in chronological order, where the sampling point index i ranges from 0 to n-1. For the historical sequence S_sus of the suspected abnormal transmitter, the first m consecutive sampling points at the beginning of S_sus are extracted to form the first anchoring set A_sus, and the last m consecutive sampling points at the end of S_sus are extracted to form the tail anchoring set B_sus, where m=[k×n], k is the proportional coefficient, and its value range is (0,0.5). For the first anchor set A_sus, calculate HL respectively. A_sus With Qn A_sus Similarly, for the tail anchoring set B_sus, HL is calculated. B_sus With Qn B_sus ; The calculation of HL A_sus With Qn A_sus The specific rules are as follows: HL A_sus Calculate the arithmetic mean of all pairwise sampling points within the initial anchor set A_sus, and take the median of all means; this is HL. A_sus ; Qn A_sus Calculate the absolute difference between all pairwise sampling points within the initial anchor set A_sus, and take the first quartile of all absolute differences. Multiply this quartile by a fixed constant factor d to obtain Qn. A_sus ; HL based on the first and last anchored sets A_sus HL B_sus The main trend direction identifier Dsus of S_sus is determined by the following rules: If HL B_sus >HL A_sus If HL is increasing, then Dsus is increasing; if HL is increasing, then Dsus is increasing. B_sus <HL A_sus If HL is decreasing, then Dsus is decreasing; B_sus =HL A_sus If so, then Dsus is flat; Using the same processing rules as S_sus described above, the same operation is performed on the historical sequence S_can of the candidate temperature transmitters to obtain the first anchoring set A_can and the last anchoring set B_can, and corresponding to the statistics HL. A_can Qn A_can HL B_can Qn B_can And the main trend direction indicator Dcan; If the candidate temperature transmitter fails the stability pre-verification, or if the main trend direction indicators Dsus and Dcan of the two sequences are not equal, the trend homogeneity filtering is deemed to fail; otherwise, it enters the intermediate set dual-dimensional verification. The specific rule for the stability pre-verification is as follows: if the Qn of the candidate temperature transmitter... A_can >QnA_sus or Qn B_can >Qn B_sus If so, the stability pre-check fails.

[0010] As a further aspect of the present invention, the specific steps for performing the two-dimensional verification of the intermediate set are as follows: After removing the first and last anchor sets from S_sus and S_can, we obtain the intermediate set of the two, and divide the intermediate set into p consecutive subsets of the same length, where p≥2 and are integers. Based on p subsets of S_sus, the robust level sequence of subsets Msus=[HL] is obtained. sus1 HL sus2 ,...,HL susp Based on the main trend direction of Dsus, a monotonicity consistency check is performed on Msus: if Dsus is increasing, check whether Msus is monotonically increasing; if Dsus is decreasing, check whether Msus is monotonically decreasing; if Dsus is flat, check whether all elements of Msus are equal. Using the same rules as described above, the robustness sequence McCan = [HL] is calculated for each of the p subsets of S_can. can1 HL can2 ,...,HL canp ], and perform the same monotonicity consistency check as described above based on Dcan; If the robust level sequences of the subsets of the two sequences S_sus and S_can both pass their respective monotonicity consistency checks, then the same-direction filtering is considered to pass and the sequence is included in the final collaborative diagnostic transmitter cluster. Otherwise, the same-direction filtering is considered to fail and the sequence is not included in the final collaborative diagnostic transmitter cluster.

[0011] As a further aspect of the present invention, the specific steps for constructing the global judgment matrix are as follows: The temperature transmitters in the collaborative diagnostic transmitter cluster are sorted in ascending order of physical distance. The top N temperature transmitters are used to construct a near-range virtual reference collaborative body. For each physical time t after global alignment, the arithmetic mean of the temperature change rate of these N temperature transmitters at that time is calculated as the reference change characteristic value V[t] of the virtual reference collaborative body at that time. The specific expression of V[t] is as follows: , of which S k [t] represents the temperature measurement value of the k-th temperature transmitter at time t; For each temperature transmitter Trans_i in the collaborative diagnostic transmitter cluster, after sorting, based on its globally aligned real-time temperature sequence, the synchronization response coefficient of that temperature transmitter relative to the nearby virtual reference collaborator is calculated point by point, generating a coefficient sequence aligned with the time axis. This coefficient sequence is the element of the i-th row of the matrix. The specific expression for the synchronization response coefficient is as follows: , where C i [t] represents the synchronization response coefficient of Trans_i at time t; According to the established ascending order rule of physical distance, the synchronous response coefficient sequence of each temperature transmitter is filled into the corresponding position point by point to form the final global judgment matrix.

[0012] As a further aspect of the present invention, the specific operation for fault determination by combining the real-time temperature time-series sequence of the suspected abnormal transmitter is as follows: For the real-time temperature time-series stream sequence S_sus of the suspected abnormal transmitter that has been globally aligned with the same timestamp, the synchronization response coefficient to be verified is calculated point by point using the calculation rule that is completely consistent with the synchronization response coefficient of the temperature transmitter in the global judgment matrix, and the synchronization response coefficient sequence C_sus to be verified is generated. The abnormal time-series anchor point t_ab locked in the initial diagnosis is reused. Taking t_ab as the center, w time intervals are taken forward and w time intervals are taken backward to form a fixed core verification window. The time range covered by this window is [t_start, t_end], where t_start=max(1, t_ab-w) and t_end=min(T-1, t_ab+w), where w is a predetermined fixed window half-width constant and T is the acquisition period. For each physical time t covered by the core verification window, verify whether the positive and negative signs of all synchronization response coefficients in the corresponding column of the global judgment matrix at that time are completely consistent. If they are consistent, the cluster consistency pre-verification is completed. After completing the cluster consistency pre-verification, perform the final fault determination: Within the core verification window, if the synchronous response coefficient sequence C_sus to be verified has at least one coefficient sign at a time that is opposite to the signs of all synchronous response coefficients in the corresponding column of the global judgment matrix at that time, then the suspected abnormal transmitter is determined to have a sampling hardware fault. Within the core verification window, if the signs of the coefficients in the synchronous response coefficient sequence C_sus at all times are consistent with the signs of the corresponding elements in the same column of the global judgment matrix, then the suspected abnormal transmitter is determined to be fault-free, and the abnormal state of the initial diagnostic flag is removed.

[0013] A fault self-diagnosis system for a temperature transmitter includes: The anomaly initial judgment module collects the temperature time-series sequence of the temperature transmitter under test in real time, and performs temperature change detection and time sequence integrity verification on the temperature time-series sequence simultaneously. If both temperature change detection and time sequence integrity verification pass, the temperature transmitter under test is determined to be operating normally. If either verification fails, the temperature transmitter under test is determined to have a suspected fault and is marked as a suspected abnormal transmitter. The cluster filtering module takes suspected abnormal transmitters as the core, retrieves the physical installation coordinate data of transmitters in the industrial field, generates a set of nearest candidate transmitters based on the actual physical distance, and synchronously obtains the historical temperature time series sequence of each candidate temperature transmitter in the set of nearest candidate transmitters. It performs temperature trend filtering with the historical temperature time series sequence of suspected abnormal transmitters to form a collaborative diagnostic transmitter cluster. The fault diagnosis module synchronously collects the real-time temperature time-series stream sequence of suspected abnormal transmitters and each temperature transmitter in the collaborative diagnostic transmitter cluster. At the same time, it performs global timestamp alignment on all real-time temperature time-series stream sequences to construct a global judgment matrix, and completes the fault determination by combining the real-time temperature time-series stream sequence of suspected abnormal transmitters.

[0014] This invention provides a self-diagnosis method and system for temperature transmitter faults, which has the following advantages compared with the prior art: (1) The present invention adopts a dual-dimensional synchronous initial judgment mechanism of temperature change detection and timing integrity verification. Abnormal changes are identified by reversing the change direction and verifying the isolated direction. Combined with the acquisition position ID to verify timing integrity, it can accurately distinguish between transmitter hardware abnormalities and normal operating condition fluctuations, effectively reducing the probability of misjudgment and missed detection. (2) This invention combines the topology and physical distance of the thermal monitoring system to select collaborative devices, and uses HL robust position estimator and Qn dispersion estimator to complete trend homogeneity filtering, and simultaneously completes stability pre-verification and monotonicity verification, eliminating candidate devices with no thermal correlation and unstable data, and constructing a highly adaptable collaborative diagnostic cluster. (3) This invention constructs a standardized global judgment matrix by global alignment of timestamps, calculates the synchronization response coefficient with a virtual reference body, and completes fault judgment based on the consistency of core window symbols, thereby improving the accuracy and practicality of fault diagnosis in industrial fields. Attached Figure Description

[0015] Figure 1 This is a flowchart of the steps of the present invention; Figure 2 This is a flowchart illustrating the steps involved in performing temperature mutation detection according to the present invention. Figure 3 This is the system principle block diagram of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] like Figure 1 This invention provides a self-diagnosis method for temperature transmitter faults; As an embodiment of this application, it includes: S1. Real-time acquisition of the temperature time-series stream sequence of the temperature transmitter under test. Simultaneous temperature change detection and timing integrity verification are performed on the temperature time-series stream sequence. If both temperature change detection and timing integrity verification pass, the temperature transmitter under test is determined to be operating normally. If either verification fails, the temperature transmitter under test is determined to have a suspected fault and is marked as a suspected abnormal transmitter. S2. Taking the suspected abnormal transmitter as the core, retrieve the physical installation coordinate data of the transmitters in the industrial field, generate a set of nearest candidate transmitters based on the actual physical distance, and for each candidate temperature transmitter in the set of nearest candidate transmitters, simultaneously obtain its historical temperature time series sequence, and perform temperature trend filtering with the historical temperature time series sequence of the suspected abnormal transmitter to form a collaborative diagnostic transmitter cluster. S3. Real-time synchronous acquisition of the real-time temperature time-series stream sequence of each temperature transmitter in the suspected abnormal transmitter and the collaborative diagnostic transmitter cluster. At the same time, perform global timestamp alignment on all real-time temperature time-series stream sequences to construct a global judgment matrix, and complete the fault determination by combining the real-time temperature time-series stream sequence of the suspected abnormal transmitter.

[0018] As a second embodiment of this application, it is implemented based on the first embodiment, except that this embodiment includes: S1. Real-time acquisition of the temperature time-series stream sequence of the temperature transmitter under test; The core purpose of initial diagnosis is to detect abnormalities as soon as possible and quickly trigger subsequent collaborative diagnosis. Temperature transmitter failures can occur at any time, and only real-time data acquisition can achieve real-time monitoring and screening. Meanwhile, since subsequent temperature mutation detection and temporal integrity verification cannot rely solely on a single sampling point, but must be based on a continuous range of temperature data to identify mutations and frame drops / skips, it is necessary to acquire a continuous short-period temperature time-series sequence in real time. The specific operation is as follows: A fixed real-time acquisition period T is pre-configured for the temperature transmitter under test. Temperature sampling values ​​are continuously acquired at fixed time intervals within this period, and all sampling values ​​within the period are combined in chronological order to form a complete temperature time series sequence. Using period T as a sliding window, each time a new temperature sampling point is collected, the oldest old sampling point in the time series is removed, and the time series is refreshed in real time while keeping the period length of the temperature time series unchanged, so as to achieve continuous rolling acquisition. Constructing a sliding window update allows the temperature time series sequence to always be in the latest real-time state, ensuring the timeliness of the initial diagnosis without increasing the computational load due to the infinite accumulation of data; Temperature abrupt change detection and temporal integrity verification are performed synchronously on the temperature time-series stream. The specific operation for performing time integrity verification on the temperature time series is as follows: Extract the sampling position IDs of the first and last sampling points in S, respectively, and denote them as ID_start and ID_end. Then, generate a set of integers that increment by 1 consecutively, denoted as Set_std, with ID_start as the starting value and ID_end as the ending value. The acquisition sequence ID is automatically generated synchronously when the temperature transmitter acquisition hardware triggers sampling, and is bound to the sampling point for output, following the principle of "the ID increments by 1 for each acquisition trigger"; Extract the acquisition sequence IDs of all sampling points in S, and form the actual acquisition sequence set according to the order of the sampling points, denoted as Set_act; Perform a full consistency comparison between Set_act and Set_std: if the elements of Set_act and Set_std are completely identical and their order is completely matched, the timing integrity check is considered to have passed; otherwise, the timing integrity check is considered to have failed. If both the temperature change detection and timing integrity verification pass, the temperature transmitter under test is determined to be operating normally, and the diagnostic process is terminated; if either verification fails, the temperature transmitter under test is determined to have a suspected fault, and it is marked as a suspected abnormal transmitter.

[0019] S2. Taking the suspected abnormal transmitter as the core, retrieve the physical installation coordinate data of the transmitters in the industrial field and generate a set of nearest candidate transmitters based on the actual physical distance. The initial diagnostic process identifies the target object to be verified through temperature mutation detection and timing integrity verification. All subsequent collaborative transmitter selection and multi-device comparison operations must revolve around this suspected abnormal transmitter. Only in this way can the entire process be targeted and meaningless full-scale equipment inspection and computing power waste be avoided. The specific operation for generating the nearest neighbor candidate transmitter set is as follows: Extract the inherent thermal monitoring system attributes of the suspected abnormal transmitter, identify the thermal system A directly monitored by the suspected abnormal transmitter, and retrieve the fixed process medium flow path topology data of the industrial site to locate two types of related objects: first, all temperature transmitters installed inside thermal system A; second, temperature transmitters in adjacent process units that have direct medium communication and direct heat exchange with thermal system A and have direct thermal influence related to thermal system A. The thermal system is a fixed process unit in the industrial site with continuous heat conduction and heat exchange relationship, including closed reaction chambers, continuous medium pipelines, complete sets of heat exchange equipment, etc. By combining the two types of temperature transmitters mentioned above, a basic set of related area transmitters is formed. For the set of transmitters in the basic associated area, the three-dimensional physical installation coordinates of each temperature transmitter are extracted, and the actual physical distance L between each temperature transmitter and suspected abnormal transmitters is calculated. Transmitters with L > L are then eliminated. th The temperature transmitter generates the final nearest neighbor candidate transmitter set, where L th This is the spacing threshold, which needs to be set according to the actual situation; Temperature, as a typical spatially correlated physical quantity, exhibits clear spatial conduction and attenuation characteristics in its changes: the closer the physical installation distance between two temperature transmitters, the higher the probability that they are in the same temperature field, under the same heat source, in the same environmental conditions, and affected by the same process disturbances, and the stronger the basis for the comparability of their temperature data; conversely, if the distance between them is too far, it indicates that their temperature fields and process environments are completely independent, and their temperature data have no reference value. For each temperature transmitter in the nearest candidate transmitter set, its historical temperature time series sequence is acquired synchronously, and filtered in the same direction of temperature trend as the historical temperature time series sequence of suspected abnormal transmitters to form a collaborative diagnostic transmitter cluster. The temperature trend homogeneity filtering starts from the long-term trend consistency of historical time series data to verify whether the temperature transmitter is truly in the same temperature field and the same process influence system. It is the core standard for judging whether the equipment can be used as a reliable reference. For temperature transmitters in the same environment and process section, their long-term temperature change trend must be in the same direction: as the overall ambient temperature rises, the temperature transmitters in the same area will show a slight temperature rise trend in sync; and when the process section starts to heat up, the temperature transmitters in the same process section will also show an upward trend in sync; this long-term trend in the same direction is the core premise for the comparability between temperature transmitters. The specific operation for performing temperature trend unidirectional filtering is as follows: The sampling points of the historical sequence S_sus of the suspected abnormal transmitter and the historical sequence S_can of the candidate temperature transmitter are matched one-to-one in chronological order, where the sampling point index i ranges from 0 to n-1. For the historical sequence S_sus of the suspected abnormal transmitter, the first m consecutive sampling points at the beginning of S_sus are extracted to form the first anchoring set A_sus, and the last m consecutive sampling points at the end of S_sus are extracted to form the tail anchoring set B_sus, where m=[k×n], k is the proportional coefficient, and its value range is (0,0.5). For the first anchor set A_sus, calculate HL respectively. A_sus With Qn A_sus Similarly, for the tail anchoring set B_sus, HL is calculated. B_sus With Qn B_sus ; The calculation of HL A_sus With Qn A_sus The specific rules are as follows: HL A_sus Calculate the arithmetic mean of all pairwise sampling points within the initial anchor set A_sus, and take the median of all means; this is HL. A_sus , used to characterize the true temperature center level of the set; Qn A_sus Calculate the absolute difference between all pairwise sampling points within the initial anchor set A_sus, and take the first quartile of all absolute differences. Multiply this quartile by a fixed constant factor d to obtain Qn. A_sus , used to characterize the stability of data within a set; HL based on the first and last anchored sets A_sus HL B_sus The main trend direction identifier Dsus of S_sus is determined by the following rules: If HL B_sus >HL A_sus If HL is increasing, then Dsus is increasing; if HL is increasing, then Dsus is increasing. B_sus <HL A_sus If HL is decreasing, then Dsus is decreasing; B_sus =HL A_sus If so, then Dsus is flat; Using the same processing rules as S_sus described above, the same operation is performed on the historical sequence S_can of the candidate temperature transmitters to obtain the first anchoring set A_can and the last anchoring set B_can, and corresponding to the statistics HL. A_can Qn A_can HL B_can Qn B_can And the main trend direction indicator Dcan; If the candidate temperature transmitter fails the stability pre-verification, or if the main trend direction indicators Dsus and Dcan of the two sequences are not equal, the verification will be terminated directly and the trend homogeneity filter will be deemed to have failed. Otherwise, the intermediate set dual-dimensional verification will be performed. The specific rule for the stability pre-verification is as follows: if the Qn of the candidate temperature transmitter... A_can >Qn A_sus or Qn B_can >Qn B_sus If the data stability of the candidate temperature transmitter is insufficient, the stability pre-verification will fail. The specific steps for performing the intermediate set two-dimensional verification are as follows: After removing the first and last anchor sets from S_sus and S_can, we obtain the intermediate set of the two, and divide the intermediate set into p consecutive subsets of the same length, where p≥2 and are integers. Based on p subsets of S_sus, the robust level sequence of subsets Msus=[HL] is obtained. sus1 HL sus2 ,...,HL susp Based on the main trend direction of Dsus, a monotonicity consistency check is performed on Msus: if Dsus is increasing, check whether Msus is monotonically increasing; if Dsus is decreasing, check whether Msus is monotonically decreasing; if Dsus is flat, check whether all elements of Msus are equal. Using the same rules as described above, the robustness sequence McCan = [HL] is calculated for each of the p subsets of S_can. can1 HL can2 ,...,HL canp ], and perform the same monotonicity consistency check as described above based on Dcan; If the sub-interval robust horizontal sequences of the two sequences S_sus and S_can both pass their respective monotonicity consistency checks, then the same-direction filtering is deemed to have passed and the sequence is included in the final collaborative diagnostic transmitter cluster. Otherwise, the same-direction filtering is deemed to have failed and the sequence is not included in the final collaborative diagnostic transmitter cluster.

[0020] S3. Real-time synchronous acquisition of real-time temperature time series sequences of suspected abnormal transmitters and each temperature transmitter in the collaborative diagnostic transmitter cluster, and simultaneous global timestamp alignment of all real-time temperature time series sequences to construct a global judgment matrix; Temperature transmitters in industrial settings often operate continuously for 24 hours. Even if the sampling interval deviation of a single device is only at the millisecond level, significant sampling point misalignment will accumulate after long-term operation. Global alignment can correct this accumulated deviation in real time, ensuring the stability of diagnostic results under long-term operation. Meanwhile, the initial diagnostic process has marked the time-series anchor points of suspected anomalies. Only by completing the global timestamp alignment can the time-series data of the collaborative temperature transmitter accurately correspond to the anomaly anchor point, focusing on the time window when the anomaly occurred for verification, and avoiding invalid comparisons of the entire sequence. For example, the initial diagnosis marks the abnormal change point at 10:00:05. After global alignment, the synchronization timing window of all temperature transmitters before and after 10:00:05 can be directly locked, and the following global judgment matrix can be constructed based on the data in the window. The specific steps for constructing the global judgment matrix are as follows: The temperature transmitters in the collaborative diagnostic transmitter cluster are sorted in ascending order of physical distance. The top N temperature transmitters are selected to construct a near-distance virtual reference collaborative body. For each physical time t after global alignment, the arithmetic mean of the temperature change rate of these N temperature transmitters at that time is calculated as the reference change characteristic value V[t] of the virtual reference collaborative body at that time. Here, N is a predetermined constant. Because the temperature transmitters that are closest to each other have the strongest spatial correlation with the suspected abnormal transmitters and are most directly affected by the same operating conditions, they naturally have the highest reference reliability. The specific expression for V[t] is: , of which S k [t] represents the temperature measurement value of the k-th temperature transmitter at time t, and V[t] represents the relative temperature change intensity of the most reliable region of the same thermal system at that time, thus eliminating the random fluctuation interference of a single temperature transmitter; For each temperature transmitter Trans_i in the collaborative diagnostic transmitter cluster, after sorting, based on its globally aligned real-time temperature sequence, the synchronization response coefficient of that temperature transmitter relative to the nearby virtual reference collaborator is calculated point by point, generating a coefficient sequence aligned with the time axis. This coefficient sequence is the element of the i-th row of the matrix. The specific expression for the synchronization response coefficient is as follows: , where C i [t] is the synchronization response coefficient of Trans_i at time t, which characterizes the synchronization and relative intensity of temperature changes between the temperature transmitter and the virtual reference body; Under normal operating conditions, temperature transmitters within the same thermal system, affected by the same heat source / process disturbance, exhibit relative temperature changes that are synchronized with the height of the reference body. i [t] will remain stable within a fixed narrow range, and the coefficient will only deviate significantly when Trans_i itself malfunctions; According to the established ascending order rule of physical distance, the synchronous response coefficient sequence of each generated temperature transmitter is filled into the corresponding position point by point to form the final global judgment matrix. The row dimension of the matrix is ​​bound to the physical distance priority. The i-th row corresponds to Trans_i after ascending sort. The smaller the row index, the stronger the spatial correlation with the suspected abnormal transmitter and the higher the reference value. The column dimension of the matrix is ​​bound to the physical time t after global alignment. All elements in the same column are the synchronization response coefficients of different temperature transmitters relative to the virtual reference body at the same time. Fault determination is completed by combining the real-time temperature time-series sequence of the suspected abnormal transmitter. The specific operation is as follows: The near-range virtual reference co-construction body determined in the global judgment matrix construction step and the reference change feature value V[t] corresponding to each physical moment after global alignment with the timestamp of all real-time sequences are reused. For the real-time temperature time series S_sus of the suspected abnormal transmitter that has been processed by global alignment with the same timestamp, the calculation rule that is completely consistent with the temperature transmitter synchronization response coefficient in the global judgment matrix is ​​used to calculate the synchronization response coefficient to be verified point by point, and generate the synchronization response coefficient sequence C_sus to be verified. The abnormal time-series anchor point t_ab locked in the initial diagnosis is reused. Taking t_ab as the center, w time intervals are taken forward and w time intervals are taken backward to form a fixed core verification window. The time range covered by this window is [t_start, t_end], where t_start=max(1, t_ab-w) and t_end=min(T-1, t_ab+w), where w is a predetermined fixed window half-width constant and T is the acquisition period. For each physical time t covered by the core verification window, verify whether the positive and negative signs of all synchronization response coefficients in the corresponding column of the global judgment matrix at that time are completely consistent. If they are consistent, the cluster consistency pre-verification is completed. After completing the cluster consistency pre-verification, perform the final fault determination: Within the core verification window, if the synchronous response coefficient sequence C_sus to be verified has at least one coefficient sign at a time that is opposite to the signs of all synchronous response coefficients in the corresponding column of the global judgment matrix at that time, then the suspected abnormal transmitter is determined to have a sampling hardware fault. Within the core verification window, if the signs of the coefficients in the synchronous response coefficient sequence C_sus at all times are consistent with the signs of the corresponding elements in the same column of the global judgment matrix, then the suspected abnormal transmitter is determined to be fault-free, and the abnormal state of the initial diagnostic flag is removed.

[0021] As a third embodiment of this application, this embodiment further discloses a method for performing temperature sudden change detection based on embodiments one and two, such as... Figure 2 As shown, the specific content includes: For a temperature time-series sequence S, for each sampling point S[i] with index i from 1 to n-1, calculate its change direction attribute relative to its preceding adjacent sampling point S[i-1]. The specific rules are as follows: if S[i] > S[i-1], then the change direction attribute corresponding to this sampling point is increasing; if S[i] = S[i-1], then the change direction attribute corresponding to this sampling point is flat; if S[i] < S[i-1], then the change direction attribute corresponding to this sampling point is decreasing, where n is the total number of sampling points in the sequence S. According to the time sequence of the temperature time series flow sequence S, all the calculated change direction attributes are combined to generate a change direction sequence D that is aligned with the time sequence of S. The index of D is j, j∈[0,n-2] and takes integer values. D[j] uniquely corresponds to the change direction attributes of S[j+1] and S[j]. For the changing direction sequence D, perform adjacent direction reversal verification and isolated direction verification respectively, and output the original sequence sampling point positions of all nodes marked as abnormal direction nodes. If there are no abnormal direction nodes, output an empty result. The specific rules for the adjacent direction reversal verification are as follows: Traverse the sequence of changing directions D. For each element D[j] with index j from 1 to n-2, check the direction combination of D[j] and the previous adjacent element D[j-1]. If the two are directly adjacent combinations of "ascending and descending" or "descending and ascending" (i.e. there are no equal attribute elements between them), then mark the original sequence sampling point position corresponding to the adjacent elements, i.e. S[j], as an abnormal direction node. The specific rules for the isolated direction verification are as follows: Traverse the sequence of changing directions D. For each element D[j] with index j from 1 to n-2, check the consistency between D[j] and its adjacent elements D[j-1] and D[j+1]. If the directions of D[j] are different from those of D[j-1] and D[j+1], and the directions of D[j-1] and D[j+1] are completely consistent, then mark the original sequence sampling point position corresponding to this element, i.e., S[j], as an abnormal direction node. If any marked abnormal direction node exists, the temperature change detection is deemed to fail; otherwise, the temperature change detection is deemed to pass.

[0022] like Figure 3 This invention provides a fault self-diagnosis system for temperature transmitters; As a fourth embodiment of this application, it includes: The anomaly initial judgment module collects the temperature time-series sequence of the temperature transmitter under test in real time, and performs temperature change detection and time sequence integrity verification on the temperature time-series sequence simultaneously. If both temperature change detection and time sequence integrity verification pass, the temperature transmitter under test is determined to be operating normally. If either verification fails, the temperature transmitter under test is determined to have a suspected fault and is marked as a suspected abnormal transmitter. The cluster filtering module takes suspected abnormal transmitters as the core, retrieves the physical installation coordinate data of transmitters in the industrial field, generates a set of nearest candidate transmitters based on the actual physical distance, and synchronously obtains the historical temperature time series sequence of each candidate temperature transmitter in the set of nearest candidate transmitters. It performs temperature trend filtering with the historical temperature time series sequence of suspected abnormal transmitters to form a collaborative diagnostic transmitter cluster. The fault diagnosis module synchronously collects the real-time temperature time-series stream sequence of suspected abnormal transmitters and each temperature transmitter in the collaborative diagnostic transmitter cluster. At the same time, it performs global timestamp alignment on all real-time temperature time-series stream sequences to construct a global judgment matrix, and completes the fault determination by combining the real-time temperature time-series stream sequence of suspected abnormal transmitters.

[0023] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0024] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A self-diagnosis method for a temperature transmitter, characterized in that, include: S1. Real-time acquisition of the temperature time-series stream sequence of the temperature transmitter under test. Simultaneous temperature change detection and timing integrity verification are performed on the temperature time-series stream sequence. If both temperature change detection and timing integrity verification pass, the temperature transmitter under test is determined to be operating normally. If either verification fails, the temperature transmitter under test is determined to have a suspected fault and is marked as a suspected abnormal transmitter. S2. Taking the suspected abnormal transmitter as the core, retrieve the physical installation coordinate data of the transmitters in the industrial field, generate a set of nearest candidate transmitters based on the actual physical distance, and for each candidate temperature transmitter in the set of nearest candidate transmitters, simultaneously obtain its historical temperature time series sequence, and perform temperature trend filtering with the historical temperature time series sequence of the suspected abnormal transmitter to form a collaborative diagnostic transmitter cluster. S3. Real-time synchronous acquisition of the real-time temperature time-series stream sequence of each temperature transmitter in the suspected abnormal transmitter and the collaborative diagnostic transmitter cluster. At the same time, perform global timestamp alignment on all real-time temperature time-series stream sequences to construct a global judgment matrix, and complete the fault determination by combining the real-time temperature time-series stream sequence of the suspected abnormal transmitter.

2. The fault self-diagnosis method for a temperature transmitter according to claim 1, characterized in that, The specific operation for performing temperature abrupt change detection on a temperature time series is as follows: For a temperature time-series sequence S, for each sampling point S[i] with index i from 1 to n-1, calculate its change direction attribute relative to its preceding adjacent sampling point S[i-1]. The specific rules are as follows: if S[i] > S[i-1], then the change direction attribute corresponding to this sampling point is increasing; if S[i] = S[i-1], then the change direction attribute corresponding to this sampling point is flat; if S[i] < S[i-1], then the change direction attribute corresponding to this sampling point is decreasing, where n is the total number of sampling points in the sequence S. According to the time sequence of the temperature time series flow sequence S, all the calculated change direction attributes are combined to generate a change direction sequence D that is aligned with the time sequence of S. The index of D is j, j∈[0,n-2] and takes integer values. D[j] uniquely corresponds to the change direction attributes of S[j+1] and S[j]. For the changing direction sequence D, perform adjacent direction reversal verification and isolated direction verification respectively, and output the original sequence sampling point positions of all nodes marked as abnormal direction nodes. If there are no abnormal direction nodes, output an empty result. If any marked abnormal direction node exists, the temperature change detection is deemed to fail; otherwise, the temperature change detection is deemed to pass.

3. The fault self-diagnosis method for a temperature transmitter according to claim 2, characterized in that, The specific rules for the adjacent direction reversal verification are as follows: traverse the changing direction sequence D, and for each element D[j] with index j from 1 to n-2, verify the direction combination of D[j] and the preceding adjacent element D[j-1]. If the two are directly adjacent combinations of "ascending and descending" or "descending and ascending", then mark the original sequence sampling point position S[j] corresponding to this group of adjacent elements as an abnormal direction node. The specific rules for the isolated direction verification are as follows: traverse the changing direction sequence D, and for each element D[j] with index j from 1 to n-2, verify the consistency of D[j] with the preceding and following adjacent elements D[j-1] and D[j+1]. If the directions of D[j] are not the same as those of D[j-1] and D[j+1], and the directions of D[j-1] and D[j+1] are completely consistent, then mark the original sequence sampling point position S[j] corresponding to this element as an abnormal direction node.

4. The fault self-diagnosis method for a temperature transmitter according to claim 1, characterized in that, The specific steps for performing time sequence integrity verification on a temperature-time-series stream are as follows: Extract the sampling position IDs of the first and last sampling points in S, respectively, and denote them as ID_start and ID_end. Then, generate a set of integers that increment by 1 consecutively, denoted as Set_std, with ID_start as the starting value and ID_end as the ending value. The acquisition sequence ID is automatically generated synchronously by the temperature transmitter acquisition hardware when sampling is triggered, and is bound to the sampling point for output, following the principle of "the ID increments by 1 for each acquisition triggered"; Extract the acquisition sequence IDs of all sampling points in S, and form the actual acquisition sequence set according to the order of the sampling points, denoted as Set_act; Perform a full consistency comparison between Set_act and Set_std: if the elements of Set_act and Set_std are completely identical and their order is completely matched, the timing integrity check is considered to have passed; otherwise, the timing integrity check is considered to have failed.

5. The fault self-diagnosis method for a temperature transmitter according to claim 1, characterized in that, The specific operations for generating a set of nearest neighbor candidate transmitters are as follows: Extract the inherent thermal monitoring system attributes of the suspected abnormal transmitter, identify the thermal system A directly monitored by the suspected abnormal transmitter, and retrieve the fixed process medium flow path topology data of the industrial site to locate two types of related objects: first, all temperature transmitters installed inside thermal system A; second, temperature transmitters in adjacent process units that have direct medium communication and direct heat exchange with thermal system A and have direct thermal influence related to thermal system A. The thermal system is a fixed process unit in the industrial site with continuous heat conduction and heat exchange relationship. The two types of temperature transmitters mentioned above are summarized to form a basic associated region transmitter set. At the same time, the three-dimensional physical installation coordinates of each temperature transmitter in the set are extracted, the actual physical distance L between each temperature transmitter and suspected abnormal transmitters is calculated, and temperature transmitters with L > Lth are eliminated to generate the final nearest neighbor candidate transmitter set, where Lth is the distance threshold.

6. The fault self-diagnosis method for a temperature transmitter according to claim 1, characterized in that, The specific steps for performing temperature trend unidirectional filtering are as follows: The sampling points of the historical sequence S_sus of the suspected abnormal transmitter and the historical sequence S_can of the candidate temperature transmitter are matched one-to-one in chronological order, where the sampling point index i ranges from 0 to n-1. For the historical sequence S_sus of the suspected abnormal transmitter, the first m consecutive sampling points at the beginning of S_sus are extracted to form the first anchoring set A_sus, and the last m consecutive sampling points at the end of S_sus are extracted to form the tail anchoring set B_sus, where m=[k×n], k is the proportional coefficient, and its value range is (0,0.5). For the first anchor set A_sus, calculate HL respectively. A_sus With Qn A_sus Similarly, for the tail anchoring set B_sus, HL is calculated. B_sus With Qn B_sus ; The calculation of HL A_sus With Qn A_sus The specific rules are as follows: HL A_sus Calculate the arithmetic mean of all pairwise sampling points within the initial anchor set A_sus, and take the median of all means; this is HL. A_sus ; Qn A_sus Calculate the absolute difference between all pairwise sampling points within the initial anchor set A_sus, and take the first quartile of all absolute differences. Multiply this quartile by a fixed constant factor d to obtain Qn. A_sus ; HL based on the first and last anchored sets A_sus HL B_sus The main trend direction identifier Dsus of S_sus is determined by the following rules: If HL B_sus >HL A_sus If HL is increasing, then Dsus is increasing; if HL is increasing, then Dsus is increasing. B_sus <HL A_sus If HL is decreasing, then Dsus is decreasing; B_sus =HL A_sus Then Dsus is flat; Using the same processing rules as S_sus described above, the same operation is performed on the historical sequence S_can of the candidate temperature transmitters to obtain the first anchoring set A_can and the last anchoring set B_can, and corresponding to the statistics HL. A_can Qn A_can HL B_can Qn B_can And the main trend direction indicator Dcan; If the candidate temperature transmitter fails the stability pre-verification, or if the main trend direction indicators Dsus and Dcan of the two sequences are not equal, the trend homogeneity filtering is deemed to fail; otherwise, it enters the intermediate set dual-dimensional verification. The specific rule for the stability pre-verification is as follows: if the Qn of the candidate temperature transmitter... A_can >Qn A_sus or Qn B_can >Qn B_sus If so, the stability pre-check fails.

7. A fault self-diagnosis method for a temperature transmitter according to claim 6, characterized in that, The specific steps for performing two-dimensional validation of the intermediate set are as follows: After removing the first and last anchor sets from S_sus and S_can, we obtain the intermediate set of the two, and divide the intermediate set into p consecutive subsets of the same length, where p≥2 and are integers. Based on p subsets of S_sus, the robust level sequence of subsets Msus=[HL] is obtained. sus1 HL sus2 ,...,HL susp Based on the main trend direction of Dsus, a monotonicity consistency check is performed on Msus: if Dsus is increasing, check whether Msus is monotonically increasing; if Dsus is decreasing, check whether Msus is monotonically decreasing; if Dsus is flat, check whether all elements of Msus are equal. Using the same rules as described above, the robustness sequence McCan = [HL] is calculated for each of the p subsets of S_can. can1 HL can2 ,...,HL canp ], and perform the same monotonicity consistency check as described above based on Dcan; If the robust level sequences of the subsets of the two sequences S_sus and S_can both pass their respective monotonicity consistency checks, then the same-direction filtering is considered to pass and the sequence is included in the final collaborative diagnostic transmitter cluster. Otherwise, the same-direction filtering is considered to fail and the sequence is not included in the final collaborative diagnostic transmitter cluster.

8. The fault self-diagnosis method for a temperature transmitter according to claim 1, characterized in that, The specific steps for constructing the global judgment matrix are as follows: The temperature transmitters in the collaborative diagnostic transmitter cluster are sorted in ascending order of physical distance. The top N temperature transmitters are used to construct a near-range virtual reference collaborative body. For each physical time t after global alignment, the arithmetic mean of the temperature change rate of these N temperature transmitters at that time is calculated as the reference change characteristic value V[t] of the virtual reference collaborative body at that time. The specific expression of V[t] is as follows: , among which, S k [t] represents the temperature measurement value of the k-th temperature transmitter at time t; For each temperature transmitter Trans_i in the collaborative diagnostic transmitter cluster, after sorting, based on its globally aligned real-time temperature sequence, the synchronization response coefficient of that temperature transmitter relative to the nearby virtual reference collaborator is calculated point by point, generating a coefficient sequence aligned with the time axis. This coefficient sequence is the element of the i-th row of the matrix. The specific expression for the synchronization response coefficient is as follows: , where C i [t] represents the synchronization response coefficient of Trans_i at time t; According to the established ascending order rule of physical distance, the synchronous response coefficient sequence of each temperature transmitter is filled into the corresponding position point by point to form the final global judgment matrix.

9. A fault self-diagnosis method for a temperature transmitter according to claim 1, characterized in that, The specific steps for fault diagnosis based on the real-time temperature time-series stream of suspected faulty transmitters are as follows: For the real-time temperature time-series stream sequence S_sus of the suspected abnormal transmitter that has been globally aligned with the same timestamp, the synchronization response coefficient to be verified is calculated point by point using the calculation rule that is completely consistent with the synchronization response coefficient of the temperature transmitter in the global judgment matrix, and the synchronization response coefficient sequence C_sus to be verified is generated. The abnormal time-series anchor point t_ab locked in the initial diagnosis is reused. Taking t_ab as the center, w time intervals are taken forward and w time intervals are taken backward to form a fixed core verification window. The time range covered by this window is [t_start, t_end], where t_start=max(1, t_ab-w) and t_end=min(T-1, t_ab+w), where w is a predetermined fixed window half-width constant and T is the acquisition period. For each physical time t covered by the core verification window, verify whether the positive and negative signs of all synchronization response coefficients in the corresponding column of the global judgment matrix at that time are completely consistent. If they are consistent, the cluster consistency pre-verification is completed. After completing the cluster consistency pre-verification, perform the final fault determination: Within the core verification window, if the synchronous response coefficient sequence C_sus to be verified has at least one coefficient sign at a time that is opposite to the signs of all synchronous response coefficients in the corresponding column of the global judgment matrix at that time, then the suspected abnormal transmitter is determined to have a sampling hardware fault. Within the core verification window, if the signs of the coefficients in the synchronous response coefficient sequence C_sus at all times are consistent with the signs of the corresponding elements in the same column of the global judgment matrix, then the suspected abnormal transmitter is determined to be fault-free, and the abnormal state of the initial diagnostic flag is removed.

10. A fault self-diagnosis system for a temperature transmitter, used to execute the fault self-diagnosis method for a temperature transmitter according to any one of claims 1-9, characterized in that, include: The anomaly initial judgment module collects the temperature time-series sequence of the temperature transmitter under test in real time, and performs temperature change detection and time sequence integrity verification on the temperature time-series sequence simultaneously. If both temperature change detection and time sequence integrity verification pass, the temperature transmitter under test is determined to be operating normally. If either verification fails, the temperature transmitter under test is determined to have a suspected fault and is marked as a suspected abnormal transmitter. The cluster filtering module takes suspected abnormal transmitters as the core, retrieves the physical installation coordinate data of transmitters in the industrial field, generates a set of nearest candidate transmitters based on the actual physical distance, and synchronously obtains the historical temperature time series sequence of each candidate temperature transmitter in the set of nearest candidate transmitters. It performs temperature trend filtering with the historical temperature time series sequence of suspected abnormal transmitters to form a collaborative diagnostic transmitter cluster. The fault diagnosis module synchronously collects the real-time temperature time-series stream sequence of suspected abnormal transmitters and each temperature transmitter in the collaborative diagnostic transmitter cluster. At the same time, it performs global timestamp alignment on all real-time temperature time-series stream sequences to construct a global judgment matrix, and completes the fault determination by combining the real-time temperature time-series stream sequence of suspected abnormal transmitters.

Images