Abnormal fault alarm device and alarm method for energy storage power station
Through data collection and machine learning models, the voltage, current and temperature correlation models of energy storage power stations are built, which solves the problems of fault detection delay and misjudgment of energy storage power stations, and realizes the timely accuracy of fault alarms and the efficiency of fault positioning. , reducing the failure rate and operation and maintenance costs.
Patent Information
- Application Number
- CN202410965060.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-22
AI Technical Summary
The existing technology cannot accurately determine abnormal faults of energy storage power plants, resulting in delays in fault detection or misjudgment, affecting power plants' safety and economic losses.
Through data collection and processing, the energy storage power station area is divided, and the machine learning prediction model and extreme random forest algorithm are used to construct a correlation model between voltage, current and temperature, and monitor and issue abnormal fault alarms in real time.
It realizes timely and accurate fault alarms, narrows the scope of troubleshooting, improves fault positioning accuracy, reduces fault rate and operation and maintenance costs, and protects the safety of life and property.
Smart Images

Figure CN119007395A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of abnormal fault detection, and in particular relates to an abnormal fault alarm device and an alarm method for an energy storage power station. Background Art
[0002] With the promotion of carbon neutrality, carbon peak, energy conservation and consumption reduction, the energy storage market has entered a fast lane of development. However, the risk of safety accidents in energy storage power stations has also increased. If the battery is in a state of high temperature, high voltage, and abnormal current for a long time, it will not only affect its service life and performance, but may also cause serious electrical fires and economic losses. Therefore, ensuring the safe and stable operation of energy storage power stations and preventing the occurrence and spread of abnormal faults are important directions for the development of energy storage technology;
[0003] Energy storage power stations store a large number of batteries in a limited space. These batteries have the characteristics of large fire load and high fire hazard. Once the battery thermal runaway occurs, it is difficult to extinguish, and it is particularly easy to form a systemic fire, and even cause damage to the entire energy storage power station. Therefore, it is particularly important to issue timely and accurate alarms for abnormal failures in energy storage power stations. Summary of the invention
[0004] The object of the present invention is to provide an abnormal fault alarm device and an alarm method for an energy storage power station, which are used to solve the technical problem that the abnormal fault of the energy storage power station cannot be accurately judged in the prior art.
[0005] In order to achieve the above object, the present invention adopts the following technical solutions:
[0006] Step 1: Collect and process relevant data of key parts of the energy storage power station through relevant data collection modules; collect the operating status, power, voltage, and temperature information of the power generation facilities and key equipment of the energy storage device of the energy storage power station, clarify the monitoring needs of the energy storage power station, and determine the key parts to be monitored and the data to be monitored. The key parts specifically refer to: energy storage battery packs, radiators and connectors of energy storage equipment; the relevant data specifically refer to: historical and real-time voltage, current, and temperature data of battery packs and energy storage equipment; clean and normalize the collected data to convert data of different units such as voltage, current, and temperature to the same scale.
[0007] Step 2: Divide the energy storage power station area through the regional division module, classify the processed data and mark the regional source, and transmit it to the data processing cloud remotely; determine the installation location, quantity and data transmission method of the sensor. The data transmission method adopts mobile communication data transmission, WIFI or optical fiber transmission. The area is divided according to the data status of key parts under the normal operation of the energy storage power station. The normal operation state refers to the operation state without any abnormal faults. The specific method of dividing the area is as follows:
[0008] According to the collected data on the key parts of the energy storage power station, the time range t is evenly divided into n time periods to obtain the voltage, current and temperature values of each time period. The variation range of the voltage, current and temperature values of the key part A under the normal operation of the energy storage power station is obtained as follows:
[0009]
[0010] Among them, a min with a max They represent the minimum and maximum voltages of the key part A within time t, and b min With b max They represent the minimum and maximum current of the key part A within time t, c min With c max They represent the minimum and maximum temperature of the key part within time t, respectively, and set the deviation value n of voltage, current and temperature respectively. a 、n b With n c If there is a key part B that meets the corresponding conditions, B and A are divided into one area. If there is no B that meets the corresponding conditions, A is divided into a separate area. The corresponding conditions specifically include:
[0011] Taking A above as a reference, the range of voltage, current and temperature in the same area is determined as follows:
[0012]
[0013] If a key part is connected to A, and the variation range of voltage, current and temperature belongs to the above range, then the key part and A belong to the same area.
[0014] For the interval (x min , x max ) is calculated as follows:
[0015]
[0016] Where n means there are n values, x i Indicates that in the interval (x min , x max ) in the i-th time period, x = {a, b, c}, deviation value n a 、n b With n c Obtained through the above formula and the corresponding variation range.
[0017] After dividing the areas, an abnormal fault alarm device is installed in each area. The collected data is classified according to voltage, current and temperature types, and the source area of the data is marked before sending it to the data processing cloud.
[0018] Step 3: Establish a prediction model based on machine learning in the data processing cloud, set the correlation coefficient, and improve the accuracy of the model through the correlation coefficient; divide the collected data sets of different types into training sets and test sets, and perform parallel distributed predictions based on different regions of data sources. Use the extreme random forest algorithm to build prediction models for each data type, and use the correlation coefficient to evaluate the correlation between voltage, current, temperature and abnormal faults of energy storage power stations. The specific method is as follows:
[0019] Perform correlation mining on abnormal fault data of energy storage power stations, determine the number of abnormal faults N that occur within a certain area within a time range of T, and calculate the number of abnormal faults T before the abnormal fault occurs. a The average values of voltage, current and temperature within a certain period of time, and the abnormal fault related data set Y are obtained as follows:
[0020]
[0021] Where a, b, and c represent the time before the abnormal fault occurs. a The average value of voltage, current and temperature over time.
[0022] The dataset X that is set as input to the model is as follows:
[0023]
[0024] in Represents T a The average value of voltage, current, and temperature of the input model data set over time.
[0025] The calculation formula of the voltage correlation coefficient is as follows:
[0026]
[0027] where u a Indicates that (a1, ..., a N ), Indicates that the Y data set The average value, σ a represents (a1, ..., a N ), express The standard deviation of the current and temperature is calculated by the same method. The correlation coefficient is used to evaluate the correlation between the relevant data and the abnormal fault, which can improve the prediction accuracy of the model. The relevant weights of voltage, current and temperature are set according to the correlation coefficient. For example, if p a >p b >p c , then set the relevant weights to satisfy k a >k b >k c , and ensure k a +k b +k c =1.
[0028] Determine the number of parallel decision trees x1, the maximum number of split features x2, and the minimum number of samples for split nodes x3. In this embodiment, the maximum number of split features is set to 3, and the minimum number of samples for split nodes is set to 2. When the number of decision trees is small, the generalization ability of the model is poor. Increasing the number of decision trees can improve the generalization ability of the model. However, when the number of decision trees reaches a certain level, continuing to increase the number of decision trees will not have any beneficial effect on the model. The number of decision trees is determined to be 60 through trial and error, and a prediction model that meets the prediction accuracy requirements is obtained. The output set is used as the prediction set. According to the real-time related data set, it is determined whether the mapping relationship is accurate, so that the output is given based on the given input sample set, and parallel prediction is performed to obtain the prediction value of each data type in each region.
[0029] Step 4: Set the difference threshold and comprehensive difference threshold of relevant data, calculate the difference between the predicted value and the actual value, and make abnormal fault alarm through the difference threshold judgment; select area Q, based on the historical voltage, current and temperature data of area Q, take the time point of each abnormal fault state in history as the time node, calculate the average value of voltage, current and temperature in the time period v before and after the time node, and record are the average values of voltage, current and temperature in the time period v after the time node, are the average values of voltage, current, and temperature in the v time period after the time node, and the corresponding voltage, current, and temperature differences are Δa, Δb, and Δc. According to this method, the voltage, current, and temperature differences under M abnormal fault conditions are collected, and the voltage, current, and temperature difference ranges under abnormal fault conditions are obtained as follows:
[0030]
[0031] Each difference interval contains M values. Each difference interval is analyzed and calculated, and the node value of each interval is selected, and the node value is set as the threshold of the corresponding interval. The node value of each interval is selected according to the probability of the true difference in the difference interval. If there is a difference f in a certain interval, and more than 80% of the M values contained in the interval are greater than or equal to f, f is called the node value of the interval. The node values corresponding to the voltage, current, and temperature difference intervals are f, respectively. a 、f b With f c .
[0032] The predicted value of time period v is obtained through the prediction model, and the real average value of voltage, current and temperature in time period v before the prediction is calculated. The predicted difference between the real average value of voltage, current and temperature in time period v and the predicted average value of voltage, current and temperature in time period v is obtained as follows: when Greater than or equal to f a When the voltage value is abnormal, a voltage alarm message is issued. Greater than or equal to f b When the current value is judged to be abnormal, a current alarm message is issued. When Δ~c is greater than or equal to f c When the temperature value is judged to be abnormal, a temperature alarm message is issued, and the abnormal fault alarm device in the area executes the corresponding alarm command.
[0033] Set the comprehensive alarm value W. The specific calculation formula is as follows:
[0034]
[0035] The voltage, current and temperature difference intervals of the comprehensive abnormal fault state are smaller than the part of each node value, and the comprehensive alarm threshold w is set. When the comprehensive alarm value W is greater than or equal to w, it is judged to be abnormal, and the abnormal information is sent, and the normal fault alarm device in the area executes the corresponding alarm command.
[0036] An abnormal fault alarm device for an energy storage power station includes the following contents:
[0037] Start switch, used to start and shut down the abnormal fault alarm device;
[0038] The voltage indicator light is used to judge whether the voltage in the installation area of the abnormal fault alarm device is normal. When the abnormal fault alarm device is started, the voltage indicator light is on when the voltage is normal. When the abnormal fault alarm device receives the voltage alarm information, the voltage indicator light flashes.
[0039] The current indicator light is used to judge whether the current in the installation area of the abnormal fault alarm device is normal. When the abnormal fault alarm device is started, the voltage indicator light is on when the current is normal. When the abnormal fault alarm device receives the current alarm information, the current indicator light flashes.
[0040] The temperature indicator light is used to judge whether the temperature of the installation area of the abnormal fault alarm device is normal. When the abnormal fault alarm device is started, the temperature indicator light is on when the temperature is normal. When the abnormal fault alarm device receives the temperature alarm information, the temperature indicator light flashes;
[0041] The abnormal indicator light is used to judge whether the abnormal fault alarm device installation area has received abnormal information and to start the abnormal fault alarm device. Under normal conditions, the abnormal indicator light is off. When the abnormal fault alarm device receives abnormal alarm information, the abnormal indicator light is on.
[0042] The voltage indicator is used to monitor the real-time voltage in the installation area of the abnormal fault alarm device under normal conditions. When receiving the alarm information, the voltage indicator does not work;
[0043] The current indicator is used to monitor the real-time current in the installation area of the abnormal fault alarm device under normal conditions. When receiving the alarm information, the current indicator does not work;
[0044] The temperature indicator is used to monitor the real-time temperature of the installation area of the abnormal fault alarm device under normal conditions. When receiving the alarm information, the temperature indicator does not work;
[0045] The operation indicator light is used to judge whether the abnormal fault alarm device is in operation or in shutdown state. When the abnormal fault alarm device is started, the operation indicator light is on, and when the abnormal fault alarm device is shut down, the operation indicator light is off.
[0046] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:
[0047] 1. The present invention makes fault alarm more timely and accurate through regional division, can quickly narrow the scope of fault investigation, and improve the accuracy of fault location. In an energy storage power station, multiple areas may operate at the same time. Through regional division, parallel processing of faults can be achieved, that is, faults in different areas can be repaired at the same time, thereby improving the overall fault processing efficiency;
[0048] 2. The present invention obtains the difference interval of abnormal fault status through data analysis, obtains the predicted value according to the actual value of real-time data through the prediction model, and calculates whether the difference between the predicted value and the actual value belongs to the difference interval of abnormal fault status. It can realize real-time monitoring and fault early warning of the operation status of the energy storage power station. When the system detects abnormal prediction data, it can immediately trigger a fault alarm and notify relevant personnel to handle it, which provides sufficient time for subsequent fault handling work;
[0049] 3. The present invention adopts the extreme random forest algorithm to construct prediction models for each data type respectively, and adopts the correlation coefficient to evaluate the correlation between voltage, current, temperature and abnormal failure of energy storage power stations, so as to obtain more stable and accurate prediction results, which greatly improves the detection effect of the algorithm model, and can predict the signs of failure of energy storage power stations in advance, reduce the failure rate and operation and maintenance costs, and protect the safety of life and property. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0051] Figure 1 A flow chart of an abnormal fault alarm method for an energy storage power station is shown;
[0052] Figure 2 A module diagram of an abnormal fault alarm device for an energy storage power station is shown;
[0053] Figure 3 A structural diagram of abnormal fault alarm equipment for an energy storage power station is shown.
[0054] The above drawings include the following reference numerals:
[0055] 201. Start switch, 202. Abnormal indicator light, 201. Start switch, 202. Abnormal indicator light, 203. Voltage indicator light, 204. Current indicator light, 205. Temperature indicator light, 206. Voltage indicator meter, 207. Current indicator meter, 208. Temperature indicator meter, 209. Operation indicator light. DETAILED DESCRIPTION
[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0057] Example 1
[0058] like Figure 1 As shown, a method for abnormal fault alarm of an energy storage power station in this embodiment includes the following contents:
[0059] Step 1: Collect and process relevant data of key parts of the energy storage power station through relevant data collection modules; collect relevant data of key parts of the energy storage power station, clarify the monitoring requirements of the energy storage power station, determine the key parts to be monitored and the data to be monitored. The key parts specifically refer to: energy storage battery packs, radiators and connectors of energy storage equipment; relevant data specifically refer to: historical and real-time voltage, current and temperature data of battery packs and energy storage equipment; clean and normalize the collected data to convert data of different units such as voltage, current and temperature to the same scale.
[0060] Step 2: Divide the energy storage power station area through the regional division module, classify the processed data and mark the regional source, and transmit it to the data processing cloud remotely; determine the installation location, quantity and data transmission method of the sensor. The data transmission method adopts mobile communication data transmission, WIFI or optical fiber transmission. The area is divided according to the data status of key parts under the normal operation of the energy storage power station. The normal operation state refers to the operation state without any abnormal faults. The specific method of dividing the area is as follows:
[0061] According to the collected data on the key parts of the energy storage power station, the time range t is evenly divided into n time periods to obtain the voltage, current and temperature values of each time period. The variation range of the voltage, current and temperature values of the key part A under the normal operation of the energy storage power station is obtained as follows:
[0062]
[0063] Among them, a min with a max They represent the minimum and maximum voltages of the key part A within time t, and b min With b max They represent the minimum and maximum current of the key part A within time t, c min With c max They represent the minimum and maximum temperature of the key part within time t, respectively, and set the deviation value n of voltage, current and temperature respectively. a 、n b With n c If there is a key part B that meets the corresponding conditions, B and A are divided into one area. If there is no B that meets the corresponding conditions, A is divided into a separate area. The corresponding conditions specifically include:
[0064] Taking A above as a reference, the range of voltage, current and temperature in the same area is determined as follows:
[0065]
[0066] If a key part is connected to A, and the variation range of voltage, current and temperature belongs to the above range, then the key part and A belong to the same area.
[0067] For the interval (x min , x max ) is calculated as follows:
[0068]
[0069] Where n means there are n values, x i Indicates that in the interval (x min , x max ) in the i-th time period, x = {a, b, c}, deviation value n a 、n b With n c Obtained through the above formula and the corresponding variation range.
[0070] After dividing the areas, an abnormal fault alarm device is installed in each area. The collected data is classified according to voltage, current and temperature types, and the source area of the data is marked before sending it to the data processing cloud.
[0071] Step 3: Establish a prediction model based on machine learning in the data processing cloud, set the correlation coefficient, and improve the accuracy of the model through the correlation coefficient; divide the collected data sets of different types into training sets and test sets, and perform parallel distributed predictions based on different regions of data sources. Use the extreme random forest algorithm to build prediction models for each data type, and use the correlation coefficient to evaluate the correlation between voltage, current, temperature and abnormal faults of energy storage power stations. The specific method is as follows:
[0072] Perform correlation mining on abnormal fault data of energy storage power stations, determine the number of abnormal faults N that occur within a certain area within a time range of T, and calculate the number of abnormal faults T before the abnormal fault occurs. a The average values of voltage, current and temperature within a certain period of time, and the abnormal fault related data set Y are obtained as follows:
[0073]
[0074] Where a, b, and c represent the time before the abnormal fault occurs. a The average value of voltage, current and temperature over time.
[0075] The dataset X that is set as input to the model is as follows:
[0076]
[0077] in Represents T aThe average value of voltage, current, and temperature of the input model data set over time.
[0078] The calculation formula of the voltage correlation coefficient is as follows:
[0079]
[0080] where u a Indicates that (a1, ..., a N ), Indicates that the Y data set The average value, σ a represents (a1, ..., a N ), express The standard deviation of the current and temperature is calculated by the same method. The correlation coefficient is used to evaluate the correlation between the relevant data and the abnormal fault, which can improve the prediction accuracy of the model. The relevant weights of voltage, current and temperature are set according to the correlation coefficient. For example, if p a >p b >p c , then set the relevant weights to satisfy k a >k b >k c , and ensure k a +k b +k c =1.
[0081] Determine the number of parallel decision trees x1, the maximum number of split features x2, and the minimum number of samples for split nodes x3. In this embodiment, the maximum number of split features is set to 3, and the minimum number of samples for split nodes is set to 2. When the number of decision trees is small, the generalization ability of the model is poor. Increasing the number of decision trees can improve the generalization ability of the model. However, when the number of decision trees reaches a certain level, continuing to increase the number of decision trees will not have any beneficial effect on the model. The number of decision trees is determined to be 60 through trial and error, and a prediction model that meets the prediction accuracy requirements is obtained. The output set is used as the prediction set. According to the real-time related data set, it is determined whether the mapping relationship is accurate, so that the output is given based on the given input sample set, and parallel prediction is performed to obtain the prediction value of each data type in each region.
[0082] Step 4: Set the difference threshold and comprehensive difference threshold of relevant data, calculate the difference between the predicted value and the actual value, and make abnormal fault alarm through the difference threshold judgment; select area Q, based on the historical voltage, current and temperature data of area Q, take the time point of each abnormal fault state in history as the time node, calculate the average value of voltage, current and temperature in the time period v before and after the time node, and record are the average values of voltage, current and temperature in the time period v after the time node, are the average values of voltage, current, and temperature in the v time period after the time node, and the corresponding voltage, current, and temperature differences are Δa, Δb, and Δc. According to this method, the voltage, current, and temperature differences under M abnormal fault conditions are collected, and the voltage, current, and temperature difference ranges under abnormal fault conditions are obtained as follows:
[0083]
[0084] Each difference interval contains M values. Each difference interval is analyzed and calculated, and the node value of each interval is selected, and the node value is set as the threshold of the corresponding interval. The node value of each interval is selected according to the probability of the true difference in the difference interval. If there is a difference f in a certain interval, and more than 80% of the M values contained in the interval are greater than or equal to f, f is called the node value of the interval. The node values corresponding to the voltage, current, and temperature difference intervals are f, respectively. a 、f b With f c .
[0085] The predicted value of time period v is obtained through the prediction model, and the real average value of voltage, current and temperature in time period v before the prediction is calculated. The predicted difference between the real average value of voltage, current and temperature in time period v and the predicted average value of voltage, current and temperature in time period v is obtained as follows: when Greater than or equal to f a When the voltage value is abnormal, a voltage alarm message is issued. Greater than or equal to f b When the current value is abnormal, a current alarm message is issued. Greater than or equal to f c When the temperature value is judged to be abnormal, a temperature alarm message is issued, and the abnormal fault alarm device in the area executes the corresponding alarm command.
[0086] Set the comprehensive alarm value W. The specific calculation formula is as follows:
[0087]
[0088] The voltage, current and temperature difference intervals of the comprehensive abnormal fault state are smaller than the part of each node value, and the comprehensive alarm threshold w is set. When the comprehensive alarm value W is greater than or equal to w, it is judged to be abnormal, and the abnormal information is sent, and the normal fault alarm device in the area executes the corresponding alarm command.
[0089] Embodiment 2
[0090] like Figure 2 As shown, an abnormal fault alarm device for an energy storage power station in this embodiment includes the following modules:
[0091] Related data collection module: collects relevant data of the energy storage battery pack, radiator and connector of the energy storage equipment of the energy storage power station, and performs data cleaning and normalization on the collected data;
[0092] Area division module: Based on the collected data on key parts of the energy storage power station, the voltage, current and temperature values and the range of changes in key parts in each time period are obtained. For key parts that are interconnected and have similar ranges of voltage, current and temperature, they are judged to belong to the same area.
[0093] Prediction model training module: Divide the collected data sets of different types into training sets and test sets, and perform parallel distributed predictions based on the different regions of the data sources. Use the extreme random forest algorithm to build prediction models for each data type. Use the correlation coefficient to evaluate the correlation between voltage, current, temperature and abnormal faults of energy storage power stations and improve the accuracy of the model.
[0094] Abnormal fault alarm module: Based on the historical voltage, current and temperature data of each area, the time point of each abnormal fault state in history is taken as the time node, and the average and difference of voltage, current and temperature in the time period before and after the time node are calculated. The voltage, current and temperature difference under multiple abnormal fault states are collected to obtain the voltage, current and temperature difference range of the abnormal fault state. The predicted value is obtained through the prediction model, the predicted average value is calculated, and the predicted difference of voltage, current and temperature is obtained. By comparing the corresponding thresholds, it is judged whether the voltage, current and temperature in the area are abnormal, and the abnormal fault alarm device in the area executes the corresponding alarm command.
[0095] like Figure 3 As shown, an abnormal fault alarm device for an energy storage power station includes the following structure:
[0096] A start switch 201 is used to start and stop the abnormal fault alarm device;
[0097] The abnormal indicator light 202 is used to determine whether the abnormal fault alarm device installation area has received abnormal information and start the abnormal fault alarm device. Under normal conditions, the abnormal indicator light is off. When the abnormal fault alarm device receives abnormal alarm information, the abnormal indicator light is on.
[0098] The voltage indicator light 203 is used to determine whether the voltage in the installation area of the abnormal fault alarm device is normal, and to start the abnormal fault alarm device. When the voltage is normal, the voltage indicator light is on, and when the abnormal fault alarm device receives voltage alarm information, the voltage indicator light flashes;
[0099] The current indicator light 204 is used to determine whether the current in the abnormal fault alarm device installation area is normal, and to start the abnormal fault alarm device. When the current is normal, the voltage indicator light is on, and when the abnormal fault alarm device receives the current alarm information, the current indicator light flashes;
[0100] The temperature indicator light 205 is used to determine whether the temperature of the abnormal fault alarm device installation area is normal, and to start the abnormal fault alarm device. When the temperature is normal, the temperature indicator light is on, and when the abnormal fault alarm device receives the temperature alarm information, the temperature indicator light flashes;
[0101] The voltage indicator 206 is used to monitor the real-time voltage in the installation area of the abnormal fault alarm device under normal conditions. When receiving the alarm information, the voltage indicator does not work;
[0102] The current indicator 207 is used to monitor the real-time current in the installation area of the abnormal fault alarm device under normal conditions. When receiving the alarm information, the current indicator does not work;
[0103] The temperature indicator 208 is used to monitor the real-time temperature of the installation area of the abnormal fault alarm device under normal conditions. When receiving the alarm information, the temperature indicator does not work;
[0104] The operation indicator light 209 is used to judge whether the abnormal fault alarm device is in operation or in shutdown state. When the abnormal fault alarm device is started, the operation indicator light is on, and when the abnormal fault alarm device is shut down, the operation indicator light is off.
[0105] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
[0106] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the contents of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. Those of ordinary skill in the art should understand that those skilled in the art can modify or replace the specific implementation methods of the present invention with equivalents, but these modifications or changes are within the scope of protection of the claims to be approved.
Claims
1. A method for abnormal fault alarm of energy storage power station, characterized in that: include: Step 1: Collect and process relevant data of key parts of the energy storage power station through relevant data collection modules; Step 2: Divide the energy storage power station area through the regional division module, classify the processed data and mark the regional source, and transmit it remotely to the data processing cloud; Step 3: Establish a prediction model based on machine learning in the data processing cloud, set the correlation coefficient, and improve the accuracy of the model through the correlation coefficient; Step 4: Set the difference threshold and comprehensive difference threshold of relevant data, calculate the difference between the predicted value and the actual value, and make abnormal fault alarm based on the difference threshold judgment.
2. The abnormal fault alarm method of an energy storage power station according to claim 1, characterized in that: The key parts in step 1 specifically refer to: energy storage battery pack, radiator and connector of energy storage equipment; The relevant data specifically refers to: historical and real-time voltage, current, and temperature data of battery packs and energy storage devices; The collected relevant data are cleaned and normalized.
3. The abnormal fault alarm method of an energy storage power station according to claim 1, characterized in that: Step 2: The specific method of dividing the energy storage power station area through the area division module is as follows: According to the collected data on the key parts of the energy storage power station, the time range t is evenly divided into n time periods to obtain the voltage, current and temperature values of each time period. The variation range of the voltage, current and temperature values of the key part A under the normal operation of the energy storage power station is obtained as follows: Among them, a min with a max They represent the minimum and maximum voltages of the key part A within time t, and b min With b max They represent the minimum and maximum current of the key part A within time t, c min With c max They represent the minimum and maximum temperature of the key part within time t, respectively, and set the deviation value n of voltage, current and temperature respectively. a 、n b With n c , taking A as a reference, the range of voltage, current and temperature in the same area is determined as follows: The variation range of the set voltage, current and temperature belongs to the above range, and the key part connected with A belongs to the same area as A.
4. The abnormal fault alarm method of an energy storage power station according to claim 3, characterized in that: For the interval (x min , x max ) is calculated as follows: Where n means there are n values, x i Indicates that in the interval (x min , x max ) in the i-th time period, x = {a, b, c}, deviation value n a 、n b With n c Obtained through the above formula and the corresponding variation range.
5. The abnormal fault alarm method of an energy storage power station according to claim 1, characterized in that: The specific method for calculating the correlation coefficient in step 3 is as follows: Perform correlation mining on abnormal fault data of energy storage power stations, determine the number of abnormal faults N that occur within a certain area within a time range of T, and calculate the number of abnormal faults T before the abnormal fault occurs. a The average values of voltage, current and temperature within a certain period of time, and the abnormal fault related data set Y are obtained as follows: Where a, b, and c represent the time before the abnormal fault occurs. a Average values of voltage, current and temperature over time; The dataset X that is set as input to the model is as follows: in Represents T a The average values of voltage, current and temperature of the input model data set within the time period; The calculation formula of the voltage correlation coefficient is as follows: where u a Indicates that (a1, ..., a N ), Indicates that the Y data set The average value, σ a represents (a1, ..., a N ), express The standard deviation of current and temperature are calculated using the same method.
6. The abnormal fault alarm method of an energy storage power station according to claim 1, characterized in that: The specific contents of step 4 are as follows: Select area Q, based on the historical voltage, current and temperature data of area Q, take the time point of each abnormal fault state in history as the time node, calculate the average value of voltage, current and temperature in the time period v before and after the time node and find the difference, the corresponding voltage, current and temperature difference is Δa, Δb, Δc, collect the voltage, current and temperature difference under M abnormal fault states, and get the difference interval of abnormal fault state as follows: Analyze and calculate each difference interval, select the node value of each interval, each interval has M values, there is a difference f in a certain interval, when there is a difference exceeding R% in the interval, its value is greater than or equal to f, f is called the node value of the interval, the node values corresponding to the voltage, current, and temperature difference intervals are f respectively a 、f b With f c , set the node value to the threshold of the corresponding interval; The prediction model is used to obtain the predicted value of time period v, and the true average values of voltage, current, and temperature in time period v before and after the prediction are calculated to obtain the predicted difference of voltage, current, and temperature. when Greater than or equal to f a When the voltage value is abnormal, a voltage alarm message is issued. Greater than or equal to f b When the current value is abnormal, a current alarm message is issued. Greater than or equal to f c When the temperature value is judged to be abnormal, a temperature alarm message is issued, and the abnormal fault alarm device in the area executes the corresponding alarm command; Set the comprehensive alarm value W. The specific calculation formula is as follows: The voltage, current and temperature difference intervals of the comprehensive abnormal fault state are smaller than the part of each node value, and the comprehensive alarm threshold w is set. When the comprehensive alarm value W is greater than or equal to w, it is judged to be abnormal, and the abnormal information is sent, and the normal fault alarm device in the area executes the corresponding alarm command.
7. An abnormal fault alarm device for an energy storage power station, applied to an abnormal fault alarm method for an energy storage power station according to any one of claims 1 to 6, characterized in that: include: Related data collection module: collects relevant data of the energy storage battery pack, radiator and connector of the energy storage equipment of the energy storage power station, and performs data cleaning and normalization on the collected data; Area division module: Based on the collected data on key parts of the energy storage power station, the voltage, current and temperature values and the range of changes in key parts in each time period are obtained. For key parts that are interconnected and have similar ranges of voltage, current and temperature, they are judged to belong to the same area. Prediction model training module: Divide the collected data sets of different types into training sets and test sets, and perform parallel distributed predictions based on the different regions of the data sources. Use the extreme random forest algorithm to build prediction models for each data type. Use the correlation coefficient to evaluate the correlation between voltage, current, temperature and abnormal faults of energy storage power stations and improve the accuracy of the model. Abnormal fault alarm module: Based on the historical voltage, current and temperature data of each area, the time point of each abnormal fault state in history is taken as the time node, and the average and difference of voltage, current and temperature in the time period before and after the time node are calculated. The voltage, current and temperature difference under multiple abnormal fault states are collected to obtain the voltage, current and temperature difference range of the abnormal fault state. The predicted value is obtained through the prediction model, the predicted average value is calculated, and the predicted difference of voltage, current and temperature is obtained. By comparing the corresponding thresholds, it is judged whether the voltage, current and temperature in the area are abnormal, and the abnormal fault alarm device in the area executes the corresponding alarm command.
Citation Information
Cited By
Mobile substation power interface plugging abnormity detection circuit and method
CN119535058A
Real-time control system for automatic loading and stacking of bulk materials
CN119599546A
