Energy-Recovery Pressure-Reducer in District Heating System

: Already existing man-made infrastructures that create water ﬂow and unused pressure are interesting energy sources to which micro-hydropower plants can be applied. Apart from water supply systems (WSSs), which are widely described in the literature, signiﬁcant hydropower potential can also be found in district heating systems (DHSs). In this paper, a prototype, a so-called energy-recovery pressure-reducer (ERPR), utilized for a DHS, is presented. It consisted of a pump as a turbine coupled to a permanent magnet synchronous generator (PMSG). The latter was connected to the power grid through the power electronic unit (PEU). The variable-speed operation allowed one to modify the turbine characteristics to match the substation’s hydraulic conditions. The proposed ERPR device could be installed in series to the existing classic pressure reducing valve (PRV) as an independent device that reduces costs and simpliﬁes system installation. The test results of the prototype system located in a substation of Cracow’s DHS are presented. The steady-state curves and regulation characteristics show the prototype’s operating range and efﬁciency. In this study, the pressure-reducer impact on the electrical and hydraulic systems, and on the environment, were analyzed. The operation tests during the annual heating season revealed an average system’s efﬁciency of 49%. exceed 7.1 mm/s RMS are deﬁned as unacceptable, and are a danger for the machine. All measurements satisﬁed the conditions, and the difference in vibration results of the diode bridge rectiﬁer and sinusoidal rectiﬁer was small. However, the substantial increase of the vibration values of the return pipe were visible. This shows that the ERPR vibrations were transmitted with a high magnitude through the feed pipe and heat exchanger to the return pipe.


Introduction
Micro-hydropower plants (MHPs), whose power output is typically below 100 kW, have gained increasing attention due to their energy recovery capability in existing infrastructures [1]. A potential energy source can be found in any system that comprises water flow and unused pressure, i.e., water networks dealing with drinking water, wastewater (raw or treated), irrigation water, runoff water (rain or storm water), as well as cooling and heating systems [2,3]. Besides respecting basic water parameters (flow rate and pressure), the MHP can be designed according to water quality (wastewater) and temperature (cooling and heating systems). The generation of electrical energy is a supplementary function of water infrastructures. Therefore, a bypass valve and a secure shutdown system for the turbine are required. MHPs as renewable energy sources provide benefits for both the supplier (reduction of transmission losses) and the customer (reliability and reduced price) [4].
In recent literature, water supply systems (WSSs) are emphasized as the main man-made infrastructure that offers micro-hydropower potential. WSSs that are located in lower topographic areas suffer from excess pressure. This energy surplus, which is commonly dissipated via pressure-reducing valves (PRVs) [5,6], can be recovered by applying MHPs. The excess pressure is mainly present in transmission pipes and between sub-grids, but also inside the network and at the building entrance. Pressure reduction also leads to a decreasing leakage problem because water loss increases with noise, as well as electrical and vibration signals, were investigated in detail. Furthermore, the ERPR operation during the annual heating season is presented and analyzed.

Energy Potential
The main function of a DHS is to supply the customer with heating energy. Considering the initial investment, maintenance, and operation costs, the most prevalent solution is a two-pipe system, as presented in Figure 1a. Here, hot water is transported from the heat source to a substation through one pipe and returned by another [25]. In indirect systems, the parameters of the district heating medium (water), which is delivered from the heat source, are converted into suitable parameters (demanded by the customer) by a heat exchanger in the substation (Figure 1c). The heat exchangers transfer heat from the primary loop (district heating) to the secondary loop (customer) without mixing the fluids. This process requires a specific flow rate and water temperature of the primary loop. The supply temperature is changed (from 75 • C up to 120 • C), depending on the outdoor temperature, by the heat source according to a specific profile. The flow rate is adjusted by the substation's differential pressure (between input and output). The heat source should provide a differential pressure in the last substation (usually the worst-situated) equal to or higher than the minimal disposal pressure ∆p min_dis (usually 0.15-0.2 MPa). The latter is needed to cover the conduction losses of the heat exchanger and the control valve requirement (Figure 1b). The differential disposal pressure of the remaining substations is higher.
Water 2018, 10, x FOR PEER REVIEW 3 of 16 analyzed. Therefore, noise, as well as electrical and vibration signals, were investigated in detail. Furthermore, the ERPR operation during the annual heating season is presented and analyzed.

Energy Potential
The main function of a DHS is to supply the customer with heating energy. Considering the initial investment, maintenance, and operation costs, the most prevalent solution is a two-pipe system, as presented in Figure 1a. Here, hot water is transported from the heat source to a substation through one pipe and returned by another [25]. In indirect systems, the parameters of the district heating medium (water), which is delivered from the heat source, are converted into suitable parameters (demanded by the customer) by a heat exchanger in the substation (Figure 1c). The heat exchangers transfer heat from the primary loop (district heating) to the secondary loop (customer) without mixing the fluids. This process requires a specific flow rate and water temperature of the primary loop. The supply temperature is changed (from 75 °C up to 120 °C), depending on the outdoor temperature, by the heat source according to a specific profile. The flow rate is adjusted by the substation's differential pressure (between input and output). The heat source should provide a differential pressure in the last substation (usually the worst-situated) equal to or higher than the minimal disposal pressure ∆pmin_dis (usually 0.15-0.2 MPa). The latter is needed to cover the conduction losses of the heat exchanger and the control valve requirement (Figure 1b). The differential disposal pressure of the remaining substations is higher. The differential excess pressure of the substation n (∆pn = ∆pdis_n − ∆pmin_dis) can be used by the ERPR to produce the electrical active power P (W) according to the formula: (1) where γ (N/m 3 ) is the specific weight of water, Q (m 3 /s) is the volumetric flow rate, H (m) is the head drop, η is the total ERPR efficiency, and ∆pn = γ H (MPa) is the differential hydrostatic excess pressure of substation n, where H = 106.4·∆pn under water temperature equals 100 °C.
The total ERPR efficiency is the product of the turbine efficiency (ηt = 0.5-0.85), which depends on the flow range and rotational speed, generator efficiency (ηg = 0.8-0.97), and efficiency of the PEU (ηp = 0.9-0.98): The actual ERPR efficiency depends on the operating parameters (flow rate and differential pressure) and applied technical devices (mainly the turbine type). The efficiency variation may be The differential excess pressure of the substation n (∆p n = ∆p dis_n − ∆p min_dis ) can be used by the ERPR to produce the electrical active power P (W) according to the formula: where γ (N/m 3 ) is the specific weight of water, Q (m 3 /s) is the volumetric flow rate, H (m) is the head drop, η is the total ERPR efficiency, and ∆p n = γ H (MPa) is the differential hydrostatic excess pressure of substation n, where H = 106.4·∆p n under water temperature equals 100 • C. The total ERPR efficiency is the product of the turbine efficiency (η t = 0.5-0.85), which depends on the flow range and rotational speed, generator efficiency (η g = 0.8-0.97), and efficiency of the PEU (η p = 0.9-0.98): Water 2018, 10, 787 4 of 16 The actual ERPR efficiency depends on the operating parameters (flow rate and differential pressure) and applied technical devices (mainly the turbine type). The efficiency variation may be significant (η = 0.35-0.8) [26][27][28][29]. Thus, the ERPR design process has key importance for the system's profitability.

Topology of Substation with ERPR
The heat demand of the customer is highly dependent on the outdoor temperature. Usually, the flow rate of the secondary loop remains constant, whereas the temperature is adjusted. The secondary loop's temperature as a function of the outdoor temperature, the so-called heating curve, is implemented in a substation-control system, which adjusts the PRV (Figure 2). A proper control of the PRV requires minimal disposal pressure. Each excess pressure is recovered by the ERPR in the form of electrical active power (Equation (1)).

Topology of Substation with ERPR
The heat demand of the customer is highly dependent on the outdoor temperature. Usually, the flow rate of the secondary loop remains constant, whereas the temperature is adjusted. The secondary loop's temperature as a function of the outdoor temperature, the so-called heating curve, is implemented in a substation-control system, which adjusts the PRV (Figure 2). A proper control of the PRV requires minimal disposal pressure. Each excess pressure is recovered by the ERPR in the form of electrical active power (Equation (1)). To maximize the energy production, two system-regulation modes can be used. As presented in Reference [19], better features (wider range of operating conditions) are obtained using the hydraulic regulation mode, which is realized by a series-parallel combination of PAT and PRV. However, this regulation mode requires two regulated PRVs and a modification to the substation's control system.
By applying the electrical regulation (ER) mode [20], the ERPR can be added in series to the existing PRV as an independent device without any modifications to the control system ( Figure 2). The system modification concerns only the ERPR bypass, which is required for a safe operation. In case of a situation where the differential excess pressure of the substation exceeds the operation range of the ERPR, or the ERPR breaks, the PRV takes over the pressure control. The ER enables the variable speed operation method. Therefore, the characteristic PAT curves are modified by changing the turbine speed to match the load conditions (instant flow rate and differential pressure). The low costs of the electromechanical components and engineering can shorten the payback period of this system. Furthermore, the installation process is simple, and does not require modifications to the existing control system. Therefore, this solution is proposed in this paper.
The ERPR should be carefully designed regarding the variability of the hydraulic conditions of the specific substation to maximize the average efficiency, which corresponds to the system capability [19]. This implies that the ranges of the most frequently occurring flow rates and differential excess pressures have to be covered by the PAT curves of the highest efficiency.

ERPR Unit
To reduce costs, the ERPR unit was based on the PAT. The turbine was connected to the PMSG, which has a higher and more stable efficiency under load variation than an induction machine. The To maximize the energy production, two system-regulation modes can be used. As presented in Reference [19], better features (wider range of operating conditions) are obtained using the hydraulic regulation mode, which is realized by a series-parallel combination of PAT and PRV. However, this regulation mode requires two regulated PRVs and a modification to the substation's control system.
By applying the electrical regulation (ER) mode [20], the ERPR can be added in series to the existing PRV as an independent device without any modifications to the control system ( Figure 2). The system modification concerns only the ERPR bypass, which is required for a safe operation. In case of a situation where the differential excess pressure of the substation exceeds the operation range of the ERPR, or the ERPR breaks, the PRV takes over the pressure control. The ER enables the variable speed operation method. Therefore, the characteristic PAT curves are modified by changing the turbine speed to match the load conditions (instant flow rate and differential pressure). The low costs of the electromechanical components and engineering can shorten the payback period of this system. Furthermore, the installation process is simple, and does not require modifications to the existing control system. Therefore, this solution is proposed in this paper.
The ERPR should be carefully designed regarding the variability of the hydraulic conditions of the specific substation to maximize the average efficiency, which corresponds to the system capability [19]. This implies that the ranges of the most frequently occurring flow rates and differential excess pressures have to be covered by the PAT curves of the highest efficiency.

ERPR Unit
To reduce costs, the ERPR unit was based on the PAT. The turbine was connected to the PMSG, which has a higher and more stable efficiency under load variation than an induction machine. The regulation process (control of turbine speed) was performed by a full-scale PEU. The PEU matches the parameters of the electrical energy generated by the PMSG to the power system requirements. This enabled an on-grid operation [30]. The structure of the proposed ERPR is presented in Figure 3. regulation process (control of turbine speed) was performed by a full-scale PEU. The PEU matches the parameters of the electrical energy generated by the PMSG to the power system requirements. This enabled an on-grid operation [30]. The structure of the proposed ERPR is presented in Figure 3. The ERPR controller measures the actual inlet and outlet pressure, as well as the flow rate. Different ERPR control targets are possible. It can regulate one of the hydraulic parameters (pressure drop, outlet pressure, or flow rate), maximal electrical power, or maximal efficiency. Furthermore, by integrating the ERPR into the substation control system, the water temperature of the secondary loop can be adjusted. However, it should be noted that the main function of the ERPR is the maximization of the energy recovery. Considering the limited operation range of the PAT, the substation control should be performed by the PRV. Thus, the control mode that provides maximal electrical active power and keeps the differential pressure higher or equal to the minimal disposal value seems to be the most appropriate.
Nowadays, there are many PEU solutions on the market that can be used for the proposed ERPR. The most popular are the two types presented in Figure 4.
The first system presented in Figure 4a consists of a diode bridge rectifier (AC/DC), a circuit that raises the DC voltage (DC/DC booster), and a PWM (pulse width modulation) inverter controlled by the DPC-SVM (direct power control with space vector modulation) algorithm [31,32]. The second type, shown in Figure 4b, employs a PWM rectifier (also called the sinusoidal rectifier) instead of a diode rectifier and a DC/DC booster.
The generator load can be controlled with the current limits limit s I and limit DC I of the PEU. The current limit changes allow for the modification of the regulation curves as a function of rotational speed [33,34]. This is presented in Figure 5.
The current limit of the inverter ( limit  Figure 6. It can be noticed that under some conditions the PEU and PAT torque curves can be parallel to each other. This situation may result in speed oscillations and even in an unstable control. Therefore, the ERPR control should be performed at the limit I limit. The presented figures The ERPR controller measures the actual inlet and outlet pressure, as well as the flow rate. Different ERPR control targets are possible. It can regulate one of the hydraulic parameters (pressure drop, outlet pressure, or flow rate), maximal electrical power, or maximal efficiency. Furthermore, by integrating the ERPR into the substation control system, the water temperature of the secondary loop can be adjusted. However, it should be noted that the main function of the ERPR is the maximization of the energy recovery. Considering the limited operation range of the PAT, the substation control should be performed by the PRV. Thus, the control mode that provides maximal electrical active power and keeps the differential pressure higher or equal to the minimal disposal value seems to be the most appropriate.
Nowadays, there are many PEU solutions on the market that can be used for the proposed ERPR. The most popular are the two types presented in Figure 4.
The first system presented in Figure 4a consists of a diode bridge rectifier (AC/DC), a circuit that raises the DC voltage (DC/DC booster), and a PWM (pulse width modulation) inverter controlled by the DPC-SVM (direct power control with space vector modulation) algorithm [31,32]. The second type, shown in Figure 4b, employs a PWM rectifier (also called the sinusoidal rectifier) instead of a diode rectifier and a DC/DC booster.
The generator load can be controlled with the current limits I limit s and I limit DC of the PEU. The current limit changes allow for the modification of the regulation curves as a function of rotational speed [33,34]. This is presented in Figure 5.
The current limit of the inverter (I limit s ) corresponds to the electrical active power of the PEU assuming a constant grid voltage and no reactive power. Furthermore, after neglecting the power losses of the PEU, it can be stated that I limit s controls the generator's active power (Figure 5a). The second control parameter I limit DC is proportional to the generator current, and in the case of a PMSG, it corresponds to the generator torque (Figure 5b). also contain the optimal operation curve that provides the maximal turbine power and, therefore, the control target.   The speed value, located at the cross point of the PEU and PAT torque characteristics, is set by for a specific reduced pressure, as presented in Figure 7. The main function of the ERPR controller is to adjust the turbine speed according to the set reduced pressure to provide the maximal electrical active power of the PEU. According to the chosen control mode, the reduced pressure is equal to the actual differential excess pressure of the substation. also contain the optimal operation curve that provides the maximal turbine power and, therefore, the control target.   The speed value, located at the cross point of the PEU and PAT torque characteristics, is set by for a specific reduced pressure, as presented in Figure 7. The main function of the ERPR controller is to adjust the turbine speed according to the set reduced pressure to provide the maximal electrical active power of the PEU. According to the chosen control mode, the reduced pressure is equal to the actual differential excess pressure of the substation. The selection of the control parameter must be done by considering the shape of the turbine curves. Examples of curves for different values of reduced pressure in the speed domain are presented in Figure 6. It can be noticed that under some conditions the PEU and PAT torque curves can be parallel to each other. This situation may result in speed oscillations and even in an unstable control. Therefore, the ERPR control should be performed at the I limit DC limit. The presented figures also contain the optimal operation curve that provides the maximal turbine power and, therefore, the control target.
The speed value, located at the cross point of the PEU and PAT torque characteristics, is set by I limit DC for a specific reduced pressure, as presented in Figure 7. The main function of the ERPR controller is to adjust the turbine speed according to the set reduced pressure to provide the maximal electrical active power of the PEU. According to the chosen control mode, the reduced pressure is equal to the actual differential excess pressure of the substation.

ERPR Prototype and Steady-State Characteristics
The ERPR prototype was installed in a DHS substation that contained a heat exchanger of 1.8 MW thermal power, and had a nominal flow rate of 0.0064 m 3 /s and an available differential disposal pressure of 0.7 MPa. The chosen turbine was a three-stage vertical PAT with a nominal flow rate of 0.0058 m 3 /s and a reduced pressure of 0.5 MPa (H = 53.2 m). The PAT was connected through boxcoupling to the PMSG with a synchronous speed of 3000 rpm. The PEU of 5.5 kW contained the two optional rectifier types (diode bridge rectifier with DC/DC booster and sinusoidal PWM rectifier), as presented in Figure 4. The rated values and nominal parameters are listed in the Appendix A (Table  A1). The ERPR prototype installed in the substation of the Cracow DHS is shown in Figure 8. All measurements were performed according to the ISO 9906 standard (grade 2B) [35]. The test installation was equipped with additional valves allowing for the control of the substation input pressure.
The ERPR performance was identified by the steady-state characteristics of the PEU active power (Figure 9a), estimated turbine torque (neglecting PMSG and PEU efficiency) (Figure 9b), and flow rate (Figure 9c) under four different values of reduced pressure (0.45, 0.55, 0.6, and 0.68 MPa). Additionally, the total ERPR efficiency was calculated (Equation (1)) and shown in Figure 9d. Figure  10 presents the electrical energy production and total efficiency in the flow-rate domain. The measurement points are marked by crosses.

ERPR Prototype and Steady-State Characteristics
The ERPR prototype was installed in a DHS substation that contained a heat exchanger of 1.8 MW thermal power, and had a nominal flow rate of 0.0064 m 3 /s and an available differential disposal pressure of 0.7 MPa. The chosen turbine was a three-stage vertical PAT with a nominal flow rate of 0.0058 m 3 /s and a reduced pressure of 0.5 MPa (H = 53.2 m). The PAT was connected through boxcoupling to the PMSG with a synchronous speed of 3000 rpm. The PEU of 5.5 kW contained the two optional rectifier types (diode bridge rectifier with DC/DC booster and sinusoidal PWM rectifier), as presented in Figure 4. The rated values and nominal parameters are listed in the Appendix A (Table  A1). The ERPR prototype installed in the substation of the Cracow DHS is shown in Figure 8. All measurements were performed according to the ISO 9906 standard (grade 2B) [35]. The test installation was equipped with additional valves allowing for the control of the substation input pressure.
The ERPR performance was identified by the steady-state characteristics of the PEU active power (Figure 9a), estimated turbine torque (neglecting PMSG and PEU efficiency) (Figure 9b), and flow rate (Figure 9c) under four different values of reduced pressure (0.45, 0.55, 0.6, and 0.68 MPa). Additionally, the total ERPR efficiency was calculated (Equation (1)) and shown in Figure 9d. Figure  10 presents the electrical energy production and total efficiency in the flow-rate domain. The measurement points are marked by crosses.

ERPR Prototype and Steady-State Characteristics
The ERPR prototype was installed in a DHS substation that contained a heat exchanger of 1.8 MW thermal power, and had a nominal flow rate of 0.0064 m 3 /s and an available differential disposal pressure of 0.7 MPa. The chosen turbine was a three-stage vertical PAT with a nominal flow rate of 0.0058 m 3 /s and a reduced pressure of 0.5 MPa (H = 53.2 m). The PAT was connected through box-coupling to the PMSG with a synchronous speed of 3000 rpm. The PEU of 5.5 kW contained the two optional rectifier types (diode bridge rectifier with DC/DC booster and sinusoidal PWM rectifier), as presented in Figure 4. The rated values and nominal parameters are listed in the Appendix A (Table A1). The ERPR prototype installed in the substation of the Cracow DHS is shown in Figure 8. All measurements were performed according to the ISO 9906 standard (grade 2B) [35]. The test installation was equipped with additional valves allowing for the control of the substation input pressure.
The ERPR performance was identified by the steady-state characteristics of the PEU active power (Figure 9a), estimated turbine torque (neglecting PMSG and PEU efficiency) (Figure 9b), and flow rate (Figure 9c) under four different values of reduced pressure (0.45, 0.55, 0.6, and 0.68 MPa). Additionally, the total ERPR efficiency was calculated (Equation (1)) and shown in Figure 9d. Figure 10 presents the electrical energy production and total efficiency in the flow-rate domain. The measurement points are marked by crosses.  To estimate the ERPR's operation range under different working conditions of the substation, Figure 11 is presented. It shows the speed-regulation characteristics of the ERPR in the reduced pressure-flow-rate domain with isolines of constant efficiency. The presented measurement results enabled an estimation of the operation range and capability of the ERPR prototype. The useful operation area lies between 80% and 140% of the nominal reduced pressure and between 85% and 120% of the nominal flow rate. Considering the maximal power-control mode, the turbine speed should be adjusted according to the differential excess disposal pressure (equals the reduced pressure) variation (optimal curve marked with black bold line). This operation leads to a total ERPR efficiency of more than 50%.   To estimate the ERPR's operation range under different working conditions of the substation, Figure 11 is presented. It shows the speed-regulation characteristics of the ERPR in the reduced pressure-flow-rate domain with isolines of constant efficiency. The presented measurement results enabled an estimation of the operation range and capability of the ERPR prototype. The useful operation area lies between 80% and 140% of the nominal reduced pressure and between 85% and 120% of the nominal flow rate. Considering the maximal power-control mode, the turbine speed should be adjusted according to the differential excess disposal pressure (equals the reduced pressure) variation (optimal curve marked with black bold line). This operation leads to a total ERPR efficiency of more than 50%. To estimate the ERPR's operation range under different working conditions of the substation, Figure 11 is presented. It shows the speed-regulation characteristics of the ERPR in the reduced pressure-flow-rate domain with isolines of constant efficiency. The presented measurement results enabled an estimation of the operation range and capability of the ERPR prototype. The useful operation area lies between 80% and 140% of the nominal reduced pressure and between 85% and 120% of the nominal flow rate. Considering the maximal power-control mode, the turbine speed should be adjusted according to the differential excess disposal pressure (equals the reduced pressure) variation (optimal curve marked with black bold line). This operation leads to a total ERPR efficiency of more than 50%.

ERPR Impact Analysis
The ERPR device was installed between a hydraulic system and an electrical power system. Thus, the negative influence of the ERPR on these systems needed to be analyzed. Additionally, the influence on the environment (sound measurements) needed to be estimated.
The prototype PEU contained two optional rectifiers, as described in Section 2.3. These two most popular structures differed in price, control complexity, and signal quality of the generator. The bad quality of electrical signals may have induced additional noise and vibrations of the ERPR during operation. This issue was investigated in this section by analyzing the quality of the generator's electrical signals, sound, and ERPR vibrations.
The curves of the generator's and PEU's electrical signals (voltage and current) over time are presented below. Figure 12 presents the results of the PEU with a diode bridge rectifier (Figure 4a), whereas Figure 13 presents those of the system with a sinusoidal rectifier (Figure 4b). The presented measurements were performed under similar electrical generator conditions (1 kW power and 2100 rpm). In order to evaluate the higher harmonics, an FFT (Fast Fourier Transform) analysis was

ERPR Impact Analysis
The ERPR device was installed between a hydraulic system and an electrical power system. Thus, the negative influence of the ERPR on these systems needed to be analyzed. Additionally, the influence on the environment (sound measurements) needed to be estimated.
The prototype PEU contained two optional rectifiers, as described in Section 2.3. These two most popular structures differed in price, control complexity, and signal quality of the generator. The bad quality of electrical signals may have induced additional noise and vibrations of the ERPR during operation. This issue was investigated in this section by analyzing the quality of the generator's electrical signals, sound, and ERPR vibrations.
The curves of the generator's and PEU's electrical signals (voltage and current) over time are presented below. Figure 12 presents the results of the PEU with a diode bridge rectifier (Figure 4a), whereas Figure 13 presents those of the system with a sinusoidal rectifier (Figure 4b). The presented measurements were performed under similar electrical generator conditions (1 kW power and 2100 rpm). In order to evaluate the higher harmonics, an FFT (Fast Fourier Transform) analysis was

ERPR Impact Analysis
The ERPR device was installed between a hydraulic system and an electrical power system. Thus, the negative influence of the ERPR on these systems needed to be analyzed. Additionally, the influence on the environment (sound measurements) needed to be estimated.
The prototype PEU contained two optional rectifiers, as described in Section 2.3. These two most popular structures differed in price, control complexity, and signal quality of the generator. The bad quality of electrical signals may have induced additional noise and vibrations of the ERPR during operation. This issue was investigated in this section by analyzing the quality of the generator's electrical signals, sound, and ERPR vibrations.
The curves of the generator's and PEU's electrical signals (voltage and current) over time are presented below. Figure 12 presents the results of the PEU with a diode bridge rectifier (Figure 4a), whereas Figure 13 presents those of the system with a sinusoidal rectifier (Figure 4b). The presented measurements were performed under similar electrical generator conditions (1 kW power and 2100 rpm). In order to evaluate the higher harmonics, an FFT (Fast Fourier Transform) analysis was performed. As expected, the generator current for the case of the sinusoidal rectifier contained fewer higher harmonics in the range below 1 kHz. Furthermore, the harmonics' amplitudes were about 10 dB lower (with a reference level of 10 −3 A) compared to those of the diode bridge rectifier's system. The additional drawbacks of the diode rectifier were the voltage spikes during the commutation process and the angle shift between the generator voltage and current that results in reactive power. The latter increased the generator current. Therefore, the active power needed to be limited. To estimate the impact of the PEU on the electrical power system, the total harmonic distortion parameter (THD) of the PEU output voltage needed to be analyzed. For both system types, the voltage quality (THD U ≈ 1.5%) fulfils the standard requirements [36].
To estimate the influence of the generator current distortion on the vibrations of the ERPR and hydraulic pipes, the measurements of the vibration velocity in three directions were analyzed. Three measurement points located at the ERPR unit, feed, and return pipes, which are presented in Figure 14a, were selected. Three operating points (2200, 2500, and 2800 rpm) with the best efficiency were analyzed. The dominant vibrations occured in the x-and z-directions for the generator and return pipe. To evaluate the vibration severity, the classification introduced by the ISO 10816-1 standard was used [37]. The ERPR belonged to Class I (up to 15 kW), for which the satisfactory condition is defined in the vibration-velocity range below 2.8 mm/s RMS (root mean square). Values that exceed 7.1 mm/s RMS are defined as unacceptable, and are a danger for the machine. All measurements satisfied the conditions, and the difference in vibration results of the diode bridge rectifier and sinusoidal rectifier was small. However, the substantial increase of the vibration values of the return pipe were visible. This shows that the ERPR vibrations were transmitted with a high magnitude through the feed pipe and heat exchanger to the return pipe. performed. As expected, the generator current for the case of the sinusoidal rectifier contained fewer higher harmonics in the range below 1 kHz. Furthermore, the harmonics' amplitudes were about 10 dB lower (with a reference level of 10 −3 A) compared to those of the diode bridge rectifier's system. The additional drawbacks of the diode rectifier were the voltage spikes during the commutation process and the angle shift between the generator voltage and current that results in reactive power. The latter increased the generator current. Therefore, the active power needed to be limited. To estimate the impact of the PEU on the electrical power system, the total harmonic distortion parameter (THD) of the PEU output voltage needed to be analyzed. For both system types, the voltage quality (THDU ≈ 1.5%) fulfils the standard requirements [36].    To estimate the influence of the generator current distortion on the vibrations of the ERPR and hydraulic pipes, the measurements of the vibration velocity in three directions were analyzed. Three measurement points located at the ERPR unit, feed, and return pipes, which are presented in Figure 14a, were selected. Three operating points (2200, 2500, and 2800 rpm) with the best efficiency were analyzed. The dominant vibrations occured in the x-and z-directions for the generator and return pipe. To evaluate the vibration severity, the classification introduced by the ISO 10816-1 standard was used [37]. The ERPR belonged to Class I (up to 15 kW), for which the satisfactory condition is defined in the vibration-velocity range below 2.8 mm/s RMS (root mean square). Values that exceed 7.1 mm/s RMS are defined as unacceptable, and are a danger for the machine. All measurements satisfied the conditions, and the difference in vibration results of the diode bridge rectifier and sinusoidal rectifier was small. However, the substantial increase of the vibration values of the return pipe were visible. This shows that the ERPR vibrations were transmitted with a high magnitude through the feed pipe and heat exchanger to the return pipe.
The ERPR worked at variable speed depending on the substation operating conditions. Therefore, the vibration values in the speed domain needed to be analyzed. Figure 14b presents the vibration velocity of the generator in the x-direction as a function of speed. The two resonant speeds (2500 rpm and 3000 rpm) were visible, and both were within the ERPR regulation area ( Figure 11). Thus, they are especially dangerous. The vibration values for the sinusoidal rectifier were about 15% lower than those of the diode bridge rectifier in the second resonance area. Moreover, the resonance vibration for the diode bridge rectifier exceeded the satisfactory level. These phenomena must be considered while designing the mounting technique of the ERPR. In the presented ERPR prototype, two main noise sources existed. Apart from the ERPR and pipes vibrations, the PEU generated sound resulting from the semiconductor-switching elements that caused higher harmonics in the electrical signals. The measurements performed at a 1 m distance from the device, in accordance with the measurement standard in [38], showed similar noise values for both PEU structures. The noise level increased with load from 65 dB (background noise) to 68 dB under 1.1 kW, and further to 72 dB for the nominal ERPR power. These values fulfil European standards regarding device noise limits [39] and international safety standards for workers exposed to noise [40].
The main common noise source in the DHS is the cavitation phenomenon in the PRV. It occurs when pressure decreases immediately below the vapor pressure directly behind the PRV. One of the methods that minimizes the cavitation risk is reducing the differential pressure over the PRV [41]. The proposed ERPR decreased this pressure to the minimal disposal value in the region before the The ERPR worked at variable speed depending on the substation operating conditions. Therefore, the vibration values in the speed domain needed to be analyzed. Figure 14b presents the vibration velocity of the generator in the x-direction as a function of speed. The two resonant speeds (2500 rpm and 3000 rpm) were visible, and both were within the ERPR regulation area ( Figure 11). Thus, they are especially dangerous. The vibration values for the sinusoidal rectifier were about 15% lower than those of the diode bridge rectifier in the second resonance area. Moreover, the resonance vibration for the diode bridge rectifier exceeded the satisfactory level. These phenomena must be considered while designing the mounting technique of the ERPR.
In the presented ERPR prototype, two main noise sources existed. Apart from the ERPR and pipes vibrations, the PEU generated sound resulting from the semiconductor-switching elements that caused higher harmonics in the electrical signals. The measurements performed at a 1 m distance from the device, in accordance with the measurement standard in [38], showed similar noise values for both PEU structures. The noise level increased with load from 65 dB (background noise) to 68 dB under 1.1 kW, and further to 72 dB for the nominal ERPR power. These values fulfil European standards regarding device noise limits [39] and international safety standards for workers exposed to noise [40].
The main common noise source in the DHS is the cavitation phenomenon in the PRV. It occurs when pressure decreases immediately below the vapor pressure directly behind the PRV. One of the methods that minimizes the cavitation risk is reducing the differential pressure over the PRV [41]. The proposed ERPR decreased this pressure to the minimal disposal value in the region before the PRV. Thus, it prevented cavitation. An implementation of the ERPR enabled an exact sizing of the PRV, improving control conditions, and prolonging the service life of the control equipment.

Operation Analysis during Annual Heating Season
The presented ERPR prototype, which was located in the substation in the Cracow DHS, was tested during five months of the annual heating season (from December to April). Figure 15 shows the curves of the main system parameters (grey line) with a one-day average filter curve (thick line) over time.

Operation Analysis during Annual Heating Season
The presented ERPR prototype, which was located in the substation in the Cracow DHS, was tested during five months of the annual heating season (from December to April). Figure 15 shows the curves of the main system parameters (grey line) with a one-day average filter curve (thick line) over time.
The operation of the DHS substation was mainly dependent on outdoor temperatures. The nominal values of the ERPR (∆p = 0.5 MPa, Q = 0.0058 m 3 /s) were achieved when the outdoor temperature measured approximately 2 °C (dashed line). This is the usual average temperature for this location in the chosen period. The analyzed heating season was characterized by a quite high average temperature of 6.6 °C, which lead to only 2.42 MW of recovered energy. The average flow (Qavg = 0.0045 m 3 /s) and average reduced pressure (∆pavg = 0.25 MPa) were much below the nominal values. Most of the time the operating conditions were outside the ERPR regulation area ( Figure 11). Nevertheless, the average ERPR efficiency (49%) was satisfactory.  The operation of the DHS substation was mainly dependent on outdoor temperatures. The nominal values of the ERPR (∆p = 0.5 MPa, Q = 0.0058 m 3 /s) were achieved when the outdoor temperature measured approximately 2 • C (dashed line). This is the usual average temperature for this location in the chosen period. The analyzed heating season was characterized by a quite high average temperature of 6.6 • C, which lead to only 2.42 MW of recovered energy. The average flow (Q avg = 0.0045 m 3 /s) and average reduced pressure (∆p avg = 0.25 MPa) were much below the nominal values. Most of the time the operating conditions were outside the ERPR regulation area ( Figure 11). Nevertheless, the average ERPR efficiency (49%) was satisfactory.

Conclusions
In this paper, we proposed an ERPR device for a DHS. The chosen system structure, which consisted of an ERPR in series with a PRV, lowered installation costs and simplified system installation. To separate the ERPR control from the substation regulation system, an ERPR operation mode was proposed that provided the highest electrical active power, and kept the differential pressure equal to the minimal disposal value of the substation.
The presented analysis of the ERPR operation showed the low impact on electrical and hydraulic systems, as well as on the environment. The resulting vibration and noise values were similar for both PEU structures. Therefore, the cheaper solution with a diode bridge rectifier can be used. The vibration analysis in the speed domain revealed the resonance phenomenon, which must be considered while designing the mounting technique for the ERPR.
The presented example for energy recovery in a DHS in the city of Cracow indicated significant potential. Energy-recovery profitability depends on the equipment price and system efficiency. The economic analysis presented in Reference [21], resulting in a payback period of 5 to 7 years, assumes costs of approximately 1200 Euro for a standardized ERPR device (1 kW), and 1700 Euro if an extra design process is necessary. The total costs of the presented prototype were 1900 Euro/kW. Considering large ERPR numbers and equipment optimization, these cost levels are achievable. Furthermore, the presented analysis of the ERPR prototype in the substation of the Cracow DHS showed that the average total efficiency in the heating season is 49% despite the unusual operation conditions resulting from unusually high outdoor temperatures. The efficiency value confirmed the assumption made in Reference [21] (average efficiency value of 50%).
In conclusion, the installation of the ERPR in the Cracow DHS was economically feasible. To apply these results to other DHSs, factors such as the specific network structure, local market, and legal nature conditions should be considered.
Author Contributions: D.B. performed the data collection, analysed the data and wrote the paper; T.W. supervised the work and reviewed the final manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.