Water feed pump fault early warning diagnosis processing method and device

Abstract

The invention provides a water feed pump fault early warning and diagnosis processing method and device. The technical problem that effective fault early warning and diagnosis cannot be formed through diversified field monitoring data in the existing monitoring process is solved. The method comprises the following steps: forming a potential fault list and a fault function module list associated with each potential fault according to prior data; extracting presentation monitoring data related to potential fault performance from the field monitoring data, and binding the presentation monitoring data with the potential faults; extracting entity monitoring data related to the entity of the associated function module from the field monitoring data, and binding the entity monitoring data with the potential fault; and forming a fault diagnosis process according to the presentation monitoring data and the entity monitoring data, and forming a fault removal process according to a fault diagnosis result. Faults are divided into potential faults reflecting overall abnormity and entity faults needing to be clearly positioned, and then effective clustering of mass field monitoring data to fault types and optimization of a fault diagnosis process are formed. And the accuracy of early warning and evaluation of the fault is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a method and device for early warning diagnosis and processing of feed pump failures. Background Art

[0002] The feed pump group is an important auxiliary machine in the steam-water circulation system of a thermal power plant. By adjusting the rotational speed of the feed pump, the feed water flow rate and pressure supplied to the boiler are changed, and desuperheated water or emergency spray water is provided to the reheater and superheater to suit the start-stop, load change, variable pressure, and sliding pressure operation of the unit, meet the needs of the power grid for unit regulation and frequency modulation, and improve the economy, safety, and reliability of the unit under variable operating conditions.

[0003] In the prior art, the main components of a feed pump usually include a balance disk, a pump shaft, a shaft seal, bearings for supporting the pump shaft, a lubricating oil system supporting the bearings, a cold oil cooler, and a coupling for torque transmission between the pump shaft and an electric motor or a steam turbine, etc. For the operating conditions of each functional module during the operation of the feed pump group, the state acquisition of the targeted signals of the physical components in the module can be realized through mature sensor technology, combined with wired or wireless communication technology to form feedback on-site monitoring data, and computer technology is used to form data comparison and alarms.

[0004] However, the existing monitoring technology solutions can only make direct judgments based on the set thresholds of each data for the massive on-site monitoring data to form alarm signals, unable to identify the underlying logical meaning of the alarm status, unable to form a judgment on the overall operating condition stability and failure rate of the feed pump group, and unable to effectively diagnose and evaluate potential faults during the operation process. Summary of the Invention

[0005] In view of the above problems, embodiments of the present invention provide a method and device for early warning diagnosis and processing of feed pump failures, which solve the technical problem that the existing monitoring process cannot form effective fault early warning and diagnosis by using diverse on-site monitoring data.

[0006] The method for early warning diagnosis and processing of feed pump failures in the embodiments of the present invention includes:

[0007] Form a list of potential faults and a list of fault functional modules associated with each potential fault according to prior data;

[0008] Extract the appearance monitoring data related to the manifestation of potential faults from the on-site monitoring data and bind it to the potential faults;

[0009] Extract the entity monitoring data related to the entity of the associated functional module from the on-site monitoring data and bind it to the potential faults;

[0010] Form a fault diagnosis process according to the appearance monitoring data and the entity monitoring data, and form a fault elimination process according to the fault diagnosis result.

[0011] In one embodiment of the present invention, the prior data is statistically quantified from the past failure data of the feed water pump equipment or the failure data of equipment with the same configuration under the same usage environment.

[0012] In one embodiment of the present invention, the normal operating conditions of the feed water pump include 0.5 times frequency, 1 times frequency, and 2 times frequency of the rated speed.

[0013] In one embodiment of the present invention, the potential failure list includes abnormal oil temperature alarm in the lubricating oil tank, low oil level alarm in the lubricating oil tank, low lubricating oil pressure alarm, high bearing temperature alarm, large bearing vibration alarm of the feed water pump group, excessive shaft seal leakage, and wear of the balance disk device.

[0014] In one embodiment of the present invention, in the list of failure function modules associated with each potential failure, the mapping relationship between the potential failure and the failure entity includes:

[0015] Abnormal oil temperature in the lubricating oil tank:

[0016] Lubricating oil system [electric heater, oil cooler, lubricating oil pipeline, lubricating oil];

[0017] Cooling water system [cold oil pump, closed cooling water pipeline];

[0018] Low oil level alarm in the lubricating oil tank:

[0019] Lubricating oil system [lubricating oil pump, lubricating oil flow];

[0020] Low lubricating oil pressure alarm:

[0021] Lubricating oil system [lubricating oil pump, lubricating oil flow];

[0022] High bearing temperature alarm:

[0023] Lubricating oil system [lubricating oil pump, lubricating oil flow, lubricating oil quality, temperature control door of the lubricating oil station];

[0024] Cooling water system [closed cooling water supply, closed cooling water pipeline];

[0025] Bearing [bearing bush];

[0026] Large bearing vibration alarm of the feed water pump group:

[0027] Dynamic balance [rotor, rotor blade, local high temperature of the main shaft, balance weight];

[0028] Bearing seat [foundation soleplate, bearing seat];

[0029] Coupling [concentricity, coupling opening, pin hole, rubber ring];

[0030] Static-dynamic friction [main shaft, steam-driven feed water pump];

[0031] Specific state [hydraulic system];

[0032] Excessive shaft seal leakage:

[0033] Shaft seal [shaft seal];

[0034] Wear of balance disk device:

[0035] Balance disk [balance disk].

[0036] In one embodiment of the present invention, the process of forming a fault diagnosis based on surface monitoring data and entity monitoring data, and forming a fault elimination process according to the fault diagnosis result includes:

[0037] Predict potential fault types through real-time surface monitoring data;

[0038] Determine the fault entity in the fault function module that causes the potential fault through real-time entity monitoring data;

[0039] Determine the fault elimination process according to the fault entity.

[0040] In one embodiment of the present invention, the predicting potential fault types through real-time surface monitoring data includes:

[0041] Determine the abnormal values of the surface monitoring data;

[0042] Predict the occurrence of potential faults according to the frequency or duration of the abnormal values;

[0043] Determine the potential fault type according to the abnormal value type and form the corresponding warning information.

[0044] The feed pump fault early warning diagnosis processing device according to an embodiment of the present invention includes:

[0045] A memory for storing the program code in the process of the feed pump fault early warning diagnosis processing method as described above;

[0046] A processor for executing the program code.

[0047] The feed pump fault early warning diagnosis processing device according to an embodiment of the present invention includes:

[0048] A fault factor classification module for forming a list of potential faults and a list of fault function modules associated with each potential fault according to prior data;

[0049] A prediction data binding module for extracting surface monitoring data related to the manifestation of potential faults from on-site monitoring data and binding it to the potential faults;

[0050] A diagnostic data binding module, configured to extract entity monitoring data related to the entities of the associated functional modules from on-site monitoring data and bind it with potential faults;

[0051] A fault prediction and diagnosis module, configured to form a fault diagnosis process based on appearance monitoring data and entity monitoring data, and form a fault troubleshooting procedure according to the fault diagnosis result.

[0052] In one embodiment of the present invention, the fault prediction and diagnosis module includes:

[0053] A fault prediction module, configured to predict potential fault types through real-time appearance monitoring data;

[0054] A fault diagnosis module, configured to determine the faulty entity in the faulty functional module that causes the potential fault through real-time entity monitoring data;

[0055] A fault troubleshooting module, configured to determine a fault troubleshooting procedure according to the faulty entity.

[0056] The feed pump fault early warning and diagnosis processing method and device according to the embodiments of the present invention effectively integrate discrete massive on-site monitoring data and isolated data judgment and diagnosis processes. Faults are divided into potential faults reflecting overall anomalies and entity faults that require precise positioning, thereby forming an effective clustering of massive on-site monitoring data into fault types and an optimization of the fault diagnosis process. It is possible to fully utilize and update the existing discrete data judgment and diagnosis process to construct a new fault early warning and diagnosis processing process, improving the accuracy of fault early warning and assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 The figure shows a schematic flow chart of the feed pump fault early warning and diagnosis processing method according to an embodiment of the present invention.

[0058] Figure 2 The figure shows a schematic fault diagnosis flow formed in the practical application of the feed pump fault early warning and diagnosis processing method according to an embodiment of the present invention Figure 1 .

[0059] Figure 3 The figure shows a schematic fault diagnosis flow formed in the practical application of the feed pump fault early warning and diagnosis processing method according to an embodiment of the present invention Figure 2 .

[0060] Figure 4 The figure shows a schematic architecture diagram of the feed pump fault early warning and diagnosis processing device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0062] A method for early warning diagnosis and treatment of feed pump faults in an embodiment of the present invention is as Figure 1 shown. In Figure 1 this, this embodiment includes:

[0063] Step 100: Form a list of potential faults and a list of fault function modules associated with each potential fault according to prior data.

[0064] Those skilled in the art can understand that the fault manifestations and types of the feed pump can be statistically quantified and analyzed from the past fault data of the feed pump equipment or the fault data of equipment with the same configuration in the same usage environment (i.e., using prior data). The on-site monitoring data is also formed by collecting physical signals through the layout of various types of sensors for the working conditions of each functional module's hardware according to the feedback requirements under the actual working conditions of the feed pump. The layout of the same type of sensor can be for different positions and different hardware entities, and can feedback the signal status of this type locally or as a whole. The feed pump as a whole includes various operating states, and the abnormal states in the overall operating state performance can often be quantified as the existence of relevant potential faults. The basic distinction of each potential fault is formed according to the differences in the abnormal performance of the overall operating state. For example, the potential fault list is determined through the preliminary abnormal manifestations such as local overheating, liquid level reduction, pressure abnormality, vibration, and liquid leakage shown by the overall signal feedback of the feed pump. According to the statistical quantification analysis of prior data, the main and clear functional modules that cause the preliminary abnormal manifestations in the feed pump can be determined, and a list of fault function modules corresponding to the potential faults can be formed. The functional modules include but are not limited to the equipment component hardware, functional subsystems, assembly and fixing structures, and the coupling state of the working conditions between functional hardwares. The on-site monitoring data of the functional modules is usually collected according to the sensors arranged on-site and formed by fitting or converting the relevant collected data.

[0065] Step 200: Extract the appearance monitoring data related to the potential fault manifestations from the on-site monitoring data and bind it to the potential faults.

[0066] The potential faults have relatively direct appearance monitoring data reflected in the overall working conditions of the feed pump. Binding these types of appearance monitoring data as the data source to the potential fault types can form a clear basis for pre-judging data for potential faults. The occurrence probability of potential faults can be predicted according to the estimated value and trend value of the pre-judging data.

[0067] Step 300: Extract the entity monitoring data related to the entities of the associated functional module from the on-site monitoring data and bind it to potential faults.

[0068] Those skilled in the art can understand that there is a large amount of monitoring data of module components in the on-site monitoring data of the functional module, which is used to represent different attributes or parameters in the working condition state of the functional module. Different types of entity monitoring data of the same functional module will form correlations with different potential faults. It can be used as the entity monitoring data of the most relevant part of the functional module with fault correlations existing in the determined potential faults.

[0069] Step 400: Form a fault diagnosis process based on the appearance monitoring data and the entity monitoring data, and form a fault elimination process according to the fault diagnosis result.

[0070] Those skilled in the art can understand that for different fault types and fault location diagnoses, evaluation logics can be preset, and prediction or diagnosis evaluation conclusions are formed through threshold comparison, data logic comparison, and type data fitting. And according to the diagnosis result, targeted elimination means are determined to form the construction of the fault elimination process. The appearance monitoring data is used to form the qualitative prediction basis of the potential fault diagnosis process, and the entity monitoring data is used to form the quantitative diagnosis basis of the potential fault diagnosis process.

[0071] The feed pump fault early warning diagnosis and processing method of the embodiments of the present invention effectively integrates discrete massive on-site monitoring data and isolated data judgment and diagnosis processes. The faults are divided into potential faults reflecting overall abnormalities and entity faults that need to be clearly located, thereby forming an effective clustering of massive on-site monitoring data into fault types and an optimization of the fault diagnosis process. It can make full use of the existing discrete data judgment and diagnosis process that can be updated to construct a new fault early warning diagnosis and processing process, improving the accuracy of fault early warning and evaluation.

[0072] A fault diagnosis process formed by using the feed pump fault early warning diagnosis and processing method of the embodiments of the present invention is as Figure 2 and Figure 3 shown. In Figure 2 and Figure 3 , the on-site monitoring data is bound to the working conditions of the feed pump at different levels to form a fault early warning diagnosis and processing process that can be mapped by using the data storage structure of the relational database. In Figure 2 and Figure 3Among them, the normal operating conditions of the feed water pump include three rated speeds of 0.5 times frequency, 1 times frequency, and 2 times frequency. The formed list of potential faults includes abnormal oil temperature alarm in the lubricating oil tank, low oil level alarm in the lubricating oil tank, low lubricating oil pressure alarm, high bearing temperature alarm, large bearing vibration alarm of the feed water pump set, excessive shaft seal leakage, wear of the balance disc device, etc. Each potential fault corresponds to a list of fault function modules that may form potential faults. The corresponding fault function modules map the possible fault entities (lists). At the same time, according to the types of potential faults and function modules, the fault location logic for using on-site monitoring data at different levels such as the overall level, module level, and entity level is given. At the same time, the corresponding troubleshooting procedures selected for fault location are given.

[0073] For the fault entities corresponding to different potential faults, they are often only partial component entities of the function modules. The list of fault entities has a strong correlation with the type of potential faults. By effectively diagnosing and processing the on-site monitoring data of the fault entities through preset logic, accurate fault judgments can be formed, providing effective support for fault prediction. In Figure 2 and Figure 3 In one embodiment of the present invention, the mapping relationship between potential faults and fault entities includes:

[0074] Abnormal oil temperature in the lubricating oil tank:

[0075] Lubricating oil system [electric heater, oil cooler, lubricating oil pipeline, lubricating oil];

[0076] Cooling water system [cold oil pump, closed cooling water pipeline];

[0077] Low oil level alarm in the lubricating oil tank:

[0078] Lubricating oil system [lubricating oil pump, lubricating oil flow];

[0079] Low lubricating oil pressure alarm:

[0080] Lubricating oil system [lubricating oil pump, lubricating oil flow];

[0081] High bearing temperature alarm:

[0082] Lubricating oil system [lubricating oil pump, lubricating oil flow, lubricating oil quality, temperature control valve of lubricating oil station];

[0083] Cooling water system [closed cooling water supply, closed cooling water pipeline];

[0084] Bearing [bearing bush];

[0085] Large bearing vibration alarm of the feed water pump set:

[0086] Dynamic balance [rotor, rotor blade, local high temperature of main shaft, balance weight];

[0087] Bearing housing [foundation soleplate, bearing housing];

[0088] Coupling [concentricity, coupling opening, pin hole, rubber ring];

[0089] Dynamic and static friction [main shaft, steam-driven feed water pump];

[0090] Specific state [hydraulic system];

[0091] Excessive shaft seal leakage:

[0092] Shaft seal [shaft seal];

[0093] Wear of balance disk device:

[0094] Balance disk [balance disk].

[0095] Using the data results formed during the processing of the feed water pump fault early warning diagnosis and processing method of the above embodiments as the data basis, specific processing methods for feed water pump fault early warning diagnosis can be formed.

[0096] As Figure 1 shown, in an embodiment of the present invention, step 400 includes:

[0097] Step 410: Predict potential fault types through real-time appearance monitoring data.

[0098] Perform potential fault prediction based on the real-time working condition state of the feed water pump at the overall level feedback by the appearance monitoring data. In an embodiment of the present invention, the prediction process includes:

[0099] Determine the abnormal values of the appearance monitoring data. The determination process of the abnormal values uses the fault location logics of different levels such as the preset overall level, module level, and entity level to utilize the on-site monitoring data.

[0100] Predict the occurrence of potential faults based on the frequency or duration of the abnormal values. The prediction results are formed through the abnormal value analysis preset in the time domain or frequency domain.

[0101] Determine the potential fault types according to the abnormal value types and form corresponding warning messages. Determine the types of isolated or associated potential faults and form corresponding warning messages.

[0102] Step 420: Determine the fault entity in the fault function module that causes the potential fault through real-time entity monitoring data.

[0103] Obtain the list of fault function modules related thereto according to the determination of the potential fault types, and obtain the real-time entity monitoring data corresponding to the fault function modules. In an embodiment of the present invention, the diagnostic process formed according to the real-time entity monitoring data includes:

[0104] Determine the outliers in the entity monitoring data. The process of determining outliers uses the pre-set fault location logic that utilizes on-site monitoring data at different levels such as the overall level, module level, and entity level.

[0105] Diagnose the faulty entity and faulty functional module based on the frequency or duration of the outliers. The diagnostic result is formed through the pre-set outlier analysis in the time domain or frequency domain.

[0106] Form the alarm information according to the diagnostic result. Determine the isolated or associated faulty entities and faulty functional modules and form the corresponding alarm information.

[0107] Step 430: Determine the troubleshooting process according to the faulty entity.

[0108] Form the scheduling order of the troubleshooting process according to the relevance of the faulty entities.

[0109] The feed pump fault early warning and diagnosis processing method of the embodiment of the present invention uses the mapping of data levels in on-site monitoring data to form the predictive judgment of potential faults and the diagnostic judgment of faulty entities. Through the pre-set logic processing of the surface monitoring data, the potential fault range is focused, and through the entity monitoring data, the faulty functional module is determined, and the recommended means and process arrangements for troubleshooting are obtained. It effectively improves the efficiency of fault identification and evaluation in the face of complex fault scenarios and massive on-site monitoring data.

[0110] An embodiment of the feed pump fault early warning and diagnosis processing device of the present invention includes:

[0111] A memory for storing the program code in the processing process of the feed pump fault early warning and diagnosis processing method of the above embodiment;

[0112] A processor for executing the program code in the processing process of the feed pump fault early warning and diagnosis processing method of the above embodiment.

[0113] The processor can adopt a DSP (Digital Signal Processor) digital signal processor, an FPGA (Field-Programmable Gate Array) field programmable gate array, an MCU (Microcontroller Unit) system board, an SoC (system on a chip) system board, a PLC (Programmable Logic Controller) minimum system including I / O, or cloud computing power.

[0114] An embodiment of the feed pump fault early warning and diagnosis processing device of the present invention is as Figure 4 shown. In Figure 4 it, this embodiment includes:

[0115] A fault factor classification module 10 is configured to form a list of potential faults and a list of fault function modules associated with each potential fault according to prior data;

[0116] A prediction data binding module 20 is configured to extract appearance monitoring data related to the manifestation of potential faults from on-site monitoring data and bind it to the potential faults;

[0117] A diagnosis data binding module 30 is configured to extract entity monitoring data related to the entity of the associated function module from on-site monitoring data and bind it to the potential faults;

[0118] A fault prediction and diagnosis module 40 is configured to form a fault diagnosis process based on the appearance monitoring data and the entity monitoring data, and form a fault troubleshooting procedure according to the fault diagnosis result.

[0119] As Figure 4 shown, in an embodiment of the present invention, the fault prediction and diagnosis module 40 includes:

[0120] A fault prediction module 41 is configured to predict the type of potential faults through real-time appearance monitoring data;

[0121] A fault diagnosis module 42 is configured to determine the fault entity in the fault function module that causes the potential fault through real-time entity monitoring data;

[0122] A fault troubleshooting module 43 is configured to determine a fault troubleshooting procedure according to the fault entity.

[0123] As described above for the feed pump fault warning and diagnosis processing device in an embodiment of the present invention, it is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A water pump fault early warning diagnosis and processing method, characterized in that: include: Forming a potential fault list and a fault function module list associated with each potential fault according to prior data; Extract the surface monitoring data related to the potential fault manifestation from the on-site monitoring data and bind it to the potential fault; Extracting entity monitoring data related to entities of associated functional modules from the field monitoring data and binding them with potential faults; A fault diagnosis process is formed based on the surface monitoring data and the entity monitoring data, and a fault elimination process is formed based on the fault diagnosis results.

2. The water supply pump fault early warning diagnosis and processing method according to claim 1, characterized in that: The priori data is statistically quantified from past failure data of the water supply pump equipment or failure data of the equipment with the same configuration under the same use environment.

3. The water supply pump fault early warning diagnosis and processing method according to claim 1, characterized in that: The normal operating conditions of the water supply pump include 0.5 times, 1 times, and 2 times the rated speed.

4. The water supply pump fault early warning diagnosis and processing method according to claim 1, characterized in that: The potential fault list includes alarms for abnormal oil temperature in the lubricating oil tank, low oil level in the lubricating oil tank, low lubricating oil pressure, high bearing temperature, large bearing vibration alarm for the water pump group, excessive shaft seal leakage, and wear of the balancing disc device.

5. The water supply pump fault early warning diagnosis and processing method according to claim 1, characterized in that: In the list of fault function modules associated with each potential fault, the mapping relationship between the potential fault and the fault entity includes: Abnormal oil temperature in the lubricating oil tank: Lubricating oil system [electric heater, oil cooler, lubricating oil pipeline, lubricating oil]; Cooling water system [cold oil pump, closed cold water pipeline]; Lubricating oil tank low oil level alarm: Lubricating oil system [lubricating oil pump, lubricating oil flow]; Low lubricating oil pressure alarm: Lubricating oil system [lubricating oil pump, lubricating oil flow]; High bearing temperature alarm: Lubricating oil system [lubricating oil pump, lubricating oil flow, lubricating oil quality, lubricating oil station temperature control door]; Cooling water system [closed cold water supply, closed cold water pipeline]; bearings; Large vibration alarm of water pump group bearing: Dynamic balancing [rotor, rotor blades, local high temperature of the main shaft, balancing blocks]; Bearing seat [base plate, bearing seat]; Coupling [concentricity, coupling opening, pin hole, rubber ring]; Dynamic and static friction [spindle, steam-driven water pump]; specific state [hydraulic system]; Excessive water leakage from shaft seal: shaft seal [shaft seal]; Balance disc device wear: Balance plate [Balance plate].

6. The water supply pump fault early warning diagnosis and processing method according to claim 1, characterized in that: The fault diagnosis process is formed according to the appearance monitoring data and the entity monitoring data, and the fault elimination process is formed according to the fault diagnosis result, including: Predict potential fault types through real-time phenomenon monitoring data; Determine the faulty entity in the faulty functional module causing the potential fault through real-time entity monitoring data; Determine the troubleshooting process based on the fault entity.

7. The water supply pump fault early warning diagnosis and processing method according to claim 6, characterized in that: The prediction of potential fault types through real-time appearance monitoring data includes: Identify outliers in surface monitoring data; Predict potential failures based on the frequency or duration of abnormal values; Determine the potential fault type based on the abnormal value type and generate corresponding alarm information.

8. A water supply pump fault early warning diagnosis and processing device, characterized in that: include: A memory, used to store program codes in the process of the water supply pump fault early warning diagnosis and processing method according to any one of claims 1 to 7; A processor is used to execute the program code.

9. A water pump fault early warning diagnosis and processing device, characterized in that: include: A fault factor classification module, used for forming a potential fault list and a fault function module list associated with each potential fault according to prior data; Prediction data binding module, used to extract the appearance monitoring data related to the potential fault manifestation from the field monitoring data and bind it with the potential fault; A diagnostic data binding module, used to extract entity monitoring data related to entities of associated functional modules from field monitoring data and bind them to potential faults; The fault prediction and diagnosis module is used to form a fault diagnosis process based on the appearance monitoring data and the entity monitoring data, and to form a fault elimination process based on the fault diagnosis results.

10. The water supply pump fault early warning diagnosis and processing device according to claim 9, characterized in that: The fault prediction and diagnosis module comprises: Fault prediction module, used to predict potential fault types through real-time appearance monitoring data; A fault diagnosis module, used to determine a fault entity in a faulty functional module causing a potential fault through real-time entity monitoring data; The troubleshooting module is used to determine the troubleshooting process according to the fault entity.

Images