FIELD OF INVENTION
The invention relates to intelligent boiler tube leak detection system for leak detection inthermal power plants. Tube leaks are one of the major causes for boiler outage. Early detection of leaks enables corrective measures to prevent expensive secondary damages. Presently, power plants deploy 12-30 acoustic sensors at strategic locations to detect leaks. These sensors are usually connected to a remote signal processing unit at the control room using signal wires carrying power and analog leak noise signals. The remote processing module processes these signals to detect the presence of leaks. The present invention relates to anintelligent boiler tube leak detection system which employs intelligent sensors connected over an intelligent fail-safe digital network.
BACKGROUND OF THE INVENTION & PRIOR ARTS
High pressure leaks in power plants can cause extensive damage in boiler components and unexpected downtimes. Historically, detection methods like manual inspection, detecting increase in makeup water use and temperature monitoring were used, which can only find leaks of significant size, when considerable damage would have already been caused. Later, acoustic steam leak detection systems which place multiple acoustic sensors at various strategic

locations were used for leak detection. These systems usually consist of a sensor module with the sensing element or transducer, a field amplifier module which amplifies the signal; a cable means to transmit the signal to a remote processing module; the remote processing module which processes the independent/ multiplexed signals; a display unit which provides user interface. The present invention describes an intelligent sensor which is capable of detecting leaks by itself and has automatic calibration, automatic purging, acoustic spectral analysis, and machine learning to augment the leak detection functions.
Indian Patent 217835, titled “A Sonic Tube Leak Detection Device for Boilers To Detect Steam Leakage of Boiler Tubes”relates to a sonic tube leak detecting device (STLD) for boilers to detect steam leakage of boiler tubes comprising a sonic tube assembly, an acoustic sensor, a pre amplifier unit and a system panel, said sonic tube assembly comprises a matching unit, isolating unit and a heat insulating coupling, said acoustic sensor comprises a coupling, an electronic unit, an insulated bush, a s.s.bush and a connector, said system panel comprising a modular system electronics, a monitor, a PC, an audio system, a key board and a printer, the steam leak noise being converted into electric signal in the sensor which is amplified in the preamplifier unit, the current signal and the voltage with frequency information being processed through the processor module, the out-put signal being simultaneously fed to a V/I converter, a comparator circuit, an analog remote display read out and a computer system for storage, mimic

generation, graphical read out, historical and trend reporting. The present invention uses intelligent sensors implementing digital signal processing techniques minimises analog circuitry while providing more insight for the power plant operator.
Indian Patent Application942/Kol/2014, describes a method and apparatus for automatic configuration and calibration of acoustic sensors employing acoustic sensors and calibrator with a wireless/ RF interface. Indian Patent Application 643/Kol/2015, describes intelligent sensors for leak detection which are capable of detecting leaks based on spectral processing using digital signal processing and machine learning having additional features - automatic calibration, powerless configuration, automatic purging and self-diagnostics. The current invention interconnects these sensors in an intelligent way to achieve plant wide information.
US Patent 20120093203 discloses a data communication method, communication device, and communication system wherein a master device is connected, through the first RS485/RS422 port and the second RS485/RS422 port of the master device, to the at least two slave devices by using an RS485/RS422 bus loop, and when transmitting the data from the first RS485/RS422 port of the master device to the slave device connected to the master device through the first RS485/RS422 port of the master device, the master device activates the data transmission function on the first RS485/RS422 port of the master device while inactivates the data transmission function on the second RS485/RS422 port of the

master device; and when a line interruption is detected, inactivating the data transmission function on the first RS485/RS422 port of the master device, and activating the data transmission function on the second RS485/RS422 port of the master device. The system uses two RS485 ports on the slave end while present invention implements a fail-safe communication system using only a single port at the slave end.
OBJECTS OF THE INVENTION
It is an object of the invention to have intelligent boiler tube leak detection system which is simple to commission while reducing the amount of cabling required.A further object of the invention is to have fail safe operation.
Another object of the invention is to utilize digital signal processing algorithms and machine learning algorithms to detect leaks intelligently reducing manual intervention for operation.
Yet another object of the invention is to have interconnection of intelligent sensors which detect leak independently to provide more utility to the plant.
A further object of the invention is to simplify and reduce the amount of analog circuitry or modules in an acoustic leak detection system.

Also the object of the invention is to have improved user notifications with sms, email alerts. Yet another object of the invention is to have a future proof system capable of expanding the number of sensors and the I/O interfaces to plant DCS.
SUMMARY OF THE INVENTION
The intelligent boiler tube leak detection system consists of multiple, usually 12-30, intelligent sensors connected to a remote data concentrator module using RS485 bus. The data concentrator module has two independent RS485 ports which can be configured either in the even-odd redundancy mode or the intelligent fail-safe ring mode. The data concentrator module employs machine learning and neuro-fuzzy algorithms to study the leak patterns and provide further insight to the plant operator. The data concentrator module also studies the trends in individual sensor noise levels and initiates a purge of the sensor tube automatically. The data concentrator can poll individual sensors for diagnostics, calibration requirement, etc. All the gathered information is available to the user in the form of user-friendly SMS, email and web pages. The sensors can be calibrated without user intervention using the calibrator module. The sensors can be programmed/ updated from the data concentrator module over the RS485 bus. The data concentrator module can interface to a Human Machine Interface or a power plant DCS using Ethernet/

RS485 connectivity. Through an additional RS485 port on the data concentrator module for interfacing analog/ digital I/O modules, the system can provide any number of analog or digital inputs or outputs to the DCS. The number of sensors can also be increased at any time as additional nodes in the RS485 bus.
BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
Figure 1 illustrates a the block diagram of the data concentrator module;
Figure 2 illustrates the system in even-odd redundancy configuration;
Figure 3 illustrates the system in fail-safe ring configuration;
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT OF THE INVENTION
The block diagram of the data concentrator module is shown in Figure 1. The data concentrator module, 100, consists of a microcontroller/ processor 101interfaced to the sensors through RS485 ports 104 and 105, with a keypad display interface 103, memory 106, CODEC 107 and audio output circuitry 108, interface to GSM module through serial port109, interface to HMI/ DCS through Ethernet port 110 or RS485 port 111, interface to expandable I/O modules through RS485 port 112 and on-board digital I/O’s with respective drivers 113. The module

is powered through appropriate DC-DC power supplies 102. The two sensor RS485 ports 104 and 105 can be configured to operate in an even-odd redundancy configuration or a fail-safe ring configuration. The detailed operation of these have been explained later with figures 2 & 3. The microcontroller polls the sensors for data including frequency spectrum, current noise level, diagnostics, etc. through the RS485 ports. The data concentrator module can also play the noise stream from individual sensors with data from the RS485 port using the CODEC and the audio output circuitry. This enables the operator to observe/ troubleshoot the sensors from the control room. The memory is used to log the trends, diagnostics, etc. for further analysis at a later time. A GSM module may be interfaced to get live alerts through SMS, e-mails, etc. The system can be interfaced to HMI/ DCS through Ethernet port or RS485 port to get further visual information. The on-board digital I/O’s provide basic inputs and outputs such as leak detected, system fault, soot blower on, etc. Further analog and/or digital outputs for individual channels may be derived using RS485 I/O modules connected to the expandable I/O port. The number of channels in the system can be increased by adding more sensors as new nodes the RS485 ports 104 and/ or 105 in either even-odd or fail-safe configuration.
Figure 2, illustrates an exemplary embodiment of the complete invention in the even-odd configuration. Here an equal number of sensors are connected to both of the sensors RS485 ports so that in case of a failure, at least

on set of sensors shall detect the leak. The ports are both in RS485 master configuration and may be 4-wire full duplex with higher communication speeds or 2-wire half duplex with cost effective cabling. The sensors, data concentrator modules and other modules may be powered using redundant power supplies to get power supply redundancy. The data concentrator module, GSM module, redundant power supplies, I/O modules can be housed in a panel at the control room.
Figure 3, illustrates the preferred embodiment of the complete invention in fail-safe ring mode. On power up, one port is configured as a master and the other as slave. The RS485 network is of two-wire half duplex nature. The slave port keeps receiving the command from the master and keeps verifying the responses from the slaves. If the master transmission is not received at the slave port at the end of the loop, the ring network has been disturbed and the slave port is now configured to become a master and shall initiate polling of the sensors. The sensors at one end of the broken ring shall communicate with master port 1 while the sensors at the other end shall communicate with the master port 2. The ports can revert back to a one-master other-slave configuration once the ring is re-established.

WE CLAIM
1) An intelligent boiler tube leak detection system comprising,
A) Multiple acoustic sensors (1…n)
B) Data concentrator module (100) comprising ,

a) Microcontroller/processor (101) interfaced to the sensors (1….n) through RS485 ports (104), (105)
b) Keypad display interface (103)
c) Memory (106)
d) CODEC (107)
e) Audio output circuitry (108)
f) Serial port (109) interfaced to GSM module
g) Ether net port (110) interfaced to human machine interface and distributed control system
h) RS485 port (112) interfaced to expandable I/O modules i) On board digital I/Os (113) with respective drivers Characterized by digital signal processing, machine learning techniques,
automatic calibration, powerless configuration, automatic purging, self
diagnostics, remote streaming of noise.

2) An intelligent boiler tube leak detection system as claimed in claim (1), wherein, the sensors (1,….n) connected to two independent master ports in even-odd configuration, half at one port and half at another receiving information from at least half the sensors on line in the event of cable failure
3) An intelligent boiler tube leak detection system as claimed in claim (1), wherein, the sensors connected to two independent ports in fail-safe ring configuration, one port as master and the other as slave until a cable fault detected when both behaving as masters.
4) An intelligent boiler tube leak detection system as claimed in claim (1), wherein, an audio output from the sensor, polled and streamed at the control panel for processing and detection.
5) An intelligent boiler tube leak detection system as claimed in claim (1), wherein, machine learning algorithmic techniques adapted for leak detection.
6) An intelligent boiler tube leak detection system as claimed in claim (1), wherein, the trends detected in advance by the data concentrator module (100), purging the sensors automatically.

7) An intelligent boiler tube leak detection system as claimed in claim (1), wherein nos. of sensors and I/Os expanded by adding additional nodes to the RS485 network.
8) An intelligent boiler tube leak detection system as claimed in claim (1), wherein, immediate information made available to the user through SMS, e-mail or web server.