Intelligent temperature point online detection system and method of suction dredger
By installing multiple temperature acquisition units and LoRa wireless networks on the mud suction machine, combined with an isolated forest algorithm of ARM controller and data center platform, comprehensive monitoring and intelligent analysis of the mud suction machine temperature is realized, and temperature detection in the existing technology is solved The incomplete problem improves the operating safety and life of the equipment.
Patent Information
- Application Number
- CN202411137341.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-26
AI Technical Summary
The existing mud suction machine temperature detection system cannot fully monitor the equipment temperature, lacks intelligent analysis capabilities, cannot predict temperature change trends and automatically identify abnormal situations, resulting in equipment failure and unstable operation.
An intelligent temperature point online detection system was designed, using multiple temperature acquisition units to communicate with the data terminal through the 485 bus, using the LoRa wireless network to transmit temperature data in real time, combined with the ARM controller and the data center platform, and using an isolated forest algorithm for data analysis and prediction, implement the hierarchical alarm function.
It realizes comprehensive detection of mud suction machine temperature and intelligent data analysis, timely discover potential faults, reduces the frequency of manual inspection and maintenance, reduces maintenance costs, and improves the safe operation and life of the equipment.
Smart Images

Figure CN119023096A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of electronic technology, and in particular relates to an intelligent temperature point online detection system and method for a sludge suction machine. Background Art
[0002] The sludge suction machine is one of the key equipment in the sewage treatment plant. It is used to suck out the sludge at the bottom of the sedimentation tank and transport it to the designated location. Since the sludge suction machine will generate a lot of heat during long-term operation, if the temperature is too high, it will cause equipment failure and affect normal operation. Temperature monitoring is of great significance to ensure the safe operation of the sludge suction machine, extend the life of the equipment, and improve the efficiency of sewage treatment.
[0003] The shortcomings of temperature detection of suction sludge machine in the existing technology:
[0004] 1. Traditional temperature detection systems often only install a small number of temperature sensors in key locations, which cannot fully monitor the overall temperature of the equipment and result in blind spots.
[0005] 2. Many temperature detection systems can only perform simple temperature monitoring, lack intelligent analysis capabilities, cannot predict temperature change trends, and cannot automatically identify abnormal situations. Summary of the invention
[0006] The purpose of the present invention is to provide an intelligent temperature point online detection system and method for a sludge suction machine, which solves the technical problems of comprehensive detection, intelligent data analysis and graded alarm of the temperature point detection system of the sludge suction machine.
[0007] To achieve the above object, the present invention adopts the following technical solution:
[0008] An intelligent temperature point online detection system for a sludge suction machine includes a perception layer, a transmission layer, an application layer and a data publishing layer;
[0009] The sensing layer includes several temperature acquisition units, which include data terminals and several temperature detection units. All temperature detection units communicate with the data terminals through the 485 bus.
[0010] The data terminal includes a 485 module, an ARM controller, a touch screen, a LoRa-DTU module and a switching power supply. The 485 module and the LoRa-DTU module communicate with different serial ports of the ARM controller through different serial port lines respectively. The touch screen is connected to the ARM controller through an SPI interface. The switching power supply supplies power for the touch screen, the 485 module, the ARM controller and the LoRa-DTU module.
[0011] The temperature detection unit includes an acquisition module, a power module and a temperature sensor group. The temperature sensor group includes several temperature sensors. The temperature sensor is electrically connected to the acquisition module through a wire. The acquisition module is connected to the 485 module through a 485 bus. The power module supplies power to the temperature detection unit.
[0012] The transport layer provides the LoRa wireless network;
[0013] The application layer includes industrial computers and data center platforms. The industrial computers communicate with the data center platform through network cables. The industrial computers are also connected to a wireless DTU through a serial port cable. The wireless DTU communicates with the LoRa-DTU module through the LoRa wireless network.
[0014] The data center platform publishes data to the mobile phone or client in the data publishing layer.
[0015] Preferably, the model of the 485 module is MAX485, the model of the ARM controller is STM32F103C8T6, the model of the LoRa-DTU module is USR-LG206-L LoRa wireless DTU, and the switching power supply is a switching power supply with an output voltage of 12V and 5V.
[0016] Preferably, the model of the acquisition module is TD-4015S, the model of the temperature sensor is PT100, and the power supply module is a 12V output switching power supply.
[0017] Preferably, the model of the industrial computer is WA-15ZYBGDR industrial computer, and the model of the wireless DTU is USR-LG206-L LoRa wireless DTU.
[0018] An intelligent temperature point online detection method for a sludge suction machine specifically comprises the following steps:
[0019] Step 1: Establish an intelligent temperature point online detection system for the sludge suction machine. In the perception layer, the temperature sensors of the temperature detection unit are installed at different temperature collection points on the sludge suction machine to collect temperature data. The temperature data is transmitted to the 485 bus through the collection module. The ARM controller receives the temperature data through the 485 module. The ARM controller transmits the temperature data to the industrial computer through the LoRa-DTU module;
[0020] The ARM controller generates a unique ID number for each temperature sensor. The user sets the location information of each temperature sensor through the touch screen according to the location of the temperature collection point. The ARM controller establishes a corresponding list between the location information and the ID number of the temperature sensor, generates an installation location list, and transmits the installation location list to the industrial computer through the LoRa-DTU module.
[0021] The data transmission layer is responsible for providing the LoRa wireless network, which is used for data interaction between the ARM controller and the industrial computer;
[0022] Step 2: The industrial computer receives temperature data and the installation location list through the wireless DTU, stores the temperature data locally, generates a temperature topology network based on the installation location list, puts the temperature data into the temperature topology network for display, and transmits the temperature data and the installation location list to the data center platform through the network cable;
[0023] Step 3: The data center platform receives the temperature data and installation location list from the industrial computer; stores the temperature data in the database according to the timestamp and sensor location, and generates historical temperature data;
[0024] Step 4: The data center platform preprocesses the temperature data;
[0025] Step 5: The data center platform performs anomaly detection on the preprocessed data, including establishing an isolation forest model and training the isolation forest model using historical temperature data;
[0026] The real-time temperature data is input into the trained isolation forest model and the anomaly score s(T i );
[0027] Step 6: Set the position weight for the installation position of the temperature sensor, including the following steps:
[0028] Step 6-1: Important location alarm setting, specifically including first setting the abnormal threshold value θ of the important location imp , and then for each data point at an important location, obtain its anomaly score s(T i ), and then judge one by one: If the abnormal score s(T i ) reaches the abnormal threshold θ imp , then the alarm is triggered immediately;
[0029] Step 6-2: Secondary position alarm setting, specifically including first setting the abnormal threshold value θ of the secondary position sec and the detection window size N sec , then for the secondary location, the detection window is used to calculate the comprehensive anomaly score S of multiple sensor data points sec , and then if the comprehensive anomaly score S sec Reaching the abnormal threshold θ sec , then trigger the alarm;
[0030] Step 7: At the data publishing layer, the industrial computer displays temperature data and abnormal status in real time; when an abnormality is detected, the data center platform publishes alarm information to the mobile phone or client through the network.
[0031] Preferably, when executing step 5, the specific steps are as follows:
[0032] Step 5-1: Use the preprocessed temperature data to construct a historical temperature dataset of size n;
[0033] Step 5-2: Use the historical data set to train the isolation forest model and build an isolation tree;
[0034] Step 5-3: Collect real-time temperature data and input it into the trained isolation forest model;
[0035] Step 5-4: For each real-time temperature data point T i , input it into each isolated tree, the path length from the root node to the leaf node is h j (T i );
[0036] Step 5-5: Use the average path length formula to calculate the average path length of each data point. The specific formula is as follows:
[0037]
[0038] Among them, T i is the preprocessed temperature data point, E(h(T i )) is the temperature data point T i The average path length among all isolated trees, t is the number of isolated trees in the isolation forest, and h is i (T i ) is the temperature data point T i The path length in the i-th isolated tree, that is, from the root node to the temperature data point T i The number of splits required; j is the variable involved in the calculation;
[0039] Step 5-6: Calculate the adjustment factor c(n) using the historical data set size n. The specific formula is as follows:
[0040]
[0041] Where H(i) is the i-th harmonic number, c(n) is the adjustment factor of the dataset size n, used to normalize the path length, and n is the size of the historical temperature dataset;
[0042] Step 5-7: Calculate the anomaly score for each data point using the average path length and adjustment factor. The specific formula is as follows:
[0043]
[0044] Among them, s(T i ) is the temperature data point T iThe anomaly score is between 0 and 1.
[0045] Preferably, when executing step 6-2, the comprehensive anomaly score S sec The calculation formula is as follows:
[0046]
[0047] Among them, the detection window size N sec The priority is determined based on the priority of the temperature sensor, which is grouped according to the importance of the installation location of the temperature sensor. i ) takes the value s(T i ), that is, the sample point T i The anomaly score.
[0048] The intelligent temperature point online detection system and method of the sludge suction machine described in the present invention solves the technical problems of comprehensive detection, intelligent data analysis and hierarchical alarm of the temperature point detection system of the sludge suction machine. The perception layer of the present invention includes multiple temperature acquisition units, each unit is equipped with multiple temperature sensors, and communicates with the data terminal through the 485 bus to cover various key parts of the equipment to achieve comprehensive monitoring. The LoRa wireless communication technology is adopted to achieve real-time transmission of temperature data through the LoRa-DTU module to ensure the timeliness and reliability of data transmission. The ARM controller can be responsible for data collection and preliminary processing, and upload data through the LoRa-DTU module. In the data center platform, the temperature data is analyzed and predicted using a machine learning algorithm (such as the isolation forest algorithm). The temperature change model is established through historical data, and anomalies in temperature data are detected in real time. Intelligent early warning functions are provided to timely discover and handle potential faults. Weights are set according to the location of temperature sensors. Important locations will immediately alarm once an abnormality occurs. Secondary locations will make comprehensive judgments and alarms based on the abnormal conditions of multiple points, effectively reducing the false alarm rate and improving the accuracy of alarms. Temperature monitoring data is released to mobile phones or clients through the data center platform to achieve remote monitoring and management. Users can view the operating status and temperature of the equipment anytime and anywhere. The system is modularly designed, and sensors and data acquisition units are easy to replace and maintain. Through remote monitoring and intelligent analysis, the frequency of manual inspection and maintenance is reduced, and maintenance costs are reduced. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Figure 1 is a schematic diagram of the system architecture of the present invention;
[0050] Figure 2 It is a schematic block diagram of the data terminal and the temperature acquisition unit of the present invention;
[0051] Figure 3 It is the main flow chart of the present invention. DETAILED DESCRIPTION
[0052] Embodiment 1:
[0053] like Figure 1-2 The intelligent temperature point online detection system of a sludge suction machine shown in the figure includes a perception layer, a transmission layer, an application layer and a data publishing layer;
[0054] The sensing layer includes several temperature acquisition units, which include data terminals and several temperature detection units. All temperature detection units communicate with the data terminals through the 485 bus.
[0055] The data terminal includes a 485 module, an ARM controller, a touch screen, a LoRa-DTU module and a switching power supply. The 485 module and the LoRa-DTU module communicate with different serial ports of the ARM controller through different serial port lines respectively. The touch screen is connected to the ARM controller through an SPI interface. The switching power supply supplies power for the touch screen, the 485 module, the ARM controller and the LoRa-DTU module.
[0056] In this embodiment, the touch screen model may be ILI9341 TFT LCD touch screen.
[0057] The temperature detection unit includes an acquisition module, a power module and a temperature sensor group. The temperature sensor group includes several temperature sensors. The temperature sensor is electrically connected to the acquisition module through a wire. The acquisition module is connected to the 485 module through a 485 bus. The power module supplies power to the temperature detection unit.
[0058] Each temperature sensor (PT100) collects temperature data and transmits it to the 485 bus through the acquisition module (TD-4015S).
[0059] The ARM controller (STM32F103C8T6) receives the data from all temperature detection units through the 485 module (MAX485), and transmits the data to the industrial computer (WA-15ZYBGDR industrial computer) through the LoRa-DTU module (USR-LG206-L LoRa wireless DTU).
[0060] The transport layer provides the LoRa wireless network; the LoRa wireless network transmits temperature data from the ARM controller to the industrial computer, ensuring stable data transmission and low latency.
[0061] The application layer includes industrial computers and data center platforms. The industrial computers communicate with the data center platform through network cables. The industrial computers are also connected to a wireless DTU through a serial port cable. The wireless DTU communicates with the LoRa-DTU module through the LoRa wireless network.
[0062] The data center platform publishes data to the mobile phone or client in the data publishing layer.
[0063] The temperature data is received through the wireless DTU (USR-LG206-L LoRa wireless DTU), stored and displayed locally, and the data is transmitted to the data center platform via the network cable.
[0064] The data center platform is the core of data early warning and prediction. It is responsible for data reception and storage, data preprocessing, building an isolation forest for anomaly detection, setting weights for important and secondary locations, and providing graded alarms based on important and secondary locations.
[0065] The model of the 485 module is MAX485, the model of the ARM controller is STM32F103C8T6, the model of the LoRa-DTU module is USR-LG206-L LoRa wireless DTU, and the switching power supply is a switching power supply with an output voltage of 12V and 5V.
[0066] The model of the acquisition module is TD-4015S, the model of the temperature sensor is PT100, and the power supply module is a 12V output switching power supply.
[0067] The model of the industrial computer is WA-15ZYBGDR industrial computer, and the model of the wireless DTU is USR-LG206-L LoRa wireless DTU.
[0068] Embodiment 2:
[0069] like Figure 3 As shown, the intelligent temperature point online detection method of a sludge suction machine described in Example 2 is implemented on the basis of the intelligent temperature point online detection system of a sludge suction machine described in Example 1, and specifically includes the following steps:
[0070] Step 1: Establish an intelligent temperature point online detection system for the sludge suction machine. In the perception layer, the temperature sensors of the temperature detection unit are installed at different temperature collection points on the sludge suction machine to collect temperature data. The temperature data is transmitted to the 485 bus through the collection module. The ARM controller receives the temperature data through the 485 module. The ARM controller transmits the temperature data to the industrial computer through the LoRa-DTU module;
[0071] The ARM controller generates a unique ID number for each temperature sensor. The user sets the location information of each temperature sensor through the touch screen according to the location of the temperature collection point. The ARM controller establishes a corresponding list between the location information and the ID number of the temperature sensor, generates an installation location list, and transmits the installation location list to the industrial computer through the LoRa-DTU module.
[0072] The data transmission layer is responsible for providing the LoRa wireless network, which is used for data interaction between the ARM controller and the industrial computer;
[0073] Step 2: The industrial computer receives temperature data and the installation location list through the wireless DTU, stores the temperature data locally, generates a temperature topology network based on the installation location list, puts the temperature data into the temperature topology network for display, and transmits the temperature data and the installation location list to the data center platform through the network cable;
[0074] Step 3: The data center platform receives the temperature data and installation location list from the industrial computer; stores the temperature data in the database according to the timestamp and sensor location, and generates historical temperature data;
[0075] Step 4: The data center platform preprocesses the temperature data;
[0076] In this embodiment, the moving average method is used to preprocess the data. For the temperature data of the time series {T1, T2, ..., T n}, moving average M i The specific formula is as follows:
[0077]
[0078] In this formula, w is the size of the smoothing window, that is, how many data points are taken at a time to calculate the average value. For each data point in the time series data, the average value of the point and the previous w-1 data points is calculated. After preprocessing, the stability of the data is guaranteed.
[0079] Step 5: The data center platform performs anomaly detection on the preprocessed data, including establishing an isolation forest model and training the isolation forest model using historical temperature data;
[0080] The real-time temperature data is input into the trained isolation forest model and the anomaly score s(T i );
[0081] When executing step 5, the specific steps are as follows:
[0082] Step 5-1: Use the preprocessed temperature data to construct a historical temperature dataset of size n;
[0083] Step 5-2: Use the historical data set to train the isolation forest model and build an isolation tree;
[0084] Step 5-3: Collect real-time temperature data and input it into the trained isolation forest model;
[0085] Step 5-4: For each real-time temperature data point T i, input it into each isolated tree, the path length from the root node to the leaf node is h j (T i );
[0086] Step 5-5: Use the average path length formula to calculate the average path length of each data point. The specific formula is as follows:
[0087]
[0088] Among them, T i is the preprocessed temperature data point, E(h(T i )) is the temperature data point T i The average path length among all isolated trees, t is the number of isolated trees in the isolation forest, and h is i (T i ) is the temperature data point T i The path length in the i-th isolated tree, that is, from the root node to the temperature data point T i The number of splits required; j is the variable involved in the calculation;
[0089] Step 5-6: Calculate the adjustment factor c(n) using the historical data set size n. The specific formula is as follows:
[0090]
[0091] Where H(i) is the i-th harmonic number, c(n) is the adjustment factor of the dataset size n, used to normalize the path length, and n is the size of the historical temperature dataset;
[0092] In this embodiment, H(i) is approximately ln(i)+γ, where γ is the Euler-Marshall constant.
[0093] Step 5-7: Calculate the anomaly score for each data point using the average path length and adjustment factor. The specific formula is as follows:
[0094]
[0095] Among them, s(T i ) is the temperature data point T i The anomaly score is between 0 and 1.
[0096] Step 6: Set the position weight for the installation position of the temperature sensor. In this embodiment, the weight is set according to the importance of the sensor position:
[0097] Important location: Directly use the anomaly score of the isolation forest model for judgment.
[0098] Secondary position: Combine the anomaly scores of multiple sensors for judgment.
[0099] The specific steps include:
[0100] Step 6-1: Important location alarm setting, specifically including first setting the abnormal threshold value θ of the important location imp , and then for each data point at an important location, obtain its anomaly score s(T i ), and then judge one by one: If the abnormal score s(T i ) reaches the abnormal threshold θ imp , then the alarm is triggered immediately;
[0101] In actual application, the following formula is used for judgment:
[0102] if s(x)>θ imp , then alert.
[0103] Among them, s(x) is the abnormal score s(T i ).
[0104] Step 6-2: Secondary position alarm setting, specifically including first setting the abnormal threshold value θ of the secondary position sec and the detection window size N sec , then for the secondary location, the detection window is used to calculate the comprehensive anomaly score S of multiple sensor data points sec , and then if the comprehensive anomaly score S sec Reaching the abnormal threshold θ sec , then trigger the alarm;
[0105] When executing step 6-2, the comprehensive anomaly score S sec The calculation formula is as follows:
[0106]
[0107] Among them, the detection window size N sec The priority is determined based on the priority of the temperature sensor, which is grouped according to the importance of the installation location of the temperature sensor. i ) takes the value s(T i ), that is, the sample point T i The anomaly score.
[0108] In actual application, the following formula is used for judgment:
[0109] if S sec >θ sec , then alert.
[0110] Step 7: At the data publishing layer, the industrial computer displays temperature data and abnormal status in real time; when an abnormality is detected, the data center platform publishes alarm information to the mobile phone or client through the network.
[0111] The intelligent temperature point online detection system and method of the sludge suction machine described in the present invention solves the technical problems of comprehensive detection, intelligent data analysis and hierarchical alarm of the temperature point detection system of the sludge suction machine. The perception layer of the present invention includes multiple temperature acquisition units, each unit is equipped with multiple temperature sensors, and communicates with the data terminal through the 485 bus to cover various key parts of the equipment to achieve comprehensive monitoring. The LoRa wireless communication technology is adopted to achieve real-time transmission of temperature data through the LoRa-DTU module to ensure the timeliness and reliability of data transmission. The ARM controller can be responsible for data collection and preliminary processing, and upload data through the LoRa-DTU module. In the data center platform, the temperature data is analyzed and predicted using a machine learning algorithm (such as the isolation forest algorithm). The temperature change model is established through historical data, and anomalies in temperature data are detected in real time. Intelligent early warning functions are provided to timely discover and handle potential faults. Weights are set according to the location of temperature sensors. Important locations will immediately alarm once an abnormality occurs. Secondary locations will make comprehensive judgments and alarms based on the abnormal conditions of multiple points, effectively reducing the false alarm rate and improving the accuracy of alarms. Temperature monitoring data is released to mobile phones or clients through the data center platform to achieve remote monitoring and management. Users can view the operating status and temperature of the equipment anytime and anywhere. The system is modularly designed, and sensors and data acquisition units are easy to replace and maintain. Through remote monitoring and intelligent analysis, the frequency of manual inspection and maintenance is reduced, and maintenance costs are reduced.
Claims
1. An intelligent temperature point online detection system for a sludge suction machine, characterized in that: Includes perception layer, transport layer, application layer and data publishing layer; The sensing layer includes several temperature acquisition units, which include data terminals and several temperature detection units. All temperature detection units communicate with the data terminals through the 485 bus. The data terminal includes a 485 module, an ARM controller, a touch screen, a LoRa-DTU module and a switching power supply. The 485 module and the LoRa-DTU module communicate with different serial ports of the ARM controller through different serial port lines respectively. The touch screen is connected to the ARM controller through an SPI interface. The switching power supply supplies power for the touch screen, the 485 module, the ARM controller and the LoRa-DTU module. The temperature detection unit includes an acquisition module, a power module and a temperature sensor group. The temperature sensor group includes several temperature sensors. The temperature sensor is electrically connected to the acquisition module through a wire. The acquisition module is connected to the 485 module through a 485 bus. The power module supplies power to the temperature detection unit. The transport layer provides the LoRa wireless network; The application layer includes industrial computers and data center platforms. The industrial computers communicate with the data center platform through network cables. The industrial computers are also connected to a wireless DTU through a serial port cable. The wireless DTU communicates with the LoRa-DTU module through the LoRa wireless network. The data center platform publishes data to the mobile phone or client in the data publishing layer.
2. The intelligent temperature point online detection system of a sludge suction machine according to claim 1 is characterized in that: The model of the 485 module is MAX485, the model of the ARM controller is STM32F103C8T6, the model of the LoRa-DTU module is USR-LG206-L LoRa wireless DTU, and the switching power supply is a switching power supply with an output voltage of 12V and 5V.
3. The intelligent temperature point online detection system of a sludge suction machine according to claim 1 is characterized in that: The model of the acquisition module is TD-4015S, the model of the temperature sensor is PT100, and the power supply module is a 12V output switching power supply.
4. The intelligent temperature point online detection system of a sludge suction machine according to claim 1 is characterized in that: The model of the industrial computer is WA-15ZYBGDR industrial computer, and the model of the wireless DTU is USR-LG206-L LoRa wireless DTU.
5. An intelligent temperature point online detection method for a sludge suction machine, applied to an intelligent temperature point online detection system for a sludge suction machine according to any one of claims 1 to 4, characterized in that: The specific steps include: Step 1: Establish an intelligent temperature point online detection system for the sludge suction machine. In the perception layer, the temperature sensors of the temperature detection unit are installed at different temperature collection points on the sludge suction machine to collect temperature data. The temperature data is transmitted to the 485 bus through the collection module. The ARM controller receives the temperature data through the 485 module. The ARM controller transmits the temperature data to the industrial computer through the LoRa-DTU module; The ARM controller generates a unique ID number for each temperature sensor. The user sets the location information of each temperature sensor through the touch screen according to the location of the temperature collection point. The ARM controller establishes a corresponding list between the location information and the ID number of the temperature sensor, generates an installation location list, and transmits the installation location list to the industrial computer through the LoRa-DTU module. The data transmission layer is responsible for providing the LoRa wireless network, which is used for data interaction between the ARM controller and the industrial computer; Step 2: The industrial computer receives temperature data and the installation location list through the wireless DTU, stores the temperature data locally, generates a temperature topology network based on the installation location list, puts the temperature data into the temperature topology network for display, and transmits the temperature data and the installation location list to the data center platform through the network cable; Step 3: The data center platform receives the temperature data and installation location list from the industrial computer; stores the temperature data in the database according to the timestamp and sensor location, and generates historical temperature data; Step 4: The data center platform preprocesses the temperature data; Step 5: The data center platform performs anomaly detection on the preprocessed data, including establishing an isolation forest model and training the isolation forest model using historical temperature data; The real-time temperature data is input into the trained isolation forest model and the anomaly score s(T i ); Step 6: Set the position weight for the installation position of the temperature sensor, including the following steps: Step 6-1: Important location alarm setting, specifically including first setting the abnormal threshold value θ of the important location imp , and then for each data point at an important location, obtain its anomaly score s(T i ), and then judge one by one: If the abnormal score s(T i ) reaches the abnormal threshold θ imp , then the alarm is triggered immediately; Step 6-2: Secondary position alarm setting, specifically including first setting the abnormal threshold value θ of the secondary position sec and the detection window size N sec , then for the secondary location, the detection window is used to calculate the comprehensive anomaly score S of multiple sensor data points sec , and then if the comprehensive anomaly score S sec Reaching the abnormal threshold θ sec , then trigger the alarm; Step 7: At the data publishing layer, the industrial computer displays temperature data and abnormal status in real time; when an abnormality is detected, the data center platform publishes alarm information to the mobile phone or client through the network.
6. The intelligent online temperature point detection method of a sludge suction machine as claimed in claim 5, characterized in that: When executing step 5, the specific steps are as follows: Step 5-1: Use the preprocessed temperature data to construct a historical temperature dataset of size n; Step 5-2: Use the historical data set to train the isolation forest model and build an isolation tree; Step 5-3: Collect real-time temperature data and input it into the trained isolation forest model; Step 5-4: For each real-time temperature data point T i , input it into each isolated tree, the path length from the root node to the leaf node is h j (T i ); Step 5-5: Use the average path length formula to calculate the average path length of each data point. The specific formula is as follows: Among them, T i is the preprocessed temperature data point, E(h(T i )) is the temperature data point T i The average path length among all isolated trees, t is the number of isolated trees in the isolation forest, and h is i (T i ) is the temperature data point T i The path length in the i-th isolated tree, that is, from the root node to the temperature data point T i The number of splits required; j is the variable involved in the calculation; Step 5-6: Calculate the adjustment factor c(n) using the historical data set size n. The specific formula is as follows: Where H(i) is the i-th harmonic number, c(n) is the adjustment factor of the dataset size n, used to normalize the path length, and n is the size of the historical temperature dataset; Step 5-7: Calculate the anomaly score for each data point using the average path length and adjustment factor. The specific formula is as follows: Among them, s(T i ) is the temperature data point T i The anomaly score is between 0 and 1.
7. The intelligent online temperature point detection method of a sludge suction machine according to claim 6, characterized in that: When executing step 6-2, the comprehensive anomaly score S sec The calculation formula is as follows: Among them, the detection window size N sec The priority is determined based on the priority of the temperature sensor, which is grouped according to the importance of the installation location of the temperature sensor. i ) takes the value s(T i ), that is, the sample point T i The anomaly score.
