Monitoring systems and methods for sewer and other conduit systems

TECHNICAL FIELD [0002] The invention relates generally to monitoring systems and methods for sewage systems and other conduit systems. In particular, the invention relates to sensor networks capable of predicting and detecting clogs and cracks in a sewage system, other liquid supply system or other gas supply system, so that preventive and emergency maintenance can be performed with a rational distribution of resources, and significant damage and losses can be prevented. This sensor network uses audio signals and / or radio frequency (RF).

BACKGROUND ART [0003] There are over 34,000 sewage system overflows per year in the United States, resulting in the leakage of over 850 billion gallons of untreated sewage. For example, in Charlotte, North Carolina, in 2007, more than 400 overflows of sanitary sewage systems (SSO) occurred in Charlotte-Mecklenburg Utilities (CMU), resulting in the leakage of over 1.6 million gallons of sewage. Such overflow of sewage systems is usually caused by obstructions. Obstructions cause untreated sewage to leak from wells to the streets, to public and private properties and to waterways. This results in property damage, environmental problems and, in the most serious cases, threats to public health and safety. The complexity and dynamic nature of the sewage system translates into significant maintenance challenges. Creating a basic maintenance policy is an engineering compromise: excessive maintenance results in an unreasonable distribution of resources and unnecessary costs, and inadequate maintenance results in a greater risk of overcrowding. Hence, in this and other industries there is a need for reliable and economical systems and methods for predicting and detecting clogs and cracks in a sewage system, other liquid supply system, or other gas supply system, so that preventive and emergency maintenance with rational resource allocation can be carried out , and significant damage and losses can be prevented. US 2003/033879 A1 describes channel monitoring by transmitting a channel signal in a direction parallel to the predicted flow direction and by detecting obstruction in the associated channel section based on a reflected signal measured by channel sensors.

SUMMARY OF THE INVENTION [0004] In various embodiments, the invention provides monitoring systems and methods for sewage systems and other piping systems. These systems and monitoring methods enable water and sewage companies and similar to implement Dynamic Preventive Maintenance Systems (DPMS) and set up maintenance programs at the right time to minimize combined sewage overflows (CSOs), including both wastewater and rainwater, and SSOs involving only wastewater . For example, monitoring systems and methods allow the detection of obstruction in low flow pipes due to fat clogging, root lumps and the like, as well as cracks. Maintenance efforts can then be targeted at areas where

Problems are most likely as well as where they already occur. Those skilled in the art will readily recognize that while many of the examples provided herein include sewage systems, systems and monitoring methods provided herein are applicable to any piping system such as oil supply systems, other liquid supply systems, gas supply systems and the like, there are any systems, which include a network of "pipes" for the supply of one or more "materials" (the terms being given the widest possible interpretation).

[0005] In one embodiment, the invention provides a monitoring system for pipe networks as defined in claim 1. Preferably, each of the sensor devices includes one of: an audio sensor device or a radio frequency sensor device, and communicates through the medium. Each of the sensor devices is capable of detecting obstruction or cracks in the associated channel section by measuring the attribute of the transmitted / received signal, alone or in combination with another sensor device. Each sensor device is capable of communicating information about obstruction or rupture to a central location via a wireless link. The monitoring system also includes a pipe network predictive model provided in a central location and capable of processing information about obstruction or fracture to create a dynamic predictive maintenance policy that dictates maintenance. The monitoring system also includes a historical database that parameterizes the predictive model of the pipe network. The monitoring system also includes an alarm mechanism that is triggered in a central location by information about obstruction or fracture if a predetermined obstruction or fracture threshold is exceeded. Optionally, the pipe network includes a sewage line network, pipe sections include a sewage line section, and pipe nodes include wells.

[0006] In another embodiment, the invention provides a monitoring method for a pipe network according to claim 1. 9. Preferably, each of the sensor devices includes one of: an audio sensor device or a radio frequency sensor device, and communicates through the medium. Each of the sensor devices is supported for detecting clogging or cracks in the associated pipe section by measuring the attribute of the transmitted / received signal, alone or in combination with another sensor device. Each of the sensor devices is capable of communicating information about the obstruction or rupture to a central location via a wireless link. The monitoring method also includes a pipe network predictive model provided in a central location and capable of processing information about obstruction or fracture to create a dynamic predictive maintenance policy that dictates maintenance activities. The monitoring method also includes a historical database that parameterizes the predictive model of the pipe network. The monitoring method also includes an alarm mechanism that is triggered in a central location by information about obstruction or fracture if a predetermined obstruction or fracture threshold is exceeded. Optionally, the pipe network includes a sewage line network, pipe sections include a sewage line section, and pipe nodes include wells.

[0007] In another embodiment, the invention provides a monitoring and maintenance method for a pipe network as defined in claim 1. 17. Optionally, the condition of the obstruction or rupture is assessed at a central location. Optionally, the condition of the obstruction or fracture is assessed at a central location using a predictive model of a pipe network capable of processing information about the obstruction or fracture to create a dynamic predictive maintenance policy that dictates maintenance activities. Optionally, the pipe network includes a sewage line network,

Pipe sections include sewage line sections and pipe nodes include wells.

DESCRIPTION OF THE DRAWINGS [0008] The invention has been illustrated and described with reference to various drawing figures, in which the same reference numerals are used to designate the same system elements / method steps, according to the case, in which:

Fig. 1 is a schematic diagram illustrating one embodiment of a DPMS backbone according to the invention;

Fig. 2 is a schematic diagram illustrating one embodiment of a wireless sewage line audio sensor (SL-WASN) network according to the invention (which is conceptually similar to a wireless RF sensor network sewage line according to the invention);

Fig. 3 is a graph illustrating the evaluation of the sound attenuation coefficient for an 8-inch polyvinyl chloride (PVC) pipe;

Fig. 4 is a block diagram illustrating one embodiment of the measurement process used by SL-WASN of Fig. 2 for obtaining a set of obstruction estimation features used to classify the obstruction of an endangered sewer line section;

Fig. 5 is a block diagram illustrating one embodiment of estimating the impulse response associated with the measurement process of Fig. 4.

Fig. 6 is a schematic diagram illustrating one embodiment of the acoustic sensor node in SL-WASN in Fig. 2;

Fig. 7 is a schematic diagram illustrating one embodiment of the SL-WASN protocol according to the invention, where occlusion data is communicated from an ABCD sensor node;

Fig. 8 is a schematic diagram illustrating another, more general, embodiment of a DPMS backbone according to the invention; and

Fig. 9 is a schematic diagram depicting one embodiment of an audio sensor node according to the invention, highlighting its components.

DESCRIPTION OF IMPLEMENTATION EXAMPLES [0009] With reference to Figure 1, DPMS 10 according to the invention uses real-time observation of endangered sections 12 of the sewer line to create a policy of dynamic predictive maintenance. This dynamic 14 predictive maintenance policy "at the right time" minimizes the economic losses associated with conventional maintenance scheduling of endangered sections 12 of the sewer line.

[0010] Conventional maintenance scheduling typically uses prior knowledge and informal risk assessment analysis to minimize CSO and SSO. For example, it is known that sewer sections located near restaurants and the like require regular maintenance due to improper handling of fats.

[0011] As the random distribution of CSOs and SSOs across many cities indicates, applying current best practices in assessing endangered sections 12 of sewage lines is not enough. The use of historical data, weather data, cultural patterns and leaf fall patterns allows the selection of specific areas in the water and sewage enterprise network as more at risk of CSO and SSO, but without actual insight into specific sections of the sewage lines at risk of overflow. Similarly, collection pipes, which often run along secluded streams or river basins, are particularly large

Threat. Due to their remote location and high flow capacity, obstruction or rupture may result in a million gallons of wastewater or more being leaked prior to repair. Using this knowledge facilitates the commercial viability of the DPMS backbone according to the invention by limiting the implementation of real-time DPMS observation deployment on endangered sections of 12 sewer lines with high risk of overfill or high costs if overfill occurs.

[0012] High risk sewer lines or aperiodic SSO are candidates for using the DPMS backbone of the invention. As shown in Figure 1, DPMS 10 of the invention deploys SL-WASN 16 capable of performing real-time observations. Similarly, the RF equivalent of SL-WASN 16 can be arranged. SL-WASN 16 is a key element for monitoring endangered sections of 12 sewage lines, as well as provided communication technology to transfer real-time observations back to the headquarters. Real-time observations are then introduced to the 18-sewer line predictive model, which uses two-turn analysis two port analysis), a set of neighborhood features, condition estimation, hidden Markov model and / or similar tools that provide maintenance risk assessment analysis. Both the predictive model 18 of the sewage line and the analysis 20 of the maintenance risk assessment are parameterized by the prior knowledge 22 described above; including historical data, weather data, cultural patterns, leaf fall patterns and / or the like. It should be noted that for other applications, other than sewage lines, other factors may apply. Maintenance Risk Assessment Analysis 20 is used to formulate dynamic predictive maintenance policy 14 and perform maintenance activities 24 about which information is returned back to predictive model 18 of the sewer line and maintenance risk assessment analysis 20. As described above, real-time observations 26 are used to inform about the location and operation of SL-WASN 16.

[0013] Each of the DPMS elements 10 is described in more detail below. DPMS 10 is based on the deployment of SL-WASN 16 to enable and implement the DPMS backbone. The elements of DPMS 10 are interrelated to create a turnkey solution for plumbing enterprises for managing endangered sections of 12 sewage lines, or for other units, for managing sections of other pipe networks in a similar manner. In this context, DPMS 10 according to the invention is very adaptive and robust.

[0014] The analysis of maintenance risk assessment 20 is carried out in real time for the part of the water supply and sewage system network covered by DPMS 10. This assessment is based on data transferred to the central office from SL-WASN 16. The data provided by SL-WASN 16 is a set of features characterizing the obstruction for each pipe section, it is a set of features for estimating the obstruction. Using a set of obstruction assessment features, state assessments for obstruction are performed. The specific composition of the set of obstruction estimation features depends on the application. The requirements for the complexity and implementation of maintenance risk assessment analysis 20 are based on the features of the set of features for estimating obstruction and noise processes associated with measuring data. These factors affect the ability to distinguish between different states of obstruction, i.e. the degree to which a pipe section is obstructed or cracked.

[0015] A classic Bayesian risk assessment analysis was used to present the conservation risk assessment analysis, where optimal maintenance performance is based on minimizing the following conditional risk on a set of all possible actions:

EP # 404 151 B1 # stan = XA (aj | wj) Pr ​​[wj | z] (1) t = and where:

and_t is the ith maintenance activity, for example cleaning the affected sewer section 12 during the next month.

is the jth condition of the affected sewer line section, for example, the endangered section 12 sewer line is 25% obstructed.

/ V [<<)_;| ..V | is the probability of the state at a given observation x. is the state of true obstruction, ax is the observation from which the state of true obstruction is inferred.

AND(^and> j) is a weight function representing the loss associated with the performance of the maintenance action a_t for a given state of obstruction, for example the loss associated with waiting for a month for maintenance, assuming that the endangered section 12 of the sewage line is currently 25% obstructed.

[0016] Bayesian risk assessment is only one possible approach, and is provided here as a non-limiting example. Due to the temporary feature of obstruction, it is possible that the use of the Markov model or hidden Markov model, well known to those skilled in the art, can increase the reliability of maintenance risk assessment analysis.

[0017] Real-time observations of the affected sections 12 of the sewage line provide measured parameters to estimate the current state of obstruction. Real-time observations are transferred to the headquarters, allowing for adaptive dynamic maintenance policy 14. SLWASN 16 according to the invention includes two fundamental and innovative technical innovations: (1) directly detects features of obstruction / cracks and 2) uses a wireless communication network located in the sewage network. These two fundamental and innovative technical innovations provide significant benefits over current CSO and SSO detection technologies where detection is assessed indirectly. For example, at present, flow rates are measured at divergent locations in the sewage network and / or radar is used to study obstruction / cracks locally. In the first case, each flowmeter measurement is transmitted directly through the above-ground cellular network to the exchange. At the central, flowmeter data, along with wastewater flow models, are used to assess obstructions throughout the entire sewer network. Depending on the number of flowmeters, this approach may be useful for detecting problems in the largest diameter pipes. However, conventional technologies suffer from the inability to detect problems in the power lines with the most overflows. Due to the high costs associated with both flow meters and the initial and recurring service costs for each cell modem, increasing the number of deployed flow meters is not available. Even with an increased number of flowmeters, this approach has the inherent limitation of indirect measurements that limits early detection of obstruction.

[0018] Referring to Figure 2, the ABC sewer line 30 is a typical supply line with an 8-inch diameter pipe with well A 32, well B 34 and well C 36 located approximately 240-500 feet apart. The ABC 30 sewage line feeds the CDE 38 section of the collection line. This example area of ​​a sewer network has a high risk of overflow. Hence, cheap wireless audio sensors or are located in appropriate wells at regular intervals in the sewer network

RF sensors / transceivers (i.e. sensor node A 40, sensor node B 42, sensor node C 44, sensor node D 46, sensor node E 48). For audio sensors, each of these sensor nodes includes readily available commercially available components as shown in the embodiment of Figure 9. The sensor nodes are collectively capable of exchanging audio signals that are predictably affected by obstructions / cracks. In the case of RF, each of the sensor nodes also includes commercially available components. The sensor nodes are collectively capable of exchanging RF signals that are predictably affected by obstructions / cracks. The sensor nodes configure themselves automatically in the sensor network, enabling sensor data to be transmitted from each sensor node through the sensor network to the collection point, i.e. in the example shown to the sensor node E 48. The collection points provide the SL-WASN 16 gateway to the cellular network, through which occlusion / crack assessments in each section of the sewer network covered by the sensor network are transferred back to the control panel. The sensor network topology can be configured to best support sewage overflow detection in the plumbing region. Many sensor networks can be installed in different regions of the sewer network for direct monitoring of all endangered sections 12 of the sewer line (Fig. 1). Again, other piping networks can be configured similarly to the illustrated sewer network.

[0019] By measuring both the transmission and the reflection of the audio / RF signals between the sensor nodes, it is possible to directly assess the obstruction / fracture of 50 sections of the sewage line. The effect on both the transmission and the reflection of the sound waves / RF signals will depend on fat accumulation, root blockage, other obstacles, as well as cracking. This is shown between node A 40 and node B 42, where the ABC section 30 of the sewage line is blocked by partial obstruction 50. Basic signal measurements for obstruction / crack 50 include: 1) measuring transmission and reflection characteristics of audio / RF signals between sensor nodes, 2) measuring only transmission characteristics of audio / RF signals between sensor nodes and 3) in the case of audio, passive detection of ambient sound waves. For example, the second and third measurements can be used with each other. In the assessment of obstruction / crack 50 based on these measurements, other signal attributes can be used, such as, for example: 1) signal attenuation - the least complex to estimate and probably sufficient - based on the transmission between two nodes and 2) impulse response - Fig. 4 - more complete way.

[0020] Optionally, the same audio technology that was used to detect the condition of the affected sections of the sewer line 12 (Fig. 1) can also be used for wireless communication. Audio communication is usually limited to underwater applications or data communication via a wired computer modem. Audio communication is rarely (if at all) used for wireless sensor network applications due to the limited range of sound wave propagation in free space. For pipe network applications of the invention, the pipes act as waveguides, making audio communication a viable choice.

[0021] For RF, the finite element method can be used to estimate propagation characteristics based on empirical measurements:

Pr = ^pt ^{- α}(Α, σ) · Χ ά ^— Acl (2) where the received power at distance d, PR (d), and the transmission power Pt, are expressed in dBm. Multimode attenuation α expressed in dB / m depends on conductivity, σ, and waveguide radius, a. Losses of the ACL antenna connection are

EP 2 404 151 B1 expressed in dB. The results of the measurement campaign carried out, for example, in the 2.4 GHz band are summarized in Table 2.

<td>Pipe diameter (m)</td><td>Ac antenna connection losses (dB)</td><td>Attenuation, α (dB / m)</td>

<td> 0,30</td><td> 22,93</td><td> 4,80</td>

<td> 0,76</td><td> 41,56</td><td> 0,74</td>

<td> 1,07</td><td> 44,83</td><td> 0,78</td>

<td> 1,37</td><td> 43,39</td><td> 0,76</td>

Table 2. Measurement campaign carried out in the 2.4 GHz band

The challenge associated with the use of RF signals in the SL-WASN application is presented using the results in Table 2. For reliable communication with commercially available transceivers operating in the 2.4 GHz band [Pt-Pr (d)] max = 120 dB. Hence, for a 1-foot (0.3 m) pipe dmax^and 66 feet (20 m), and 4.5 ft (1.37 m) pipe dmax^and 330 feet (100 m), the estimated maximum distance is significantly smaller than the desired maximum distance for SL-WASN application. Based on both theory and empirical measurements, as the diameter of the pipe decreases, the number of supported modes also decreases, resulting in higher signal attenuation. By using a higher frequency RF signal, it is possible to reduce attenuation, but this involves higher implementation costs. This increase may be acceptable in some applications and the use of an RF based system may not be missed.

[0022] Propagation of audio waves in pipes is a classic theme with a significant theoretical and research basis. Sewer pipes and similar pipes are natural audio waveguides. The basic theory of audio propagation based on laboratory measurements of sound absorption in liquids contained in cylindrical tubes indicates the following relationship between the pressure amplitude P (-) at distance d2 compared to the pressure amplitude P (-) at distance di:

P (d₂} = Pidje - ^ - ^ (3) where α is the sound absorption coefficient. For liquids contained in pipes, the absorption coefficients must take into account the anti-flow resistance provided on the pipe walls. The flow of liquid in the pipe causes a phase difference between the speed of the liquid particle and the sound pressure due to the walls of the pipe. In addition to attenuation, the phase difference results in the presence of acoustic scattering. Acoustic attenuation in pipes can be well modeled with:

where α is the diameter of the pipe, c is the speed of sound, ne is the modified viscosity coefficient, ω is the frequency of the audio signal, and ρ is the density of the liquid. The modified viscosity coefficient, ne, takes into account the phenomenon of thermal conductivity on the walls of the pipes:

<img file="PL2404151T3_D0001.tif" />

where Y is the ratio of specific heat, κ is the thermal conductivity, Cp is the specific heat at constant pressure. Fig. 3 shows an estimate of the audio attenuation factor using typical constants for an 8-inch PVC pipe. Initial audio propagation data was collected on an inactive section of the sewer line using first generation measuring equipment, which in the transmission system7

The receiver used a 0.5 watt audio signal and sound pressure meter as a receiver with a frequency band from 20 Hz to 20 kHz. Fig. 3 also shows a preliminary evaluation of the audio suppression factor based on measurement data. As expected, the empirically estimated ratio is greater than the analytically estimated ratio.

[0023] The feasibility of using audio signals for SL-WASN applications is motivated by the results of Fig. 3. To ensure reliable communication with commercially available audio equipment, [P (di) -P (d2)] max = 90 dB. For a typical 8-inch supply sewer pipe, based on empirical estimation of the audio attenuation coefficient, dmax^and 3475 feet (1059 m) at 600 Hz and max^and 769 feet (234 m) at 10 kHz. The estimated maximum distance is generally greater than the desired maximum distance of 500 feet required for SL-WASN applications. In addition, the audio attenuation factor is inversely proportional to the radius of the pipe; hence, for pipes with a larger diameter, the audio attenuation coefficient decreases proportionally, suggesting greater maximum transmission ranges.

[0024] Ultrasonic waves have long been used for imaging, and audio wave imaging is used to determine room acoustics for sound reproduction systems. For the SLWASN applications of the invention, a similar approach is used based on the assumption that, for a sufficiently short period of time, each section of the sewer line can be seen as a time-constant linear system (LTI). Due to the dynamics of the flow and changes in the flow level, the impulse response of the sewer line, h (t), is statistically unchanged for a limited period of time. This limited period of time is the coherence time for the sewer line section and is one of the parameters of the invention.

[0025] Fig. 4 shows a possible measurement process used in SL-WASN (Figs. 1 and 2) according to the invention to obtain the set of obstruction estimation features used to classify the obstruction. Nodes of sensors 40, 42, 44 are located at equal distances inside the wells. The sensor nodes 40, 42, 44 periodically estimate the impulse response associated with their reflections, e.g. hAA (t), as well as the impulse response associated with transmission from neighboring sensor nodes, e.g. hBA (t) and hCA (t). Each of the sensor nodes 40, 42, 44 in the vicinity performs sequential transmission and thus provides an approximate two-turn analysis for each section of the sewer line.

[0026] The impulse response can be estimated in several different ways. One more promising example method is shown in Figure 5, where the signal transmitted in the source, X (t), is the maximum length sequence (MLS). MLS have approximately constant power spectral density and autocorrelation expressed by Λ.:: Υ'Υϊ: ·). The impulse response of the tested sewer line section, i.e. the sewer line section between the source and the receiver, is:

and^T h (r} «Υ (τ) ~ 2 ^ J ^{x (t)}s^(t + ^ dt (6) assuming that T is sufficiently large and the impulse response is invariant in the range of 2T.

Again with reference to Fig. 4, obstruction in the tube affects the impulse response, e.g. hBB (t), hAA (t), hAB (t), hBA (t). The effect on the impulse response is due to obstruction, but also to a change in water levels up from the obstruction compared to those down from it. It should be noted that the variation of the impulse response up and down is also affected by obstruction in the AB section of the sewage line.

[0028] A set of obstruction estimation features based on two-turn impulse response analysis provides a significant degree of redundancy that improves detection reliability, but at a relatively low

The low cost of increasing computational complexity. An alternative approach is to assess the sound pressure level (SPL) of transmission over several frequency bands instead of assessing the impulse response. The evaluation of tradeoffs associated with different sets of obstruction estimation features for the development of a robust and computationally economical set of features is a major aspect of the invention, which includes testing to evaluate different obstruction conditions at different water flows and levels.

[0029] The flow rate and flow level values ​​change throughout the day in the sewage line section and may change in other periods of time in other piping systems. Therefore, based on the relative size of the obstruction to the water level of the sewer line, the obstruction may be obstructed by water. It is beneficial that fat obstructions can increase slowly, so statistics can be collected over many days to observe potential changes in obstruction characteristics. One of the challenges in developing SL-WASN 16 (Figures 1 and 2) is scheduling acoustic or RF detection of a sewage line with a correspondingly high occurrence rate, while minimizing the occurrence for saving limited battery power. In order for SL-WASN 16 to be economically profitable, the sensor nodes should work for at least a year when using in a sewage line, which is the frequency of inspection of entrances to manholes required by the Environmental Protection Agency (EPA). An additional aspect of the invention is to obtain estimates of daily and weekly variations in the set of features for estimating obstruction based on long-term field trials. They are used to develop an acoustic or RF detection schedule and an optimized scheduling algorithm.

[0030] Referring to Fig. 6, in one embodiment, the SLWASN acoustic sensor node 54 (Figs. 1 and 2) includes a first set of elements 56 and a second set of elements 58 that are electrically connected. Optionally, a first set of 56 elements and a second set of 58 elements are integrated. The first set of elements 56 includes a microphone 60, a microphone preamplifier 62, an analog-to-digital (A / C) converter 64, a digital-to-analog (C / A) 70 converter, an audio amplifier 72 and a speaker 74. The analog-to-digital converter 64 and the digital-to-analog (C / A) converter 70 are electrically connected to the microcontroller 66 and associated data carrier 64 of the second set of 58 elements. Optionally, the microphone and speaker subsystem is designed and manufactured using commercially available audio components. The microcontroller and data storage subsystem can be ordered as a development board to reduce development time and costs. Current technology is available for integrating the cellular modem into the embedded system; and the invention addresses the challenge of receiving a cellular RF signal inside the well entrance, or placing the cell antenna outside the well entrance.

[0031] Based on the preliminary assessment of the sewer line propagation features, the upper limit of the operating frequency may be 20 kHz. This corresponds well to the consumer market for mobile devices and mp3 as well as home cinema systems. Therefore, the cost of audio components is based on existing markets with large economies of scale. In addition, many consumer products have similar battery limitations and, therefore, have predefined energy saving modes embedded in their integrated circuits (ICs).

[0032] The main design challenge of the invention includes transceiver systems; both microphone and speaker. Due to possible exposure to the flow of water and chemicals from industrial wastewater, the microphone and speaker must be able to work in a harsh environment. For example, they can be used

EP 2 404 151 B1 microphones and speakers designed for harsh environmental conditions, with the ability to work in immersion up to 15 m under water, in addition to corrosion resistance.

[0033] Using a sewer line network as a communication channel presents many problems: thermal variability, pipe bends, manhole entrances, differences in pipe surfaces, differences in flow rates and levels, surface noises (cars, trucks and the like), flow noise water and other sources of noise in the sewage system (for example, animals and insects). The physical (PHY) audio communication layer must adapt to the current communication channel conditions to maximize channel capacity with a sufficiently low transmission error rate. A definite advantage in the case of SL-WASN (Figs. 1 and 2) - the audio signal has a double role of acoustic detection and communication. Acoustic detection results can be used to adjust the PHY layer signal processing parameters, e.g. channel equalizer and transmission power. The only limitation is that the communication transmission must occur during the coherence of the acoustic detection of the sewage line. The upper layers of the protocol are developed based on the adaptation of significant achievements from energy-efficient wireless sensor networks.

[0034] Fig. 7 illustrates one exemplary embodiment of the SL-WASN protocol according to the invention, where occlusion data is communicated from an ABCD sensor node. Transmissions are packet transmissions with error detection and acknowledgment. The MAC layer is based on combined time division multiple access (TDMA) and multi access with media state tracking and collision avoidance (CSMA / CA). Each sensor node is allocated a transmission window in which it is to successfully transmit; hence, no other sensor nodes interfere with its transmission at the intended receiving node. CSMA / CA is used to minimize the impact of co-channel noise sources. For this reason, before transmitting, the sensor node listens to the channel for noise sources and steps away to avoid transmissions when they occur. The spacing algorithm and the length of the communication window are based on the features of noise sources present in the sewage line. The length of the communication window is set to allow multiple retransmissions to ensure a reliable communication link. Time synchronization between sensor nodes is achieved through the transmission of acoustic detection. The sensor nodes work with a very short duty cycle and remain off for a large part of the time.

[0035] Referring to Figure 8, in another, more general embodiment, which summarizes the systems / methods described above, DPMS 80 according to the invention uses real-time observation 82 of endangered sewer line 84 to create a dynamic predictive maintenance policy 86. The policy of dynamic predictive maintenance at the right time minimizes the economic losses associated with conventional maintenance scheduling on endangered sections of the 84 sewage line. High risk sewer lines and aperiodic SSO are candidates for implementing the DPMS backbone of the invention. As shown in fig. 8 The DPMS 80 of the invention deploys a wireless sewage line sensor network (SL-WASN) 86 capable of performing real-time observations. SL-WASN 88 can be based on audio, based on RF or based on an equivalent medium / technology. SL-WASN 86 is a key element for monitoring endangered sections 84 of sewage lines as well as provided communication technology to transfer observations 82 in real time back to the headquarters. Real-time observations 82 are then introduced into a sewage line predictive model that uses two-turn analysis, a set of neighborhood features, state estimation, hidden Markov model and / or

EP 2 404 151 B1 similar tools that provide analysis 90 maintenance risk assessment. Both the sewer line prediction model and maintenance risk assessment analysis are parameterized by the previous knowledge described above; including historical data, weather data, cultural patterns, leaf fall patterns and / or the like. It should be noted that for other applications, other than sewage lines, other factors may apply. Maintenance risk assessment analysis is used to formulate dynamic predictive maintenance policy 86 and perform maintenance actions 92, information about which is fed back to the sewage line prediction model and maintenance risk assessment analysis. As described above, real-time observations 82 are used to inform about the location and operation of SL-WASN 88.

[0036] Although the invention has been shown and described herein with reference to preferred embodiments and specific examples thereof, those skilled in the art will readily recognize that other variants and examples are within the scope of the invention as long as they fall within the definitions of the following claims.